When 'rules of thumb' are wrong

[FlyingPete]

After considerable shooting on my 20D in low light conditions I have concluded that the old rule that you should shoot at least the inverse shutter speed of your focal length (i.e, 200mm = 1/200s or faster, or 1/320s on 1.6 cropped cameras) is wrong. If you really want a sharp image you need to go at least twice as fast as this, any images I have shot handheld like the above lack the sharpness of images with much faster shutter speeds. What is everyone else’s experience with this?

Also the good old shoot a the nearest inverse shutter speed of you ISO at f/16 (i.e, ISO 100, shoot at 1/100s) on a sunny day seems to work very well, I am sure distance will vary depending on latitude and season though.

[Belmondo]

Don't forget the X factor if you're using a camera other than one with a full-frame sensor. (1Ds)

If you use the 1/focal length rule, the 200 mm lens in your 20D becomes functionally a 320mm lens. This is significant, and your 1/200 sec. is going to be somewhat slow.

[tim]

Don't forget the X factor if you're using a camera other than one with a full-frame sensor. (1Ds)

If you use the 1/focal length rule, the 200 mm lens in your 20D becomes functionally a 320mm lens. This is significant, and your 1/200 sec. is going to be somewhat slow.

I don't believe this to be, theoretically. The 1.6X is a crop factor, not a magnification factor, so you're just getting a smaller part of the image. On the other hand because you're enlarging it more maybe you need extra sharpness to compensate.

Hmmm. Must try some tests myself some time.

[Andy_T]

Also, don't forget that some people have more steady hands / better technique than others enabling them to hold slower shutter speeds steady.

I'm afraid that I belong to the 'others' category, so I also go for as fast a shutter speed as possible normally.

Best regards,
Andy

[Belmondo]

I don't believe this to be, theoretically. The 1.6X is a crop factor, not a magnification factor, so you're just getting a smaller part of the image. On the other hand because you're enlarging it more maybe you need extra sharpness to compensate.

Hmmm. Must try some tests myself some time.

You don't have to. In this case, I'm right. I used to think that it was strictly a matter of focal length, and 200mm lens was still a 200mm lens regardless of sensor size.

But this is where the term 'crop factor' might tend to be a little misleading because what is created on the sensor is a magnification equivalency. Camera shake as seen on a full-frame sensor will be the same as on a 1D (1.3X) or D60/10D/20D (1.6X). BUT that amount of shake stated as a percentage of the sensor size will be magnified by an amount equal to the crop factor. It will be 30% greater on a 1D, and 60% greater on cameras with APS-sized sensors.

I probably haven't explained it well. It is a bit confusing at first, and I must confess to having misunderstood the concept at first myself.

Just try to visualize how equal amounts of camera shake will affect the image on different sensor sizes, and you'll get there.

Tom

[tim]

I'm not sure I understand what you mean - can anyone expand on the idea? I'm an engineer so i'm not afraid of technical terms ;)

[RockOne]

I'll have a go at explaining what I think is meant here !

Imagine you are going to print a 1D and 20D photo at 12x8. The image using a standard 50mm lens (for example) will show a samller area, which means that it has been magnified relative to the 1D image, and any motion will be magnified with it. If you were to make the image from the 20D the same size as the cropped image from the 1D then the motion blur (given idential shooting conditions) should be equal.
Confused yet ?. I think I am :-) .

[Belmondo]

Tim:
Since you're an engineer, I have every confidence that you'll figure it out.

[mdr]

I wonder how steady Skippy's paws are for handholding at low speeds :lol: .

[RockOne]

I wonder how steady Skippy's paws are for handholding at low speeds

Absolutley fantastic. Two legs + tail + tripod - problem solved :-) :-) :-) !

[sdommin]

I don't believe this to be, theoretically. The 1.6X is a crop factor, not a magnification factor, so you're just getting a smaller part of the image. On the other hand because you're enlarging it more maybe you need extra sharpness to compensate.

You answered your own question. Since you are enlarging more, any camera shake will become more visible.

[mdr]

A rule of tumb is just a rule of thumb. It only applies to those who invented it.

In fact if you have steady hands, you'll probably be able to get sharp pics with slower speeds. I can generally half or even quarter the speed based on this rule and still get sharp pics http://www.photography-on-the.net/forum/images/smilies/icon_cool.gif.

Remember, there are lots of other factors affecting camera shake.

Shake is highly dependent on how you hold the camera. Both hands on the body with a heavy telephoto lens attached is pretty shaky compared to with the left hand underneath the lens and the right hand on the body. Anyone tried it with their hands the other way round http://www.photography-on-the.net/forum/images/smilies/icon_lol.gif?

The actual weight of camera and lens has an impact. The heavier, the more difficult to keep steady, especially when sustaining the same position some time before taking the shot.

A grip with vertical shutter button will improve stability when shooting in portrait format.

How do you press the shutter button? Ever tried taking a shot with a 200mm at 1/125 or 1/60 handheld with and without the self timer. You'll definitely see a difference there.

Apart from shaky hands, body posture and temperature play their part as well. Having a relaxed pose or even leaning against a tree or wall will help steadying anyone. In freezing weather conditions, I'm shaking anyway, irrespective of what I'm doing ;).

If you have altzheimers, you'll need a tripod as well as a zimmer frame :evil:.

If you've taken plenty of pictures, you'll now at what speeds you can handhold while still getting sharp results. If you haven't, get out there and get snapping.

Maybe this is all academic, as Canon's IS lenses don't abide by the rule. In fact, it may be time to dump the Canon and switch to Konica Minolta with it's in-body imaga stabaliser technology.

[mdr]

P.S. Yes, I'm twiddling my thumbs at work...

[mdr]

... and you can always fill the pouch with stones to steady the tripod :lol: .

[scottbergerphoto]

If you have altzheimers, you'll need a tripod as well as a zimmer frame :evil:.


That's true but you won't remember you took any pictures!
Scott

[RockOne]

Apart from shaky hands, body posture
And at weddings and parties the rule of thumbs is something like shutterspeed is inversly proportion to the number of mls of alchohol consumed :-) :-).

... and you can always fill the pouch with stones to steady the tripod .
:-) :-) :-).

And with rules of thumb...there are exceptions to every rule (mostly :-) )

[Cadwell]

The applicability or otherwise of a rule of thumb depends on who’s thumb you are using. :p

[RockOne]

depends on who’s thumb you are using.
And if you are clumsy and "all thumbs", it depends on which of your own thumbs you are using :-).

[CyberDyneSystems]

Rules of thumb are about as appplicable as

"i before E except after C" ... Yeah,.. right...

[22littlereasons]

Interesting rule of thumb. Never heard of it before. I'm thinking if you can't hold the camera steady enough after 1/320 - 500 then your lens is too heavy and you need a monopod or tripod or medication. *s*

[maderito]

After considerable shooting on my 20D in low light conditions I have concluded that the old rule that you should shoot at least the inverse shutter speed of your focal length (i.e, 200mm = 1/200s or faster, or 1/320s on 1.6 cropped cameras) is wrong. If you really want a sharp image you need to go at least twice as fast as this, any images I have shot handheld like the above lack the sharpness of images with much faster shutter speeds. What is everyone else’s experience with this?
Well, I've enjoyed the humor in this thread - but the original question remains an important one.

FlyingPete's observations are well-grounded. While shooting in low light, your images are more likely to seem less sharp. The problem -- there's much that can go wrong:

1. Apertures are wider; DOF is smaller - and usually much smaller than we realize.
2. Lenses are softer at wider apertures.
3. The 1.6 crop is a factor as already pointed out.
4. High ISO images will always look less sharp on close examination - even with superb noise reduction techniques.

Typically in low light settings, you are working at the limits of what is technically achievable from the lens, camera and operator. Thus failure rates are high -- and often very frustrating.

Personally - I sometimes like the soft look of available light shots whether created by lens factors, camera shake, subject motion or some combination of all.

OK -- let's get back to the jokes. :) :)

[jfred]

What rule of thumb is programmed into the 20D in the program modes? My eos300 (35mm) goes with the 1/focal lengh of lens as far as possible

If you're at a wedding reception, perhaps a spirit level viewfinder display would be a good accessory to have to mitigate against the effects of the alcohol :) Dammit - should have patented that idea before posting!

[Monito]

The applicability or otherwise of a rule of thumb depends on who’s thumb you are using.

Indeed it does. The origin of the term is that under old English law (predating the colonization of America), a man could not be convicted of illegally beating his wife if he used a rod (stick) no thicker than his thumb. To whom it may concern: have you stopped beating your spouse?

[jfred]

...To whom it may concern: have you stopped beating your spouse?

that would probably cause camera shake :rolleyes:

[Belmondo]

Indeed it does. The origin of the term is that under old English law (predating the colonization of America), a man could not be convicted of illegally beating his wife if he used a rod (stick) no thicker than his thumb. To whom it may concern: have you stopped beating your spouse?

Men with very small thumbs were no doubt highly prized by women of the day.

[Redbird_xo]

...To whom it may concern: have you stopped beating your spouse?

Years back a Taiwanese friend of mine told me that I shouldn't start the habit of beating up the wife...it's addictive, he said.

Seriously, for the 300D and tele shots, I always use at least '1.6 x actual focal length' equivalence in shutter speed if not higher rather than the 1/focal length rule for 35mm cameras. Because the sensor on D300 is smaller than 35mm x 24mm, the image has to be magnified. Consequently, the vibration of hand holding the camera will also be magnified.

[Monito]

As mdr says, shake depends on many things and some photogs can do better than others. mdr identifies some important factors which I will summarize here and augment:


Practice.
Left hand under body & lens, right hand on body and trigger.
Torso tripod: elbows gently against body, camera against face.
Body tripod: one foot forward, other back and angled slightly.
Relaxation: gently squeeze at the bottom of a breath.
Lean against things when possible.
Hold the camera against a wall or column or rock.
Sit down to take the picture.
Use a string pod.
Use a tripod or monopod as much as is practical.
Don't be lazy and leave the tripod at home.
Stay trim, don't smoke, keep fit. An enlarged heart shakes more.
Meditate so you can be in the zone and not get excited.
Make pictures in preference to taking them.
Did I mention practice? Many problems with technique can be improved by making 100,000 pictures thinking as you make them and again as you review them later.

Successful slow speed shots is a question of playing the odds. One might be able to get sharp pictures 97 percent of the time at 1/125, 80 percent at 1/60, 50 percent at 1/30, and 20 percent at 1/15. Take extra shots to increase the odds in your favor.

Digital "film" is cheap! Don't forget to practice (shoot early and often).

[cc10d]

The crux of the issue of camera movement is this


Belmondo said:

But this is where the term 'crop factor' might tend to be a little misleading because what is created on the sensor is a magnification equivalency. Camera shake as seen on a full-frame sensor will be the same as on a 1D (1.3X) or D60/10D/20D (1.6X). BUT that amount of shake stated as a percentage of the sensor size will be magnified by an amount equal to the crop factor. It will be 30% greater on a 1D, and 60% greater on cameras with APS-sized sensors.

Thus if we are holding with the "rule of thumb", we need to use the effective magnification at full size sensor, because the "rule" was originally based on the 35 mm format.

[robertwgross]

Tim, I'm an engineer, and I would go with Belmondo's rule.

Many years ago, during a brief military period, I had spent enough time on the rifle range that it had helped my long-lens photography skills. Alas, after a long time, that has worn off.

---Bob Gross---

[Belmondo]

The crux of the issue of camera movement is this


Belmondo said:

But this is where the term 'crop factor' might tend to be a little misleading because what is created on the sensor is a magnification equivalency. Camera shake as seen on a full-frame sensor will be the same as on a 1D (1.3X) or D60/10D/20D (1.6X). BUT that amount of shake stated as a percentage of the sensor size will be magnified by an amount equal to the crop factor. It will be 30% greater on a 1D, and 60% greater on cameras with APS-sized sensors.

Thus if we are holding with the "rule of thumb", we need to use the effective magnification at full size sensor, because the "rule" was originally based on the 35 mm format.


If you're agreeing with me, then that's correct. :lol:

Depending on the size of he sensor in your camera, you will have to multiply the focal length by the appropriate 'X-factor' to get the reciprocal of the suggest shutter speed.

i.e. a 200 mm lens on a 20D with a 1.6X crop factor would suggest a shutter speed of 1/320 second.

[Monito]

Camera shake is based on angle of view. If an angular movement of the camera sweeps the edge one degree, then this is compared to the angle of view of the image on the sensor or film. A crop factor decreases the angle of view, thus increasing the effect of shake. Increasing focal length decreases the angle of view, thus increasing the effect of shake.

[Belmondo]

A crop factor decreases the angle of view, thus increasing the effect of shake. Increasing focal length decreases the angle of view, thus increasing the effect of shake.

Yup. Almost sounds like a conspiracy, doesn't it? But that really is the case. For purposes of this discussion, the first statement is most apt.

[pcasciola]

I'm not sure I understand what you mean - can anyone expand on the idea? I'm an engineer so i'm not afraid of technical terms ;)

Ok, I'll give it a shot. See if this math works.

The 20D has a 22.5mm wide sensor, with 3504 lines, so it's about 156 lines per mm of sensor. The 1Ds has a 36mm wide sensor, with 4064 lines, which is about 113 lines per mm. So, if you "shake" 1/50th of a mm, on a 1Ds that would translate to about 2 pixels of blur, and on a 20D you would get get about 3 pixels of blur.

[Andy_T]

Plus, the 1Ds has more pixels in every dimension, so a blur of x pixels would not be as obvious as the same amount of blur on the 20D.

Best regards,
Andy

[Persian-Rice]

Rules are meant to be broken.

[Jon]

I'm not sure I understand what you mean - can anyone expand on the idea? I'm an engineer so i'm not afraid of technical terms ;)

The amount of shake you record on the film/sensor plane when hand-holding a 200 will average the same regardless of what sensor format you're using. BUT, to get a given print size, you'll have to magnify the recorded image different amounts (the DR/10D/20D image will need to be enlarged 1.6X more than the 35 mm full-frame). When you magnify the image, you magnify the recorded shake. So, you need to allow for the crop factor in deciding safe hand-holding speed, since you're enlarging everything more.

[CyberDyneSystems]

Jon,.. you put that perfectly! ;)

Just to add to that.... if you were shooting 35mm high quality film,.. and had a 300mm lens attached,. and while you were framing your subject you realized that your lens was not long enough to giv the framing you wanted,. and knew that your resulting print would then have to be made from an enlarged crop taken from the full frame 35mm image... let's say a 40% crop ;) ...

... if you had the forsight to make all those decisions while shoting with your 35mm and 300mm lens,...

then you may also have the forsite to realize that when it came to enlarging that crop for printing,. you may run into trouble with maintaining detail on the enlarged crop Vs. the relatively smaller print per emulsion of an uncropped image.

You may even realize that you could perhaps keep those details clearer in the enlarged crop IF you had a faster shutter speed than your standard rule of thumb of 1/300 for a 300mm lens.
:)

[CyberDyneSystems]

Of course this was delved into specifically in a thread immortalized here;

-=10,000,000 Posts on the X-Factor=- (http://photography-on-the.net/forum/showthread.php?t=45388)

...second to last link....
X-Factor and 1/1 focal length/shutter speed rule (http://photography-on-the.net/forum/showthread.php?t=48748)

Oh what the hell,. I'll add this thread too :)

[Jon]

Jon,.. you put that perfectly! ;)

Thanks! I'm now working on the complete, concise(?) cra^hop factor explanation using the same relentless application of something (logic? triumph of hope over realism?). So far it's about 50 pages without the formulae.

;{)#

[FlyingPete]

Well this has been intresting, posting a little observation just before bed, only to wake and find this thread, I see we were unable to escape the inevidable X Factor discussions, however careful reading of my original post shows that I have taken this into account ;) We even had a discussion on wife beating, an interestig diversion that was likly to come up in refererence to 'Rule of Thumb'.

Anyway something we missed, how about IS? How much does this really help? Most of my issues were when I was doing portrait shots at long focal lengths, someones comment about a gip helping was a good point, and I believe IS doesn't work on the vertiacl axis on the 75-300IS?

Next year when I am shooting this event again, I am getting a monopod, and a faster telephoto lens (started saving for a 70-200/2.8L) :cool:

[Jon]

IS is non-direction-specific unless you're in Mode (1? - panning) AFAIK, which isn't an option on the 75-300. And I think that's effective regardless of the camera's orientation; the lens' brain just cancels out IS functionality in the direction that's seeing a steady motion. I'd just apply the same mod to Canon's stated limits - figure on it giving you about 2 stops beyond what you could normally do.

[Monito]

Since everybody and every body is different, and since cameras differ, and since sensor sizes differ, and since lenses differ, and since situations differ, there is only one real answer: get to know your equipment and your capabilities by practicing by making lots of pictures. Digital imagery is cheap so try things out and try to stretch yourself by attempting pictures you might have skipped when using film. When in doubt, shoot; and shoot multiple tries in borderline situations to increase your odds of a good take.

Did I say practice? I hope so.

[cc10d]

Belmondo, YES that was intended to be a confirmation and reqognition of your accuracy, using your statement as the solid reality of the issue.

[mdr]

This has been one of the most enjoyable threads I've read for a while :D .

Think we should ask Canon to update there firmware of the 20D to change the program modes to take account of our new calculations?

[Andy_T]

Think we should ask Canon to update there firmware of the 20D to change the program modes to take account of our new calculations?

That would be nice. My 20D keeps suggesting 1/50 second at 75 mm lens focal length in P mode, even when I have the flash open. Somehow they either got that rule wrong, or they're overestimating my handholding abilities.

Best regards,
Andy

[mdr]

What is P mode? ;)

[HKFEVER]

What is P mode? ;)

Fully Automatic, I guest.:oops:

[Scottes]

Actually, almost fully automatic. It IS automatic but you can adjust some things like EC.

[mdr]

Notice the question was asked with a ;).

Of course I know what the I-never-ever-use P mode is :lol:.

By the way, fully automatic is the green square mode.

[Redbird_xo]

P or green box modes sometimes spit out very nice pictures.

[FlyingPete]

P or green box modes sometimes spit out very nice pictures.

Note sometimes. I never use green box, not sure what it is doing on a camera of that level, infact I fell a poll comming on...

[rfreschner]

not sure what it is doing on a camera of that level

Perhaps so Canon can sell higher end cameras to folks that really don't need them, but just want a cool camera with all the bells and whistles. ;)

Rick