the 1/focal length rule

[sharky]

This is something which is probably pretty obvious, but I seem to like asking obvious questions. On digital bodies, the 1.6x crop factor makes a 200mm lens an "effective" 320mm, so what does that do to the 1/focal length rule? Would 1/250th still be adequate, or would you need to use higher shutter speeds?

I'm assuming the former because the focal length really hasn't changed, but I've noticed that a lot of shots taken at 200mm with the 10D don't seem as steady as with my old film body.

[slin100]

The rule of thumb should be refactored to take into account the effective focal length. So, a 200mm lens should used with a shutter speed no less than 1/320. Of course, this is merely a rule of thumb. YMMV.

[Chazs]

In my SLR days 20 years ago I used the 1/(2FL) rule, but I just posted a photo taken with a 200mm at 1/250 hand-held. It would have been nice to get to 1/500, or 1/300; I just had to hold me breath better.

My dad has always been able to and-hold a 50mm to 1/4s. and on a good day to 1/2s. Guess I drink too much coffee. :oops:

If you are used to the 1/FL rule with a standard SLR, you should be okay with a 1/(1.6FL) on a DSLR. Me, I'm not usually comfortable until at least 1/(2.5FL). Next lens I get may be an IS type.

[sharky]

I actually sold a 28-135IS to buy the 70-200L, and while I like a lot of things about the L, I never realised how reliant I'd become on IS.

[Steven M. Anthony]

The 1.6 crop factor associated with many digital cameras does NOT change the effective focal length of a lens. It does NOT make a 200mm lens a 320mm lens. If it did, any time you wanted a 1,000mm lens, all you would have to do is put duct tape over most of your sensor and a thrifty-50 would be magically converted into a 1,000 mm prime!

The 1.6 crop factor is not magic. It is simply a crop factor--a crop factor relative to the content that would be contained on a full frame of 35mm film.

Here's an example: Close one eye and look at something across the room. Go ahead, I'll wait... Okay, now roll the fingers of one hand up so your finger tips touch the palm of your hand. Now hold your hand up to your opened eye and look through the little tunnel in your hand at the same object across the room. Did the effective focal length of your eye change? No. Did the object across the room look closer the second time you looked at it? No. All that happened by looking through the tunnel in your hand was that you reduced (i.e., cropped) your field of view.

That's how the 1.6 crop factor works. If you put the same lens on a film camera, everything will look as far away as it did through the digital camera--but on the film camera, there would be more content in the viewfinder (above, below, to the left and to the right of what shows up in the viewfinder of the digital camera).

[Olegis]

The magnification of the lens does not change on 1.6 crop body, but the shake is more evident because the frame contains smaller portion of the original picture - that's why you should use the 1/1.6*focal length rule.

[dhbailey]

The lens relative to the sensor doesn't change -- if you have a 200mm lens on a 35mm body you would shoot as close to 1/200th as possible. If you crop that picture down to represent the same angle of view that the 1.6 crop factor of a digital body gives you, you won't see any more shake than on the image as a whole.

But actually, all this is academic -- before anybody gets to a critical shoot where they don't want to blow any pictures, a wise person would test things out for themselves rather than rely on any "rule of thumb." Try the rules-of-thumb out and decide what works for you.

As for what Chazs says about his dad, does that mean that we ALL should be able to shoot at 1/4 and expect steady pictures? I don't think so, but I do envy all the great shots his dad must have gotten that I've had to forego or have ruined by trying to shoot that slowly handheld. So his dad's rule of thumb would be "anything you can shoot at 1/4second, go for it." I don't think that applies to many photographers I know of.

So take your camera and lenses out to a place where you can get good light, clear objects to focus on, and try that 200mm lens at all the shutter speeds you want. You'll find out what YOUR rule of thumb is. You might find that you need to do 1/fl + 1 (one speed setting faster than 1/fl) or even +2 or +3. You might find you can do 1/fl -1 or -2 or -3. Only you will be able to decide appropriately.

Whenever I see those banks of professional photographers at sporting events I sometimes wonder how many "rules of thumb" they are using in their shooting. Probably none at all -- in a professional situation you have to depend on what you KNOW will work, not somebody else's "rule of thumb."

Rules of thumb are great starting points from which to begin to test your own individual abilities.

But it is sad to hear somebody say, about a ruined picture, "but I did what they said I should!"

[Andy_T]

Common sense tells me that the 1.6 crop factor should be taken into account.

Why? Because when your hands shake, the picture moves up and down (let's assume linear shake, for the sake of the agreement).

In the picture you want to print, the movement will be visible if it exceeds a certain amount ... and the further distant your subject is, the bigger the movement due to shake will be.

As is true with the question whether a 50 mm is a good 'portrait' lens on a DSLR .... the focal length does not change with the crop factor ... However, the FOV changes and so maybe the distance to your motive might change. And this will influence both distortion and shake.

Best regards,
Andy

PS: What about a practical example. My G2 actually has a 7 - 21 mm lens (it has a 5x crop factor). I haven't succeeded very well to hand-hold pictures at full tele with 1/20 shutter time.

[chris.bailey]

The 1/FL rule of thumb is less reliable with digital than film as the important thing is the size of the sensor sites. As sensor site size decreases with increased resolution for a fixed sensor size the more sensors are available to resolve the shake so on a 1DS MkII you would need a much higher speed to prevent visible evidence of camera shake than you would with a 10D. To go to extremes if you had a camera with a 100 pixel resolution, though the image would be low in detail, there is little chance that camera shake would change the detail being resolved by any of those 100 sites. There is an article on the Canon website somewhere that goes in to a lot of detail on this but then makes some pretty unbelievable claims about the associated benefits of IS. If I can find it again I will post up the link.

I prefer to not use a tripod in the studio but have found that with the 1dMkII I am having to so to achieve improved results over the 10D. Outside I aim for 1/2 x FL wherever possible and up the ISO if needs be. IS does help, particularly on the 100-400 but I treat this as a last resort rather than relying on it.

[CyberDyneSystems]

The lens relative to the sensor doesn't change -- if you have a 200mm lens on a 35mm body you would shoot as close to 1/200th as possible. If you crop that picture down to represent the same angle of view that the 1.6 crop factor of a digital body gives you, you won't see any more shake than on the image as a whole.


I disagree here,

If I am shooting film with a lens that I find too short for the tak and KNOW I will be cropping.. I also know that the cropped area will thus be enlarged when I print.. further drawing attention to any defects/camera shake than if I printed without enlarging.


In this case I would certainly try to get more that 1/200th for a 200mm lens.

Thus.. as the 1.6 "X-Factor" is all about cropping and then enlarging at the print end of the post processs.. the same idea would apply.

[Steven M. Anthony]

yes--the factor applies to printing, but not the actual amount of "shake" contained in the image. The 1/f rule wouldn't hold up for film if you were going to make a 8 foot by 12 foot print, either...

[DaveG]

The 1.6 crop factor associated with many digital cameras does NOT change the effective focal length of a lens. It does NOT make a 200mm lens a 320mm lens. If it did, any time you wanted a 1,000mm lens, all you would have to do is put duct tape over most of your sensor and a thrifty-50 would be magically converted into a 1,000 mm prime!

The 1.6 crop factor is not magic. It is simply a crop factor--a crop factor relative to the content that would be contained on a full frame of 35mm film.

Here's an example: Close one eye and look at something across the room. Go ahead, I'll wait... Okay, now roll the fingers of one hand up so your finger tips touch the palm of your hand. Now hold your hand up to your opened eye and look through the little tunnel in your hand at the same object across the room. Did the effective focal length of your eye change? No. Did the object across the room look closer the second time you looked at it? No. All that happened by looking through the tunnel in your hand was that you reduced (i.e., cropped) your field of view.

That's how the 1.6 crop factor works. If you put the same lens on a film camera, everything will look as far away as it did through the digital camera--but on the film camera, there would be more content in the viewfinder (above, below, to the left and to the right of what shows up in the viewfinder of the digital camera).

You have to ask yourself under what conditions does this rule work? Part of it must be the size of reproduction you are planning, and I'd say that an 8x10 print should be the minimal acceptable print size.

If you took an image with a 100 mm lens (on full frame 35mm) then the suggested "lowest" shutterspeed would be 1/125 or thereabouts, and most of the time will make an acceptable 8X10 print. If you cropped this image to give you the equivilent crop as a 1000 mm lens and made an 8x10 print, then you would find that it was very unsharp (excluding any grain implications), and now unacceptable.

If you leave the grain/noise out of the equation then the softness with a lens cropped to give you a 1000 mm "look" will be EXACTLY the same as if you used a 1000 mm lens.

It's not only the lens that's at issue here, it's the end use as well.

If you used a 300 mm lens and only made contact sheet sized prints then you could hand hold it down to 1/15 second mostly because you couldn't see the softness. With digital, and the 1.6 conversion you have to take the converted focal length into account because you are basically enlarging the image while maintaining the same size print.

After all this is said an done just remember that: a) the 1/focal length is a minimal not optimal standard; b) it is part of some guy's manhood to brag about how low they can handhold a camera, c) it's not true, their stuff sucks but they don't see it, & d) there are no contests for how slow you can handhold a camera, only for good images; so get over it.

[Steven M. Anthony]

David G.: See my post above yours... plus, don't forget the artistic intent behind the image. The 1,000mm crop of a handheld shot might be exactly what the artist was going for...!

All kidding aside, the point I was making is that the amount of softness in an image due to camera shake is not magnified by the crop factor--because the crop factor does not change the focal length of the lens.

And the 1/f rule is a rule of thumb. Use it or don't. And go easy on those who are proud of how steady their hands are. We are all proud of some aspects of ourselves. I'm not sure how you can claim that their stuff sucks as a general rule. H.C-B. didn't use a tripod, and I think his stuff is generally considered pretty good...

[CyberDyneSystems]

yes--the factor applies to printing, but not the actual amount of "shake" contained in the image. The 1/f rule wouldn't hold up for film if you were going to make a 8 foot by 12 foot print, either...

Yes.. I agree here 100% :wink:

The only "magnification" that results from a "Crop factor" occurs at the printing end.

If you printed all your 10D images on paper that was 40% (or is it 60% ?) smaller than standard print sizes.. there would be no "magnification" at all relative to "full frame" there would only be the crop/loss of angle of view.


Hehe heheh

Here we go again :lol: :lol:

-=CROP FACTOR=- 10,000 posts on the X-Factor (http://photography-on-the.net/forum/showthread.php?t=45388) <== CLICK ME :)

[Scottes]

Does it really matter if the 1.6 factors in? If *you* can shoot steady then shoot at *your* shutter speed. If someone else is shakey then they should shoot at *their* shutter speed.

You should take some test shots. Whether you start at 1/focal or 1/focal * 1.6 doesn't matter, because the goal is going to be to find what shutter speed YOU can use.

And then learn some techniques on steadying, and practice them. Soon you'll be better.

This has nothing to do with 35mm or 1.6 crop factors. It only depends on the shooter.

[DaveG]

David G.: See my post above yours... plus, don't forget the artistic intent behind the image. The 1,000mm crop of a handheld shot might be exactly what the artist was going for...!

All kidding aside, the point I was making is that the amount of softness in an image due to camera shake is not magnified by the crop factor--because the crop factor does not change the focal length of the lens.

And the 1/f rule is a rule of thumb. Use it or don't. And go easy on those who are proud of how steady their hands are. We are all proud of some aspects of ourselves. I'm not sure how you can claim that their stuff sucks as a general rule. H.C-B. didn't use a tripod, and I think his stuff is generally considered pretty good...

The crop factor DOES magnify softness to the point where it is EXACTLY the same as using a longer lens, so while it does not change the focal length, it has the same effect. The rule really shouldn't be "1/focal length" but "1/field of view". An 80 mm lens will be moderate tele on a 20D, a portrait lens on a 1Ds, a normal lens on a Mamiya 645 and a moderate wide angle on a 4x5. For each format there needs to be a new rule and that is 1/field of view.

Because it's new and its lifetime is expected to be short, there is no sense of "format" with the 1.6 conversion. If it was to last and become standard then we wouldn't say "... the equivilent of a 50 mm lens on 35mm." any more than I convert my Mamiya focal lengths now. If that day comes then a new rule will appear and it'll be "1/focal lenght + one shutterspeed"

As to the "stuff sucks" it's that I hear all of the time, "I've got a 16x20 on the wall that's perfectly sharp and I took it at 1/15 of a second ..." blah blah blah. But rarely do we get to see these materpieces.

The rule is actually flawed in the sense that it sounds absolute. What it should say is that with a shutterspeed of 1/focal length (in 35mm format) then 75% of your images will be sharp. With one stop less of shutterspeed perhaps 40% will be sharp, two stops 10%, and so forth. How many shots did Henri have to throw away because they weren't sharp? How well do his great ones enlarge? Hey in-focus and sharp is or should be the expected start point, but if I had a slightly unsharp shot of a guy with a rifle in a building in Dallas 41 years ago it'd be moot, wouldn't it?

Nolan Ryan had a 100 mph+ fast ball and he could throw it right down the middle of the plate. But I can assure you that he wouldn't teach anyone else to do this, and he wasn't taught that way either. He could just do it. All of us with our mediocre fastballs had better aim for the corners or someone will smack our pitches out of the park.

[Steven M. Anthony]

DaveG: How does a crop factor magnify the softness in an image? All the crop factor is doing is describing the reduced amount of the image circle coming into the camera body (through the lens) that is recorded by light-sensitive medium. The sensor in a 10D, for example, is only 62.5% the size of a full-35mm-frame sensor.

If you have a full-frame sensor camera with a 50mm lens and a 10D with a 50mm lens, and take a photo with each, you'll get the same level of camera shake (within the variability of the shooter). If you crop off 37.5% of the full-frame image (from around the edges), you will end up with THE SAME image you got with the 10D.

The size of the light-sensitive medium behind the lens has NO IMPACT on the optical qualities of the lens itself. Nor does the size of the light-sensitive medium behind the lens have any impact on the amount of camera shake that will be recorded.

Re "stuff sucks," it sounds like you are the one who needs to get over it. It also seems that you have given the "1/f rule of thumb" more credit that it is due. I've never considered it an absolute rule and disregard it all the time when lighting conditions don't cooperate!

[slin100]

I think many of us are actually agreeing about the details but disagreeing about the conclusion. Everyone agrees that cropping doesn't change the true focal length of the lens. It does change the field of view (FOV); that is what is meant by "effective" focal length.

I think we also agree that cropping the lens' image circle does not directly magnify camera shake. But the comparison with 35mm film doesn't end there. The final print is always the basis for comparison, and to meaningfully compare two formats, you have to compare identical prints. To do that, you have to magnify the cropped image by a greater amount than you would for an uncropped 35mm frame. The greater magnfication of the cropped image magnifies the camera shake. In other words, if you take a 35mm film frame shot with a 320mm lens at 1/320 and a 1.6x crop factor camera with a 200mm lens at 1/320 and magnified both images up to 4x6, both images would exhibit the same amount of camera shake.

I think Andy Thaler's reference to P&amp;S digicams makes this point incontrovertible. A digital P&amp;S takes the concept of cropping to an extreme. Crop factors are routinely in the 5x or greater range compared to 35mm film. The true focal lengths of these cameras are in the teens and single digits. You don't see many people using 1/(true FL) here. A 1.6x crop factor is no different. Crop factor should be factored into the rule of thumb.

Some people have shakier hands than others, so 1/(2x FL) on a 35mm may work out better for them. If they want similar results when switching to a 1.6x crop factor camera, then they better adjust the shutter speed to 1/(3.2x FL).

[Steven M. Anthony]

But slin100, to compare identical prints, you would need to crop the full-frame to contain only the field of view found on the smaller sensor. Then, when enlarged to a given size, they would look the same.

[Jon]

But slin100, to compare identical prints, you would need to crop the full-frame to contain only the field of view found on the smaller sensor.

. . . or use a lens with the same angle of coverage at full frame as the other lens has with the smaller sensor.

Go back to basic geometry. Using a 50 mm lens on a 1.6 crop factor DSLR or an 80 mm lens on a full-frame 35 mm, you'll be establishing congruent triangles between any reference points in the subject and the lens' nodal point, and their projections on the film/sensor through the lens' nodal point. The triangles for the two lenses will be fully congruent. That's what the "crop factor" is about. And what it results in is a greater degree of enlargement for any given 1.6x "crop" DSLR sensor image to the comparable 35 mm "full frame" sensor image to achieve a given final print size.

[slin100]

But slin100, to compare identical prints, you would need to crop the full-frame to contain only the field of view found on the smaller sensor. Then, when enlarged to a given size, they would look the same.
That's not how the comparison should be done. People don't compare shots by cropping down a 35mm frame. The point is to leverage the entire coverage available to the format. This implies that you must use different lenses. This is a meaningful comparison:
35mm film camera with 320mm (true focal length) lens
1.6x crop factor camera with 200mm (true focal length) lens
Both cameras shot from the same distance to subject will produce identical framings with no additional cropping. The image from the 1.6x crop factor camera will necessarily need to be magnified more than the 35mm image to arrive at the same final print size.

[BoySpot]

Sharky,

I bet you are really glad you asked now, aren't you! Having read everything above, plus having a view of my own, if the rule worked for you on film, factor in the 1.6 for your 10D and it should still be about right. Leave everyone else to get very upset about the rights and wrongs of it all and be happy with your shots.

[CyberDyneSystems]

Twice in two days time....

The "X-Factor" strikes again...

We are hopeless I tell you! Hopeless!


:lol: :lol: :lol:

[DaveG]

DaveG: How does a crop factor magnify the softness in an image? All the crop factor is doing is describing the reduced amount of the image circle coming into the camera body (through the lens) that is recorded by light-sensitive medium. The sensor in a 10D, for example, is only 62.5% the size of a full-35mm-frame sensor.

If you have a full-frame sensor camera with a 50mm lens and a 10D with a 50mm lens, and take a photo with each, you'll get the same level of camera shake (within the variability of the shooter). If you crop off 37.5% of the full-frame image (from around the edges), you will end up with THE SAME image you got with the 10D.

The size of the light-sensitive medium behind the lens has NO IMPACT on the optical qualities of the lens itself. Nor does the size of the light-sensitive medium behind the lens have any impact on the amount of camera shake that will be recorded.

Re "stuff sucks," it sounds like you are the one who needs to get over it. It also seems that you have given the "1/f rule of thumb" more credit that it is due. I've never considered it an absolute rule and disregard it all the time when lighting conditions don't cooperate!

"How does a crop factor magnify the softness in an image? All the crop factor is doing is describing the reduced amount of the image circle coming into the camera body (through the lens) that is recorded by light-sensitive medium. The sensor in a 10D, for example, is only 62.5% the size of a full-35mm-frame sensor."

What makes a longer lens less sharp for the given shutter speed? I mean why do you use 1/500 of a second with a 500mm lens? It's because you are looking at a very tiny field of view and any camera movement or shake will be visible. If you take a capture with a shorter lens and then crop it into the same field of view, IN THE SAME PRINT SIZE, then those same movements or shaking will look exactly the same.

"If you have a full-frame sensor camera with a 50mm lens and a 10D with a 50mm lens, and take a photo with each, you'll get the same level of camera shake (within the variability of the shooter). If you crop off 37.5% of the full-frame image (from around the edges), you will end up with THE SAME image you got with the 10D"

Yes and if I enlarge the guts out of any image (which is more or less like increasing the focal length of the lens) then the flaws will be visible on the 1DS shot exactly like they would be from the 10D. The image from the 1DS at 1/focal length is only acceptable if you use ALL of the capture. A full frame print at 1/focal length may well be acceptable for full frame but when you crop in you'll see that it isn't.

Try this: Put your longest lens on your camera and take a picture using that lenses' rule. Now use a shorter lens and its rule which should be a slower shutterspeed. Download them into your computer and open a file with the longer lensed shot so that it fills about 90% of your screen. Now open up the shorter capture and enlarge it until it looks just like the first shot. Do you see that the perspective is exactly the same (as in a longer focal length lens) and that the image blurr is much worse? You used a shorter shutterspeed which could freeze your muscle tone with the long lens (which you expect) but the shorter shutterspeed couldn't freeze your muscle tone when you enlarged the second shot to a size where you could see it.

I stand by my comment about handholding. A tripod is a much better solution.

[Steven M. Anthony]

It only makes sense to compare the camera shake of two images if the images are the same.

Let's say you take a photo of a yard stick and on the full frame image it stretches the full frame from left to right. If you stand in the same spot and use a 10D with the same lens, your frame will contain only 22.5 inches of the yard stick (inches 6.75 through 29.25). If you blow up the 10D image to 8"X12" it will look just like the full-frame image when an 8"X12" print from it only includes inches 6.75 - 29.25.

That's the only point I'm trying to make here--i.e., that the crop factor does not magnify camera shake (or anything else).

[slin100]

I really don't understand why people are pooh-poohing this rule of thumb. It's a rule of thumb not a rule, meaning that it's intended to be roughly correct, not scientifically correct.

I happen to think this rule of thumb is a good one. It simply needs to be adjusted to one's personal shakiness factor. If you shake twice as much as the average person, cut your shutter speeds to half what the rule of thumb dictates. If you're twice as stable, then double your shutter speeds. Just don't forget to include the sensor crop factor! :D

[Tom W]

All kidding aside, the point I was making is that the amount of softness in an image due to camera shake is not magnified by the crop factor--because the crop factor does not change the focal length of the lens.

And the 1/f rule is a rule of thumb. Use it or don't. And go easy on those who are proud of how steady their hands are. We are all proud of some aspects of ourselves. I'm not sure how you can claim that their stuff sucks as a general rule. H.C-B. didn't use a tripod, and I think his stuff is generally considered pretty good...

The "crop factor" does magnify camera shake as compared to its full-frame counterpart. Its not the absolute focal length that matters, but rather the field of view that determines how much camera shake will be visible.

The entire purpose of the "rule of thumb" was to allow 35 mm shooters a simple formula to give them good odds of producing a print that doesn't exhibit the effects of camera shake. As you said, it is not a hard-and-fast rule, but a guideline.

Given that the guideline was made with the final print in mind, you have to apply it in such a manner that the same-sized final print can be made with the same "shake-free" results. The only way to do that is to take into account that a 1/1.6X-sized sensor will have to have its end product enlarged by 1.6X more than the equivalent-framed output produced by 35 mm film. And when you enlarge any film or sensor output by 1.6X, you are multiplying every aspect of that image by the same amount, including the blur that is coincident with camera shake.

Since the shake-induced blur is multiplied by 1.6X more on the smaller sensor than it is on 35 mm film, you should compensate by multiplying your shutter speed by 1.6X that which you would use with 35 mm film to get the equivalent result.

[slin100]

It only makes sense to compare the camera shake of two images if the images are the same.

Let's say you take a photo of a yard stick and on the full frame image it stretches the full frame from left to right. If you stand in the same spot and use a 10D with the same lens, your frame will contain only 22.5 inches of the yard stick (inches 6.75 through 29.25). If you blow up the 10D image to 8"X12" it will look just like the full-frame image when an 8"X12" print from it only includes inches 6.75 - 29.25.

That's the only point I'm trying to make here--i.e., that the crop factor does not magnify camera shake (or anything else).
That's not in dispute. I mean this sincerely, but it's just not relevant as far as the rule of thumb is concerned. Standing in the same spot you would have to switch to a wider-angle lens on the 10D to capture the whole yard stick. That's the point of the exercise--to capture the same image, but not necessarily with the same lens. If you accept this, then the conclusion should be apparent.

I'm puzzled why the extreme crop factor example for the digital P&amp;S doesn't make this obvious?

[Steven M. Anthony]

slin100: Tom W and DaveG seem to dispute it! But it's like teaching pigs to sing--it just frustrates you and the pig never learns to sing...

The P&amp;S camera example isn't quite the same because the geometry of the camera is different from that of a 35mm dslr. Without the need for a mirror to swing in front of the lens, you can get the back element of the lens much closer to the sensor. From what I understand, this increases the magnification of a given focal length relative to the same focal length on a slr.

[Tom W]

It only makes sense to compare the camera shake of two images if the images are the same.

Let's say you take a photo of a yard stick and on the full frame image it stretches the full frame from left to right. If you stand in the same spot and use a 10D with the same lens, your frame will contain only 22.5 inches of the yard stick (inches 6.75 through 29.25). If you blow up the 10D image to 8"X12" it will look just like the full-frame image when an 8"X12" print from it only includes inches 6.75 - 29.25.

That's the only point I'm trying to make here--i.e., that the crop factor does not magnify camera shake (or anything else).

Why on earth would you want to take a picture of only part of your desired subject?

[Tom W]

slin100: Tom W and DaveG seem to dispute it! But it's like teaching pigs to sing--it just frustrates you and the pig never learns to sing...

You seem to have missed the entire purpose of the "rule of thumb", and proceeded to insult people in the process.

The P&amp;S camera example isn't quite the same because the geometry of the camera is different from that of a 35mm dslr. Without the need for a mirror to swing in front of the lens, you can get the back element of the lens much closer to the sensor. From what I understand, this increases the magnification of a given focal length relative to the same focal length on a slr.

The whole issue of the "rule of thumb" has absolutely nothing to do with backfocus spacing. You're overcomplicating the issue with irrelevent information.

[Jon]

slin100: Tom W and DaveG seem to dispute it! But it's like teaching pigs to sing--it just frustrates you and the pig never learns to sing...

The P&amp;S camera example isn't quite the same because the geometry of the camera is different from that of a 35mm dslr. Without the need for a mirror to swing in front of the lens, you can get the back element of the lens much closer to the sensor. From what I understand, this increases the magnification of a given focal length relative to the same focal length on a slr.

Then you misunderstand. Any lens of a given focal length has the exact same optical (magnifying/reducing) properties as any other lens of the same focal length. When we're using wide-angle lenses on 35 mm SLRs, the designers have had to use "retrofocus" designs, where the lens' nodal point is behind the last element of the lens. A similar situation comes up in telephoto lenses, where the correct combination of lens elements can have the apparent nodal point well forward of the front of the lens. BUT the lens still has exactly the same magnification as a simple lens of the same focal length. If what you believe were true, a 28 mm lens on a Leica IIIg or an Argus C3 (both 35 mm rangefinders) would have different magnification than a 28 mm lens on an SLR. This is demonstrably not so. Any lens of a given focal length will have exactly the same magnifying, or reducing (in the case of a negative diopter lens) power. It will have different angles/fields of view depending on the lens design and the sensor size, but the focal length is independent of angle/field of view.

[slin100]

slin100: Tom W and DaveG seem to dispute it! But it's like teaching pigs to sing--it just frustrates you and the pig never learns to sing...

The way I interpret Tom W and DaveG words, they're actually agreeing. They're backing your statement but confirming mine. I've been trying to point out that we're making two valid but different statements.

The P&amp;S camera example isn't quite the same because the geometry of the camera is different from that of a 35mm dslr. Without the need for a mirror to swing in front of the lens, you can get the back element of the lens much closer to the sensor. From what I understand, this increases the magnification of a given focal length relative to the same focal length on a slr.
Ok, this is probably beyond my ability to argue scientifically, but I don't see how this can be true. While the omission of a reflex mirror enables much shorter lens register distances (distance from lens mount to sensor plane) on a P&amp;S, the only thing affected is the design of the lens geometry. If there was some extra magnification happening here on a P&amp;S that was not present on an SLR, then I would argue that the resulting field-of-view could not be the same. After all, what is magnification but simply a change of field-of-view? Furthermore, if lens registers were a factor, then the rule of thumb would not even work in the 35mm world because different manufacturers use different lens register distances for their respective mounts. I have never heard anyone claim lens register as a factor.

[CyberDyneSystems]

At this point I'm merely playing sideline critic laughing all the way...

But;




I'm puzzled why the extreme crop factor example for the digital P&amp;S doesn't make this obvious?


...He's got you there!

My old Olympus C2100 UZ had a 7-70mm lens... thats all.. 7-70mm :?

But... it "cropped" like a 38-380mm... Yes.. nearly 400mm and it would bring in birds like a 380mm lens too..

...and it had Image Stabilazation!

I.S. for 70mm?

I don't think so.. the IS was for the 380mm "equivelent" and it was needed too.. as would be a faster shutter speed without the IS.


But don't let any of this logical thinking disturb the everlasting "X-Factor Sideshow"

:lol: :lol: :lol: :lol: :lol:

[robertwgross]

The way I interpret Tom W and DaveG words, they're actually agreeing.

We should send you to the Middle East and have you negotiate between the Israelis and the Palestinians.

---Bob Gross---

[sharky]

Living in Australia, I'm about 10 hours out of sync with a lot of the posters here, so it came as a bit of a surprise to log in this morning and see what I had started. Thanks to all the posters - I think the thing I'll take away from this is that I'm going to have to go out and shoot a lot more to figure out what X factor works for me. And start saving to buy another 28-135IS, which I was dumb to sell in the first place.

[Steven M. Anthony]

Check out this page

http://www.panoramafactory.com/equiv35/equiv35.html

Pay particular attention to the part that compares a 24mm lens to a 72mm lens (both with 35mm film).

Now, consider the 24mm lens diagram and assume the film plane is a full-framed sensor plane. The distance between the sensor plane and the focal point of the lens is 24mm. The horizontal Field Of View is 74 degrees.

Now, consider making the 36mm long sensor plane 22.5mm long (but centered in the same place as the full sensor). The focal length of the LENS does not change--it remains 24mm. The HFOV of the LENS does not change--it remains 74 degrees. The size of the subject on the SENSOR does NOT change. The proportion of the frame taken up by the subject DOES change--because there is a portion of the image circle that is no longer falling on the sensor (because the sensor got smaller--and, in effect, CROPS the image that would be recorded by the full-frame sensor).

The 24mm lens is not MAGNIFYING any differently between the full-frame and the smaller sensor (as indicated by the fact that the size of the subject on the smaller sensor would be the same as it would be on the larger sensor). Note the change in MAGNIFICATION when moving from the 24mm to the 72mm lens while maintaining the same sensor plane size. Here there is MAGNIFICATION because the narrower FOV is being spread over the same sized area--which is why you need a longer shutter speed to correctly expose as you zoom to longer focal lengths (you are spreading a smaller amount of light over the sensor as you zoom to long focal lengths).

I haven't meant to insult anyone. I've only wanted to correct those who believe having a smaller sensor MAGNIFIES anything for a given focal length.

Jon: Thanks for the clarification on P&amp;S and focal length in general. I see the error of my thinking on that and it motivated me to do the search where I found the above link.

[Tom W]

Check out this page

http://www.panoramafactory.com/equiv35/equiv35.html

Pay particular attention to the part that compares a 24mm lens to a 72mm lens (both with 35mm film).

Now, consider the 24mm lens diagram and assume the film plane is a full-framed sensor plane. The distance between the sensor plane and the focal point of the lens is 24mm. The horizontal Field Of View is 74 degrees.

I don't recall anybody arguing that the actual focal length of the lens changes based on sensor size. The field of view of the image you collect, however, changes in relation to the size of the sensor.

Now, consider making the 36mm long sensor plane 22.5mm long (but centered in the same place as the full sensor). The focal length of the LENS does not change--it remains 24mm. The HFOV of the LENS does not change--it remains 74 degrees. The size of the subject on the SENSOR does NOT change. The proportion of the frame taken up by the subject DOES change--because there is a portion of the image circle that is no longer falling on the sensor (because the sensor got smaller--and, in effect, CROPS the image that would be recorded by the full-frame sensor).

This is all true so far. But its only part of the story.

The 24mm lens is not MAGNIFYING any differently between the full-frame and the smaller sensor (as indicated by the fact that the size of the subject on the smaller sensor would be the same as it would be on the larger sensor). Note the change in MAGNIFICATION when moving from the 24mm to the 72mm lens while maintaining the same sensor plane size. Here there is MAGNIFICATION because the narrower FOV is being spread over the same sized area--which is why you need a longer shutter speed to correctly expose as you zoom to longer focal lengths (you are spreading a smaller amount of light over the sensor as you zoom to long focal lengths).

I agree - The lens isn't changing anything, and I didn't mean to imply that it was. What is changing is what the user does with both the camera and the resultant image afterwards (I tried to explain it in a previous post). The magnification takes place later when an equally-framed image is enlarged to a similar size.

Put yourself in this situation:

You're taking a head-shot portrait of a beautiful model. You have 2 cameras - a 1Ds and a 20D. You put your 135 mm lens on the 1Ds and frame the image so that you have a perfectly framed head of your subject in the viewfinder. You have an inexpensive hot studio lighting set up and you set your aperture to f/2.0, and your shutter to 125 second. You snap the image, knowing that with your steady hand, you can get a good shot with this shutter speed.

Now, you set down your 1Ds and pick up your 20D. You put your 135 mm lens on the 20D and look through the viewfinder. Obviously, the subject's head is too large for the frame now, since you're using a smaller portion of the image circle. So, you have two choices - you either move back, or you change lenses.

This presents a problem - you can't move back any farther to reframe the model as the room isn't large enough, so you have to change lenses. As luck would have it, you are the proud owner of a Canon 85 mm lens which, on the 20D, provides roughly the same field of view as the 135 does on the 1Ds. So, you put the 85 on your 20D.

Now, in this particular instance, your copy of the 85 mm is a little soft until its stopped down to about f/2.5. So, you set your aperture to f/2.5, 2/3 stop slower than you needed with the 135. The result is that you now need a 1/80 second exposure time to give the same EV. Knowing that you have an 85 mm lens, you feel comfortable with a 1/80 shutter speed. So, you snap the picture.

The customer requests 18X12 inch (~450mm X 300mm) prints to analyze the quality of the images. So you take both of your images, tweak them equally in PS, and then take them to the printer. He makes your prints, but you find that the 10D image has a little bit of motion blur - it appears that you weren't holding the camera quite as steadily with the 20D shot as you were with the 1Ds. (Yes, you should have used a tripod, but that would negate the entire point).

So, what happened?

Well, the 85 mm lens was still an 85 mm lens. And the relationship between the lens and the sensor did not change - except that the 20D sensor only covered a portion of the image approximately 1/1.6X that of the 1Ds sensor. Essentially, nothing has changed within the camera that would multiply or amplify the shake of the user (we agree on that).

But, in the process of developing and producing the final image, the 1Ds's sensor output must be enlarged from an image of 36X24 mm to 450X300 mm, while the image from the 20D must be enlarged from an original sensor image of roughly 22.5X15 mm to the same final 450X300.

The image from the 1Ds must be enlarged to 12.5 times its original size.
The image from the 20D must be enlarged to 20 times its original size.
Coincidently, the ratio of 20/12.5 is 1.6.

This is where the shake issue comes into play. The original "rule of thumb" was written with the idea that the original negative would be enlarged a certain amount to create a given size print. But the 20D negative is smaller, and therefore must be enlarged to a greater extent to acheive the same image size. That greater amount of enlargement is 1.6 times the enlargement needed from a full-frame sensor to create the same sized print.

That is where the shake gets multiplied along with every other aspect of the image - in the process of enlarging from sensor size to final print size. The more you enlarge the original, the more detail and imperfections will show.

So, utilize your 1.6X multiplier when using the "rule of thumb", not because of any changes in the camera, but because of the necessity to enlarge the smaller image to a greater extent to achieve the same final product.


I haven't meant to insult anyone. I've only wanted to correct those who believe having a smaller sensor MAGNIFIES anything for a given focal length.

Not to worry - we're looking at two different parts of the process here, I think. A little frustrating at times. :)

[Steven M. Anthony]

"I agree - The lens isn't changing anything, and I didn't mean to imply that it was. What is changing is what the user does with both the camera and the resultant image afterwards (I tried to explain it in a previous post). The magnification takes place later when an equally-framed image is enlarged to a similar size."

Okay, Tom. But comments like "The "crop factor" does magnify camera shake as compared to its full-frame counterpart. Its not the absolute focal length that matters, but rather the field of view that determines how much camera shake will be visible” (by you, and “The crop factor DOES magnify softness to the point where it is EXACTLY the same as using a longer lens, so while it does not change the focal length, it has the same effect,” by DaveG suggested otherwise. But keep in mind, the FOV of the LENS still doesn't change--just the FOV of capture does.

[slin100]

But comments like "The "crop factor" does magnify camera shake as compared to its full-frame counterpart. Its not the absolute focal length that matters, but rather the field of view that determines how much camera shake will be visible” (by you), and “The crop factor DOES magnify softness to the point where it is EXACTLY the same as using a longer lens, so while it does not change the focal length, it has the same effect,” by DaveG suggested otherwise. But keep in mind, the FOV of the LENS still doesn't change--just the FOV of capture does.
I can see why you thought Tom's and DaveG's comments suggested otherwise, but they don't if you consider their comments in the context of the various stages of magnification that occur from image capture to final output.

[Tom W]

"I agree - The lens isn't changing anything, and I didn't mean to imply that it was. What is changing is what the user does with both the camera and the resultant image afterwards (I tried to explain it in a previous post). The magnification takes place later when an equally-framed image is enlarged to a similar size."

Okay, Tom. But comments like "The "crop factor" does magnify camera shake as compared to its full-frame counterpart. Its not the absolute focal length that matters, but rather the field of view that determines how much camera shake will be visible” (by you, and “The crop factor DOES magnify softness to the point where it is EXACTLY the same as using a longer lens, so while it does not change the focal length, it has the same effect,” by DaveG suggested otherwise. But keep in mind, the FOV of the LENS still doesn't change--just the FOV of capture does.

A lot depends on how you would define "crop factor". Or, for that matter, if "crop factor" is even the best term to use. A lot of symmantics involved here, apparently.

[nosquare2003]

Try the rules-of-thumb out and decide what works for you.

It's the best rule of thumb!

[pcasciola]

There is definitely a lot of semantics regarding the crop factor, but it all comes down to geometry and resolution. If the resolution of the smaller sensor is the same as the full frame sensor (i.e. both 8MP), then you are in effect making it a longer lens on the 1.6x camera and would have to adjust for that with your 1/focal_length rule.

Here's an example. If I take a picture with a 200mm lens on my 20D, I am getting 8 megapixel resolution within the same angle of view as an 8MP full frame camera would using a 320mm lens. All things being equal, and assuming the chips have the same properties other than the sensor size, and assuming in neither case are we exceeding the resolution of our two lens, we are getting the same resolution over the same area pixel for pixel. So, why is a 200mm on a 20D not the same as a 320mm lens on an 8MP full frame camera if there is no physical way to tell the two apart, even under a microscope?

In this scenario, the 1/focal_length calculation would have to be raised to allow for the crop factor, because if both cameras in the above example shake the same amount, the blur is the same, so the guy with the 20D + 200mm lens would have to use the same shutter speed as he would with the 320mm on the full frame. I don't think it has to be raised by exactly the 1.6x factor, but I don't have the time right now to prove out all the geometry, maybe tomorrow if I get some time.

Interesting points though, and I think dhbailey has it right. Everyone has a different amount of shake, and I've found the 1/focal_length definitely does not work for me. A sniper would probably be able to use a 1000mm lens and shoot 1/60th at f/16 without a problem, but I need about 1/1000th to be able to shoot wide open with my 50mm in broad daylight :lol:

[Andy_T]

I have to disagree with CDS ...

... the 'X' factor should obvisouly be called 'Y' factor ...

... as in 'WHY :roll: :roll: :roll: '

actually, I'm really astonished as to the amount of work you guys put into this reasoning.... :wink:

Best regards,
Andy

[Tom W]

I have to disagree with CDS ...

... the 'X' factor should obvisouly be called 'Y' factor ...

... as in 'WHY :roll: :roll: :roll: '

actually, I'm really astonished as to the amount of work you guys put into this reasoning.... :wink:

Best regards,
Andy

In the spirit of boolean logic, I present:

Y-NOT

:?

[CyberDyneSystems]

You'd have to dig back even further to find this bit of CDPF history,.... but one of the major "X" Factor blow-outs we had ... over a year ago anyways..


..well one of the big "issues" became "what are you ALLOWED to call it"

You see even magazines and such used to call it "magnification factor"... then "Crop factor" became "PC" .

But there was still a big disagreement about this.

I think it was Bob Gross that suggested "lens Factor" as a nuetral alternative between the "crop" group and the "magnification" party.

.. but then he changed his mind and got mad at me for using that... :roll: (I think I'm making this up in all truth)

Anyway... this is whn I started calling it "X" factor.. so as not to upset any of the warring factions either way.. it was the "PC" thing to do.... :roll: :roll: :wink:


:lol:

[PacAce]

It only makes sense to compare the camera shake of two images if the images are the same.

Let's say you take a photo of a yard stick and on the full frame image it stretches the full frame from left to right. If you stand in the same spot and use a 10D with the same lens, your frame will contain only 22.5 inches of the yard stick (inches 6.75 through 29.25). If you blow up the 10D image to 8"X12" it will look just like the full-frame image when an 8"X12" print from it only includes inches 6.75 - 29.25.

That's the only point I'm trying to make here--i.e., that the crop factor does not magnify camera shake (or anything else).

If you take a look at an inch in the 10D 8x12 picture you'll notice that it's not the same length as an inch in the 35mm 8x12 picture. In other words, the inch got magnified in the 10D picture as will any shakes and motions captured in the picture.

But you are right about one thing. It's like trying to teach a pig to sing. But the question is, who IS the pig? :mrgreen:

[Steven M. Anthony]

Leo: If the full-frame and "partial-frame" sensors are of equal pixel density (i.e., NOT the same number of pixels accross the sensor--but the same density of pixels) the size of each inch would be the same. This is because at equal density--the only way it makes sense to compare full-frame to partial frame--the section of the full-frame shot that is cropped to show only inches 6.75 to 22.5 of the yard stick would have the same number of pixels as the partial-frame sensor. THAT would be an apples to apples comparison--and under these conditions, the camera shake would be the same.

I think that answers your question... :lol:

[CyberDyneSystems]

Ahh.. you've slipped up now... :wink:

Leo said "on the print".. not on the sensor. The measurment would be different in size on the print.. so pixel size on the sensor is not important.

If you print a 10D image on an 8X10 it will blow up the photo of the ruler as compared to an 8X10 print from full frame with the same lens and distance from subject.

This is what "crop factor" is all about. Magnification when printing.

[Scottes]

Magnification when printing.

Look Out CDS! You just went and actually joined this discussion. Now you're in for it!

[Steven M. Anthony]

no--the measurement on the print would be the same, too--6MP blown up to 8 X 12 in each case (remember, we are only using the center 6MP of the full frame--i.e., we are CROPPING the full frame by a... oh yeah, a CROP FACTOR!) :)

[Tom W]

NOOOOooooooo.........

:shock: :) :D :D

[robertwgross]

...
I think it was Bob Gross that suggested "lens Factor" as a nuetral alternative between the "crop" group and the "magnification" party.


I thought the easiest term was "the 1.6 factor."

If somebody wanted to think of that as a crop factor, then fine. If somebody else wanted to think of it as a magnification factor, then fine. If somebody wanted to think of it as a field of view factor, then fine.

I just wonder why we waste so much bandwidth about it.

---Bob Gross---

[kfong]

Since this thread won't die, I might as well put my 2 cents worth.
Let's look at this argument mathematically.

Using the simple lens formula, M1, the magnification at the sensor is
M1= i/o = f/(o-f) where
o is the object distance,
i is the image distance,
and f is the focal length.

Of course nobody is going to look at the image unmagnified at the sensor size. It will always be magnified from the sensor size (s) to a suitable viewing size (p).
M2=p/s.

The overall magnification from object size to viewing size is
M=M1*M2=pf/(s*(o-f))

Now if the full frame 35 mm sensor has size sf=S, and the 20D, 10D, D60, D30 has size ss=S/x, where x= 1.6.

The overall magnifcation M for full frame sensor is
Mf=pf/(S*(o-f)) ~= pf/(So) for o>>f

and for the smaller sensor is
Ms=pfx/(S*(o-f)) ~= pfx/(So) for o >>f

Presto, the smaller sensor has an "effective" focal length of fx!
QED.

AND this applies to the effect of camera shake on the viewing image as well as to FOV and depth of field.

However, this does NOT apply to effects that do not include M2, e.g. exposure calculation, image stabilization operation.

Ken

[pcasciola]

I have a crop factor question.

If I use a 300mm lens w/1.6x EF Extender (let's pretend there is one) on an 8MP full frame camera, or a 300mm lens on a 1.6x "crop factor" 8MP camera, what's the difference, other than with the extender you are losing a full stop.

If there is no practical difference, should people be saying they are cropping by 1.4x when they put their 1.4x TC on?

[CyberDyneSystems]

1.4X does increase the magnification of the image being projected on the film plane...

So it's differnet..

AAAAARGHHH

I'm hooked!

This was a trick wasn't it!!!!!!!

:lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol:


//I can see it now.. the perfect crime and they can't make me crack under the pressure no matter what... until...

"So.. Mr. CDS.. if the killer did not use a set of extra long bungie cords.. how did you kill him?"

"... it was simple really.. DOH!" :shock:

[kfong]

I have a crop factor question.

If I use a 300mm lens w/1.6x EF Extender (let's pretend there is one) on an 8MP full frame camera,

You have increased M1 by 1.6x.

or a 300mm lens on a 1.6x "crop factor" 8MP camera,


You have increase M2 by ~ 1.6x

what's the difference, other than with the extender you are losing a full stop. If there is no practical difference, should people be saying they are cropping by 1.4x when they put their 1.4x TC on?

The difference is that in the first case, the same number of photons are spread over an area larger by a factor of 1.6*1.6=2.56, thus you are losing more than a full stop in light intensity.
The change in optics also mean that the convergence/divergent angle into the film plane/focus detector are reduced, and if that is reduced beyond the design limit, the focuser will not receive any light no matter how bright the object is.
A host of other things are also affected adversely, e.g. diffraction, aberrations etc.

So given the same megapixels, a "crop factored" camera is good when you want large magnification, but it sure sucks in wide angle photography.

Ken

[Andy_T]

If I use a 300mm lens w/1.6x EF Extender (let's pretend there is one) on an 8MP full frame camera, or a 300mm lens on a 1.6x "crop factor" 8MP camera, what's the difference, other than with the extender you are losing a full stop.

- the full frame camera has larger photosites that might capture more light ... this is good

- the full frame version uses the corners of the lens, where the resolution normally is not as good as in the center ... this is bad

Why not put the Pro1 sensor into the DRebel?

Then *theoretically* you can either use the Pro1Rebel with 1.8/50 ($700) or the 1DMKII with 1.8/200 ($ 7000) for your wildlife shots :lol:
(OK, the math is a bit off, but you get the picture)

The Pro1Rebel would *really* suck at wide angle, though....

Best regards,
Andy

[KennyG]

This thread is getting confusing, so I will confuse it even more.

Pixel volume/density is a very important factor, never mind FOV and its effect on camera shake/movement. To quote Canon- "Smaller pixels are more sensitive to camera shake, as a smaller movement will cause the image to move across more pixels."

Some people think they have softer images in the newer higher megapixel cameras, where it could be as a result of camera shake. Does this mean you should shoot at higher speeds as pixel volume/density increases? Probably, but not always.

The 'rule of thumb' is just that. It comes from the film days and I swore by it for over 20 years. Now we have cameras that can out-resolve film and lenses (plus some cameras) with IS, you can't apply a 'fit all' rule of thumb. It comes down to what works for you with the equipment you have and the end result you are trying to achieve.

Sorry CDS, I have thrown something else into the melting pot. :twisted:

[Andy_T]

Does it really matter if the 1.6 factors in? If *you* can shoot steady then shoot at *your* shutter speed. If someone else is shakey then they should shoot at *their* shutter speed.

You should take some test shots. Whether you start at 1/focal or 1/focal * 1.6 doesn't matter, because the goal is going to be to find what shutter speed YOU can use.

And then learn some techniques on steadying, and practice them. Soon you'll be better.

This has nothing to do with 35mm or 1.6 crop factors. It only depends on the shooter.

*bold print in quote is my emphasis, not Scottes'*

I think that this was the best advice in the thread. :wink:

What a pity it is not founded by some sort of mathematical equation :roll:

Best regards,
Andy

[Scottes]

What a pity it is not founded by some sort of mathematical equation :roll:

It is, but only in the 1.6th dimension. :wink:

[pcasciola]

1.4X does increase the magnification of the image being projected on the film plane...

So it's differnet..

AAAAARGHHH

I'm hooked!

This was a trick wasn't it!!!!!!!

:lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol:


I knew we'd get ya. :wink:

Seriously though, I ran this by an engineer at my office, and he said it's very similar (crop factor vs. lens extender) given the total pixel count of the sensors are the same. They are both working with the same amount of light through the original lens. He did say the light density goes down slightly compared to the extender, like kfong shows in his calculation, but it's probably not that much different because of the different pixel densities of the sensors. Ok, so maybe there is a little more than one stop's worth of light lost on the smaller sensor, but it's still similar enough that I don't think it should be called cropping, because that implies a complete loss of resolution.

I'm not saying my any means I'd rather have this over a full frame camera, but I'd vote for calling the x-factor a 1.6x extender or magnification rather than crop.

Hmmm.. We're arguing semantics again, aren't we?

Jut my 1-1/2 cents.

[PacAce]

Seriously though, I ran this by an engineer at my office, and he said it's very similar (crop factor vs. lens extender) given the total pixel count of the sensors are the same. They are both working with the same amount of light through the original lens. He did say the light density goes down slightly compared to the extender, like kfong shows in his calculation, but it's probably not that much different because of the different pixel densities of the sensors. Ok, so maybe there is a little more than one stop's worth of light lost on the smaller sensor, but it's still similar enough that I don't think it should be called cropping, because that implies a complete loss of resolution.

I'm not saying my any means I'd rather have this over a full frame camera, but I'd vote for calling the x-factor a 1.6x extender or magnification rather than crop.

Hmmm.. We're arguing semantics again, aren't we?

Jut my 1-1/2 cents.

For the most parts, your engineer friend is correct in his assessment. However, there is a difference in the light reaching the sensors, as you pointed out, but not the way you said but the opposite. The full frame with the 1.6X TC would receive a little over 1 stop LESS light than the 1.6x camera. The light reaching the 1.6x camera would be exactly the same amount as that reaching the full frame camera without the 1.6x TC.

[Jon]

It only makes sense to compare the camera shake of two images if the images are the same.

Let's say you take a photo of a yard stick and on the full frame image it stretches the full frame from left to right. If you stand in the same spot and use a 10D with the same lens, your frame will contain only 22.5 inches of the yard stick (inches 6.75 through 29.25). If you blow up the 10D image to 8"X12" it will look just like the full-frame image when an 8"X12" print from it only includes inches 6.75 - 29.25.

That's the only point I'm trying to make here--i.e., that the crop factor does not magnify camera shake (or anything else).

But that's not the way you typically use your lenses. I'm not going to accept a smaller field of view than the photo demands just because "that's the lens I'd use in 35 mm" which is what your "test" does. In consequence, all your test proves is that the image shake captured at the sensor is constant for a given focal length lens. You're arguing, to put it another way, that my 8.5x enlargement (8" x 12" in 35 mm) full frame print and my 5" x 7" (same 8.5x) in my 20D are equally viewable. But are you going to mount and frame that 5 x 7 on the wall where you would have put the 8x12 full-frame print? I don't think so. You're going to enlarge it to 8 x 12, a 13.5 x enlargement.

What I'm going to do in the real world is use a (shorter f.l.) lens that gives me the same field of view as the 35 mm lens would. The "crop factor" comes into play when you use a 50 mm lens on your 1.6 crop factor DSLR to take the same picture, in coverage that you'd use an 80 mm lens for in full frame 35 mm. That's what crop factor's referring to - what lens you'd need to be the equivalent of a 35 mm full-frame. And that's why the "rule of thumb" needs a 1.6x applied to it. Because to get the same size print out of the "negative", you'd need to enlarge it more, so you're enlarging/magnifying the shake more as well.

[Jon]

Seriously though, I ran this by an engineer at my office, and he said it's very similar (crop factor vs. lens extender) given the total pixel count of the sensors are the same. They are both working with the same amount of light through the original lens. He did say the light density goes down slightly compared to the extender, like kfong shows in his calculation, but it's probably not that much different because of the different pixel densities of the sensors. Ok, so maybe there is a little more than one stop's worth of light lost on the smaller sensor, but it's still similar enough that I don't think it should be called cropping, because that implies a complete loss of resolution.

I'm not saying my any means I'd rather have this over a full frame camera, but I'd vote for calling the x-factor a 1.6x extender or magnification rather than crop.

Hmmm.. We're arguing semantics again, aren't we?

Jut my 1-1/2 cents.

What your engineer friend is saying is correct as long as you're considering a point light source. Unfortunately, unless we're doing astrophotography, we aren't dealing with point sources. If you apply his reasoning, the entry pupil's absolute size is the controlling factor and as long as that's unchanged, the illumination remains constant. It thus follows naturally that my 100 mm f/4, with a 25 mm entry pupil, when used wide open should be able to properly expose a photo at the exact same shutter speed as my 50 mm f/2 (with a 25 mm entry pupil). Gee, ya think a 300 f/12 will work as well, too?

A teleconverter actually changes the optics we're working with. The "crop factor" just changes how much of the image circle we capture; the "crop factor" value is determined [i]solely by the differences in sensor sizes. That's why we refer to crop factors in terms of the camera, not the lens. We then apply the crop factor to a lens to understand how the it will be affected by use on the differing cameras, and have a basis for comparison to the known (35 mm) standard photographers have been dealing with since Oskar Barnack built the first Leica and certainly since Life photographers started using the format.

[Rayz]

It seems everyone has forgotten the original question. The question wasn't, "Is the 1/FL rule of thumb applicable to me because I'm suffering from Parkinson's disease, or applicable to me because I'm a crack shot with a gun and have super steady hands?"

The question was, "Does the rule change with the smaller format (1.6 crop) camera?" And the answer is, yes it does change because the rule is only meaningful in relation to a certain size print, and that size is probably an 8x enlargement of full frame 35mm which is equivalent to a 12.8x enlargement of an APS-C format. The greater the enlargement, the faster the required shutter speed for a given size print exhibiting the same degree of camera shake. There's always some camera shake in a hand-held shot.

[Jon]

It seems everyone has forgotten the original question. The question wasn't, "Is the 1/FL rule of thumb applicable to me because I'm suffering from Parkinson's disease, or applicable to me because I'm a crack shot with a gun and have super steady hands?"

The question was, "Does the rule change with the smaller format (1.6 crop) camera?" And the answer is, yes it does change because the rule is only meaningful in relation to a certain size print, and that size is probably an 8x enlargement of full frame 35mm which is equivalent to a 12.8x enlargement of an APS-C format. The greater the enlargement, the faster the required shutter speed for a given size print exhibiting the same degree of camera shake. There's always some camera shake in a hand-held shot.

We haven't forgotten it - many of us have said exactly that. But some people are contending the contrary, due to what most of us perceive as a misunderstanding of the "crop factor". So we continue in a, perhaps vain, effort to impart our superior understanding ;{)#

[Andy_T]

It seems everyone has forgotten the original question. The question wasn't, "Is the 1/FL rule of thumb applicable to me because I'm suffering from Parkinson's disease, or applicable to me because I'm a crack shot with a gun and have super steady hands?"

The question was, "Does the rule change with the smaller format (1.6 crop) camera?" And the answer is, yes it does change because the rule is only meaningful in relation to a certain size print, and that size is probably an 8x enlargement of full frame 35mm which is equivalent to a 12.8x enlargement of an APS-C format. The greater the enlargement, the faster the required shutter speed for a given size print exhibiting the same degree of camera shake. There's always some camera shake in a hand-held shot.

:lol: :lol: :lol:

So what if the original poster had included the remark 'BTW, I'm suffering from Parkinson's desease' in his post :lol:

Would you then advise ... ' Sure you use 1.6x the focal length, because that's mathematically correct :wink: ' or would you rather say
'It's a rule of thumb ... you might try 1.6x, but maybe 3x would be better for you ... just try it out!!!'

Best regards,
Andy

[Steven M. Anthony]

Since this thread won't die, I might as well put my 2 cents worth.
Let's look at this argument mathematically.

Using the simple lens formula, M1, the magnification at the sensor is
M1= i/o = f/(o-f) where
o is the object distance,
i is the image distance,
and f is the focal length.

Of course nobody is going to look at the image unmagnified at the sensor size. It will always be magnified from the sensor size (s) to a suitable viewing size (p).
M2=p/s.

The overall magnification from object size to viewing size is
M=M1*M2=pf/(s*(o-f))

Now if the full frame 35 mm sensor has size sf=S, and the 20D, 10D, D60, D30 has size ss=S/x, where x= 1.6.

The overall magnifcation M for full frame sensor is
Mf=pf/(S*(o-f)) ~= pf/(So) for o>>f

and for the smaller sensor is
Ms=pfx/(S*(o-f)) ~= pfx/(So) for o >>f

Presto, the smaller sensor has an "effective" focal length of fx!
QED.

AND this applies to the effect of camera shake on the viewing image as well as to FOV and depth of field.

However, this does NOT apply to effects that do not include M2, e.g. exposure calculation, image stabilization operation.

Ken

Nice proof, Ken. But it makes one faulty assumption. The equations above assume the sensor is circumscribed within the image circle created by the lens. But in the case of the 10D (and others with aps-sized sensors in 35mm film-sized set-ups) the sensor corners do not touch the limit of the image circle. Instead, portions of the image circle that would normally fall on a full frame simply fall on the back of the camera body. That's why "crop factor" is such an appropriate name for the "factor" when in reference to full-frame. It's as if one cropped the center 62.5% of the full-frame image and threw the rest away.

But your math gave me a sense of why P&amp;S camera can utilize such small focal length lenses and still obtain the magnification factors (not just crop factors) they do.

Magnification is related to the field of view captured by the LENS (not the sensor) when projected on the sensor. The more FOV you cram onto the sensor (of a given size), the lower the magnification has to be (I'm probably saying that backwards, but I think you can see what I mean). Conversely, the smaller the FOV projected onto the entirety of the same size sensor, the greater the magnification.

A small-ish focal length lens, of say 48.5mm (when it's image circle is trained on the entirety of the tiny sensor in a sony 707) will retain the field of view of a 240mm lens that needs to fill a 35mm frame-sized sensor. This not only makes sense from a physics/optics perspective, but it is congruent with the tests I did between my neighbor's D300 and my Sony 707 (with its tiny sensor). What I found amazing was the difference in QUALITY of the images from camera to camera--to the point where I gave my sony to my 14 year-old and bought a 10D the next day!

I guess you can refer to it as just a "1.6 factor," but why not figure out what it really is so you know what you are really dealing with?

[Rayz]

So what if the original poster had included the remark 'BTW, I'm suffering from Parkinson's desease' in his post :lol:



Then all bets are off. There would be no point in asking the question. The 1/fl rule would be entirely irrelevant.

[kfong]

Nice proof, Ken. But it makes one faulty assumption. The equations above assume the sensor is circumscribed within the image circle created by the lens. But in the case of the 10D (and others with aps-sized sensors in 35mm film-sized set-ups) the sensor corners do not touch the limit of the image circle. Instead, portions of the image circle that would normally fall on a full frame simply fall on the back of the camera body. That's why "crop factor" is such an appropriate name for the "factor" when in reference to full-frame. It's as if one cropped the center 62.5% of the full-frame image and threw the rest away.


I have not made this particular assumption.
The size of the object and/or image and thus whether cropping occurs do not enter into the equations. Only the overall magnification, which determines the amount of blur due to camera shake, is in question.
The only assumption that I've made is that the simple lens formula applies, and in the last 2 equations o>>f.

Ken

[Tom W]

After reading and participating and even argueing in this thread, I've come to the conclusion that the tripod, mirror lockup, and remote shutter release provide the only easy solution. ;)

[PacAce]

After reading and participating and even argueing in this thread, I've come to the conclusion that the tripod, mirror lockup, and remote shutter release provide the only easy solution. ;)

Good! We have a solution! :D :D :D

:? Now, what was the problem again? :oops:

:mrgreen: :mrgreen: :mrgreen:

[Scottes]

Good! We have a solution! :D :D :D

:? Now, what was the problem again? :oops:

I think it was something about the tripod length causing mirror shake when using a remote shutter release.

So Tom was wrong, it seems, since he doesn't take tripod cropping into account, let alone the length of the shutter release cord.

[Steven M. Anthony]

Nice proof, Ken. But it makes one faulty assumption. The equations above assume the sensor is circumscribed within the image circle created by the lens. But in the case of the 10D (and others with aps-sized sensors in 35mm film-sized set-ups) the sensor corners do not touch the limit of the image circle. Instead, portions of the image circle that would normally fall on a full frame simply fall on the back of the camera body. That's why "crop factor" is such an appropriate name for the "factor" when in reference to full-frame. It's as if one cropped the center 62.5% of the full-frame image and threw the rest away.


I have not made this particular assumption.
The size of the object and/or image and thus whether cropping occurs do not enter into the equations. Only the overall magnification, which determines the amount of blur due to camera shake, is in question.
The only assumption that I've made is that the simple lens formula applies, and in the last 2 equations o>>f.

Ken

No--you didn't make the assumption explicitly. But it is implicit in the math. The clue to the assumption being there is the fact that there is no term in your equations for image size.

One way to think about it is to consider a window in your house. Consider the window to be the sensor in a camera. when you sit in a particular spot, you get a particular FOV captured by the window--think of it as an image coming through a lens and falling on a sensor. Now if you soap up the outer edges of the window (you've just made your window a smaller sensor) and sit back in that same spot, what you see through the window has a narrower FOV but has not magnified the scene 1 bit. You now have the worst of 2 worlds--narrower FOV and NO Magnification Factor.

[Tom W]

After reading and participating and even argueing in this thread, I've come to the conclusion that the tripod, mirror lockup, and remote shutter release provide the only easy solution. ;)

Good! We have a solution! :D :D :D

:? Now, what was the problem again? :oops:

:mrgreen: :mrgreen: :mrgreen:

How the crop factor affects altheimer's disease. Or something like that - I forget. 8)

[Tom W]

Good! We have a solution! :D :D :D

:? Now, what was the problem again? :oops:

I think it was something about the tripod length causing mirror shake when using a remote shutter release.

So Tom was wrong, it seems, since he doesn't take tripod cropping into account, let alone the length of the shutter release cord.

I've cropped, chopped, slammed, and lowered my tripod. It's only 12 inches tall now, but it is stylin.

[PacAce]

Steven, give it a rest already, will ya? Nobody is arguing with you about what you would get on the sensor. Yes, for a lens of a given focal length, the image that you get on the sensor, no matter what the size, will be exactly the same with equal amounts of motion blur, if any, captured in the image and the sensor X-factor has no relevance here and does not come into play. Nobody is denying that. We agree with you 100%.

But we are taking that one step further and transferring the image on the sensor to an 8x12 picture and that is where the X-factor comes into play because an image from an APS sized sensor has to be magnified by the X-factor more than a full-frame sensor in order to fit that 8x12 sheet of paper. And with this magnification come an equal amount of magnification of any movement blur captured in the image.

And if you're still not convinced, then I give up. Time to move on to other discussions.

[sharky]

I'm pretty sure I don't have Parkinson's, but thanks for asking :D . This was my first ever post to this forum, and I'm kind of scared by the monster I created. But it has reassured me that the question wasn't as dumb as I originally thought it to be.

After reading though these pages of posts I've realised that one of the big factors here is the switch to digital itself - camera shake probably wasn't such an issue before since I didn't habitually look at images as 100% crops. The same goes for lenses - in the film days consumer grade lenses suited me fine, but digital has shown me their limitations (Hence the L obsession).

[Tom W]

I'm pretty sure I don't have Parkinson's, but thanks for asking :D . This was my first ever post to this forum, and I'm kind of scared by the monster I created. But it has reassured me that the question wasn't as dumb as I originally thought it to be.

It was obviously a good question, as it sparked a lively discussion. I even got to tell one of my little parables. :)

After reading though these pages of posts I've realised that one of the big factors here is the switch to digital itself - camera shake probably wasn't such an issue before since I didn't habitually look at images as 100% crops. The same goes for lenses - in the film days consumer grade lenses suited me fine, but digital has shown me their limitations (Hence the L obsession).

The curse of the 100% crop is that you can see things that you wouldn't ordinarily see on the typical photograph. Whatever you do, don't get an "L" prime! ;)

[PacAce]

I'm pretty sure I don't have Parkinson's, but thanks for asking :D . This was my first ever post to this forum, and I'm kind of scared by the monster I created. But it has reassured me that the question wasn't as dumb as I originally thought it to be.

After reading though these pages of posts I've realised that one of the big factors here is the switch to digital itself - camera shake probably wasn't such an issue before since I didn't habitually look at images as 100% crops. The same goes for lenses - in the film days consumer grade lenses suited me fine, but digital has shown me their limitations (Hence the L obsession).

I guess some people missed the R in your handle and though it was SHAKY. :lol: :mrgreen: :lol:

[sharky]

Damn that's funny!

[Steven M. Anthony]

Steven, give it a rest already, will ya? Nobody is arguing with you about what you would get on the sensor. Yes, for a lens of a given focal length, the image that you get on the sensor, no matter what the size, will be exactly the same with equal amounts of motion blur, if any, captured in the image and the sensor X-factor has no relevance here and does not come into play. Nobody is denying that. We agree with you 100%.

But we are taking that one step further and transferring the image on the sensor to an 8x12 picture and that is where the X-factor comes into play because an image from an APS sized sensor has to be magnified by the X-factor more than a full-frame sensor in order to fit that 8x12 sheet of paper. And with this magnification come an equal amount of magnification of any movement blur captured in the image.

And if you're still not convinced, then I give up. Time to move on to other discussions.

I don't give it a rest because people still have misconceptions about what the CROP FACTOR is.

You say everyone agrees with the image magnification at the sensor. But Ken's proof suggested otherwise. Which is why I pointed out the erroneous assumption behind his formulae.

And if you took the full frame image, cropped off 37.5% around the center, and blew the remaining 62.5% up to 8 X 12, IT WOULD LOOK EXACTLY LIKE THE IMAGE FROM THE SMALLER SENSOR--if the sensors had equal pixel density.

Of course, if you take a smaller original image and enlarge it more than a larger original image, the smaller original is going to look worse. If you take the full frame image and enlarge it again to 16 X 24, it will look worse than the 8 X 12 from the same full frame image. Why would you even waste time discussing the obvious? Unless you compare apples to apples there is no point comparing...

[Scottes]

I don't give it a rest because people still have misconceptions about what the CROP FACTOR is.

As Leo said, we agree with you 100%.


Mods, time to lock this?

[kfong]

But it makes one faulty assumption.


No--you didn't make the assumption explicitly. But it is implicit in the math. The clue to the assumption being there is the fact that there is no term in your equations for image size.

But where and why is the faulty part?

Sigh,
soldering on,
The image size at the sensor is
ys=M1*yo
and the image size at the final viewing screen/port/print is
yp=M1*M2*yo=M*yo.
The above are just the definitions of magnification.
A little geometry will give
FOV=yo/o (radian) when ys=s.and substituting yo and simplifying
FOV~=s/f.

So far so good but this has got nothing to do with camera shake!
For a distance object, camera movement in the x,y,z direction has little effect on blur. The bulk of blur comes from pitch, yaw, and to a lesser degree, roll camera rotation.
Using the same assumption o>>f -> M1~=o/f, a rotation of of w radian in pitch or yaw will cause an apparent translational movement of the object (blur)
dyo=o*w
and the real image at the sensor
dys=f*w
and the magnifed viewing image
dyp=(x*f)*(p/s)*w

Note that the initial object size yo does not matter. Only the rotation ange w matter. The shutter speed t affects w because
w = t * dw, where dw is the angular velocity of the camera rotation due to operator mishanding.
The motion blur is
dyp=(x*f)*(p/s)*dw*t.
which implies as far as viewing the picture with the same print size as concerned, the crop-factored camera behaves as if a x*f focal length is used.

One way to think about it is to consider a window in your house. Consider the window to be the sensor in a camera. when you sit in a particular spot, you get a particular FOV captured by the window--think of it as an image coming through a lens and falling on a sensor. Now if you soap up the outer edges of the window (you've just made your window a smaller sensor) and sit back in that same spot, what you see through the window has a narrower FOV but has not magnified the scene 1 bit. You now have the worst of 2 worlds--narrower FOV and NO Magnification Factor.
This is all correct but how does it apply to the "faulty" assumption?

Ken

[Quinn Porter]

Mods, time to lock this?

I'll second that. All those in favor..

[sharky]

And since I started it, I third the motion to lock it off and move on.

[CyberDyneSystems]

Done.