Something I learned about "sharpness" on the 7D

I have been "playing" with my new camera learning its features etc. and, to make a long story short, I noticed something: Since this camera has such a high pixel density (18MP on APS-C) if you are using 100% crop to assess sharpness of focus etc, you really have to be careful, because 100% really "zooms" in. I found for me personally, I need more shutter speed than 1/focal length handheld even with the IS if I am going to use 100% crop and expect that view to be tack sharp. I haven't calculated what the effective "zoom" is on this camera for 100% crop, but it is extreme, and any small motion is greatly magnified, and it makes the image look soft. It is very subtle.

When I increased shutter speed, all was well. I wonder if anyone else has noticed this. At first, I thought my lens might need microadjustment but it doesnt. It was very small motions being magnified. The 1/FL rule of thumb was made in the film days, when we didn't have the luxury (or curse) of being able to look at 100% easily without a loupe.

I agree, with this resolution even the smallest amount of camera shake is magnified. Shutter speed was one of a number of things I had to adjust to get to the point where I was satisfied with my shots on the 7D. The real bear was the new AF system.

Stop looking at 100% crop and look at your prints instead--you're judging by the wrong metric.

The physical size of the blur permitted by a given shutter speed is precisely the same size on the sensor regardless of the resolution of the sensor. If 1/60 allows 0.01mm of blur across an 8mp sensor, it allows the same amount of blur across an 18mp sensor. When you enlarge both sensors to the same final display size, the amount of blur will be precisely the same.

The difference will be that areas of the image not blurred will be more highly resolved on the more highly resolving sensor. If all other factors are equal the resulting print will never be worse from the higher resolving sensor.

When you look at the individual pixels at 100%, you are being deceived as to the appearance of the final display. When you look at an APS-C 18mp image at 100%, you are looking at the APS-C frame at three times the enlargement of looking at an APS-C 6mp image at 100%.

If there is a 0.01mm blur on both images, you're looking that same sized blur at three times the enlargement on the APS-C 18mp image at 100% as you're seeing it on the APS-C 6mp image at 100%--so for sure, it looks like more blur because you're looking at a greater enlargement. Look at the rulers! That APS-C 18mp image at 100% is a greater print size.

The appearance of blur in an image varies directly with the degree of enlargement. Three times more enlargement makes the blur look three times worse. If you look at them at the same degree of enlargement, an APS-C 6mp image at 100% equals an APS-C 18mp image at 33.3%.

Props for making an important realization regarding shutter speeds. Many never make that connection and simply cry "My Camera/Lens Is Soft!"

RD's suggestion though is valid. You can make yourself nuts with excessive pixel peeping.

Yes, in film days only the very best shots were printed larger than 4x6 and you can cover a lot of shake at that 4x enlargement. Slides were projected large but usually in a room with 15-20 feet viewing distance. Not just camera shake but focus and depth of field, even lens optics must be perfect for things to look good at 100% these days. If it looks good when viewing the whole scene then you should be satisfied.

I look at 100% crops, but I have calibrated my eye to know how prints will look against them.

If you get a new camera with a very different pixel count you have to re-learn that calibtration. Some people do not understand this, which is one of the reasons that the 50D (especially) and the 7D ( to a lesser extent) took a beating as the first units hit the field.

It's okay to look at 100% crops for some editing purposes... But I usually don't evaluate images at more than 50%. It's much more as the image will print. Depending upon your monitor and the ppi you have set for the files, you may need to reduce even more to see an accurate print size. On my monitor I usually use 37% when the image is 240 ppi.

Using a Spyder3 on your eyeball?

JeffreyG wrote in post #9184661
I look at 100% crops, but I have calibrated my eye to know how prints will look against them.

If you get a new camera with a very different pixel count you have to re-learn that calibtration. Some people do not understand this, which is one of the reasons that the 50D (especially) and the 7D ( to a lesser extent) took a beating as the first units hit the field.

That's a very good point.

ASA_5 wrote in post #9184411
I have been "playing" with my new camera learning its features etc. and, to make a long story short, I noticed something: Since this camera has such a high pixel density (18MP on APS-C) if you are using 100% crop to assess sharpness of focus etc, you really have to be careful, because 100% really "zooms" in. I found for me personally, I need more shutter speed than 1/focal length handheld even with the IS if I am going to use 100% crop and expect that view to be tack sharp. I haven't calculated what the effective "zoom" is on this camera for 100% crop, but it is extreme, and any small motion is greatly magnified, and it makes the image look soft. It is very subtle.

When I increased shutter speed, all was well. I wonder if anyone else has noticed this. At first, I thought my lens might need microadjustment but it doesnt. It was very small motions being magnified. The 1/FL rule of thumb was made in the film days, when we didn't have the luxury (or curse) of being able to look at 100% easily without a loupe.

The 1/focal length rule has been updated for the Digital Era. It is now 1/(focal length x crop factor). Hence for your 7D you should multiply focal length by 1.6 to get the right "rule of thumb" shutter speed.

Thanks RD, for a very well presented explanation of that phenomenon. Very elegant.

nureality wrote in post #9184865
The 1/focal length rule has been updated for the Digital Era. It is now 1/(focal length x crop factor). Hence for your 7D you should multiply focal length by 1.6 to get the right "rule of thumb" shutter speed.

Good point. When I was doing my test, I just doubled it and it was easy to see the difference. Maybe I shake a bit more than most as well. Who knows.

RD, those are good points, and I think we are saying the same thing. I was not holding output size constant. I was simply making an observation based on what many people do with digital images: hit the 100% button, and then complain that my images are soft. Spreading 18MP out over the typical screen resolution on the computer expands the "blur" making it look softer. Many people don't do much printing, so they may not realize this issue. When I zoom to 100% or less, it is primarily to check how I did on focus.

nureality wrote in post #9184865
The 1/focal length rule has been updated for the Digital Era. It is now 1/(focal length x crop factor). Hence for your 7D you should multiply focal length by 1.6 to get the right "rule of thumb" shutter speed.

I don't think the crop factor really applies since you are only cropping the image not increasing the actual zoom ability of the lens. 300mm is still 300mm just cropped when using an aps sensor.

More
https://photography-on-the.net …1+x+focal+lengt​h+rule+aps

stellgar wrote in post #9184967
I don't think the crop factor really applies since you are only cropping the image not increasing the actual zoom ability of the lens. 300mm is still 300mm just cropped when using an aps sensor.

More
https://photography-on-the.net …1+x+focal+lengt​h+rule+aps

Its not about zoom ability. Its about effective focal range. (Lets not get into the whole crop vs. reach controversy here.) But adding the multiple gives you a better shutter speed estimate needed to eliminate camera jitter. This is especially relevant with longer focal lengths.

Try shooting a 300mm lens on a crop body without IS. If you don't stabilize the lens (on a tripod/monopod) you will see camera shake blur at shutter speeds slower than 1/500.

I've tested this with a 300mm prime (see "The Thing" in my gear list) and with a 300mm zoom (see P.O.S. EF 75-300).

I am not saying that what you suggest is wrong in fact for many it is probably right on, but with practice you can get much slower than factor of one over the focal length.

Taken from across the room using available light from a desk lamp at 1250 ISO 300mm no IS.
Elbows against ribs, breath in, hold breath, squeeze trigger, don't push or pull.

These were resized an compressed for posting.

Hand held 300mm @ 1/50 sec. ISO 1250

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

HTTP response: 404 | MIME changed to 'image/gif' | Byte size: ZERO | PHOTOBUCKET ERROR IMAGE

^ You must be very steady.

Because of increasing pixel counts, I'm constantly amazed that we use this "100% crop" metric, and compare it across cameras of differing pixel densities. I've seen discussions where people said "but 100% crops on my 20D were far sharper than 100% crops on the 7D!".

RDKirk explains why above, and I agree with Mark, it's well-written.