Further testing of the 50D

[finnbay]

Yesterday's shooting convinced me that I have to do a lot more shooting with my 50D under different conditions to take full advantage of the potential of this camera. Much of my shooting under low light is inside gyms and churches with sports and weddings. Had a chance to take some shots during a volleyball practice today, and you can look at the comparisons. All shot with the Canon 50D and a 70-200L 2.8. Av priority at f/2.8. No IS and hand held. Shot from 1600 ISO, 3200 and 6400 (I will indicate each). Shutters varied from 1/80 to 1/400th depending on ISO and which part of the gym the girls happened to be standing in. Some PP, but tried to keep to a minimum.

ISO 1600

ISO 3200

ISO 6400

[Steve Foss]

Man I am LIKING how well that 6400 performs. Thanks for providing some great examples, Ken.

[finnbay]

I should mention, Steve, that I didn't see ANY banding with this lens compared to some significant banding under low light with the 100-400L. I intentionally shot some dark backgrounds today and didn't have any that had to be taken out.

[Dbl]

Sweeeet!!!! That is what I have found as well Ken, 6400 in this camera is like 3200 in the Mark II's and 20D/30D. I haven't had a chance to try 12800 though under real world testing. Did you give that a try?

[finnbay]

Sorry, Dan. Didn't take any of the girls at 12,800 today. The only shot I took on Hi2 was the side wall of the gym that is located under an overhang without any of the ceiling lights getting to it:

Will have to try that next time out.

[Steve Foss]

I should mention, Steve, that I didn't see ANY banding with this lens compared to some significant banding under low light with the 100-400L. I intentionally shot some dark backgrounds today and didn't have any that had to be taken out.

Interesting. Food for thought.

[mmeyer]

Man that is very good at 6400. I can see why you guys are all raving over it.

By the way, I love that first shot with the girl sharp and the ball blurred. Nice catch.

[DMN]

It is interesting seeing high ISO shots like these, thanks for showing. I also like the expression of the girl in the yellow shorts.

[WCS]

Very impressive. Those shots all look great. I never would have expected the lack of noise at such high ISO settings. With my 30D, I hesitate to shoot indoor sports at anything higher than ISO 1600. Even at ISO 6400, your shots are very clean.

[MN Shutterbug]

I'm surprised the lens makes that much of a difference. The 50D with 70-200 make a heck of a combination.

[Steve Foss]

Mike, Ken and I were talking a lot about that when we were shooting together.

I was struck by how clean my high iso images with the 400 f5.6, 70-200 f2.8L and 300 f2.8L were from the 20D and 30D, compared with all the high iso sports work I did with those bodies and the 100-400. But I've got quite a few band performance images with the 20D and 100-400 from a few years back at iso3200 that were as clean as anything I shot the other day.

My reading indicates that, theoretically, there should be no difference in noise from lens to lens, because lenses don't cause noise. One thing I know for sure: Sharpening increases the look of digital noise, as does saturation and contrast. So the sharper the image is on capture, the less sharpening needed in pp. And though lenses don't cause noise, it doesn't necessarily mean that varying lenses couldn't have some type of impact that makes the noise look more apparent.

And if I remember right, Ken is going to do some side-by-side noise tests on the 50D comparing the 70-200, 100-400 and 300 to see if there really are any differences in apparent noise. It'll have to be a matter of matching iso, lighting and all camera settings from one test image to the other to remove any variables.

Dan, what's your take on apparent noise vs lens choice?

[finnbay]

Steve,

I will be doing that comparison, hopefully in the near future.

[Dbl]

I agree with you Steve, lens shouldn't make a difference in noise and with all the high ISO's I've shot the results seem to confirm that. I also think you are dead on correct when you mention PP work, sharpening, saturation, etc.

The more you have to do to that image in PP the more apparent noise becomes. So using a sharp lens, the 300 or 70-200 requires very little work in PP. I would say close to 95% of my high ISO shots are with the 70-200/2.8 the rest with various lenses. I can look at those shots and tell you which lens was used. So in theory it should make no difference. In real life results....well I think there is a difference when you start messing with the pixels.

One other factor. I think when using faster lenses we don't have the tendency to try and push the exposure that a slower lens can sometimes get you to do. I mean using a 5.6 lens you might try and just get a little more shutter speed, even though you are underexposing so that you can get the shot. Good example was your wolf encounter. You know you don't want to hand hold at 1/20s so you cheat just a bit so you use a 1/30s, or 1/60s shutter speed and you get a shot that is underexposed. With the faster lens you are already at 1/60s with proper exposure so you don't change your shutter speed because you know you can get sharp results at that shutter speed. You end up with a cleaner shot at high ISO because you exposed properly and you were not trying to fix some motion blur in post.

Just a theory based on my experiences and don't get me wrong I'm not at all saying that about your wolf shots, just using that as a type of quick encounter example where you want to get the shot and sometimes you do what it takes to get the shot, even though your equipment has limitations.

[finnbay]

Quote:

I mean using a 5.6 lens you might try and just get a little more shutter speed, even though you are underexposing so that you can get the shot.

Guilty.

[Steve Foss]

Yep, agreed on all that.

And it's quite true that some lenses have a signature "look" to the captures, though in some cases the differences are subtle. To this day, even though it's been going on a year since I owned it, I can usually spot a 100-400 image right off the bat, and it's not a sharpness telltale. Just comes from putting 80,000 shutter actuations through that lens.

Luckily, with the monopod and IS and f2.8, the wolf images turned out OK, but I'd not have hesitated to use EC to underexpose a full top for a faster shutter speed if I'd have had to. In this case, I had one more stop of ISO to rely on, but there have been times in the past that I've accepted the greater noise in order to get enough SS to make the picture sharp. And if I'd had the shutter speed to burn, I'd have gone one stop overexposed on the wolf images so the noise would have been even better. As it was, it was on evaluative metering and I just let the meter take care of business.

[MN Shutterbug]

In the case with the 100-400, wouldn't it be better to back off on the zoom to let's say 300mm, thereby giving you an effective aperture of F4? I don't know where the lens drops to F4, but I've seen somewhere that its upwards of 300mm.

[Steve Foss]

Without one in my hand I'm only going on memory, but I thought it was somewhere around 150mm where the lens bumped from max aperture of f5.6 to f4.5.

Ken could probably settle that one in a heartbeat by putting the lens on the camera and checking where it opens up.

I'm not sure what you mean by "better." It probably wouldn't have an impact on noise, but if you meant it would have been better for the wolf images on the other thread to go from f5.6 to f4.5 on Ken's 100-400, I expect that's true. Every little bit helps, and he ended up plenty close so 150mm would have been enough muscle.

[finnbay]

Just did a quick check - started at 100 and it went from f/4.5 to f/5 at about 135, and then all the way to f/5.6 at about 300.

[MN Shutterbug]

Interesting. Mine hit f/5 at about 130mm and f/5.6 at 250mm. Evidentally, there is a difference in lenses.