File :-(, x, )
Consistent setup for noise comparisons Anonymous
Can any recommend a good, consistent way to set up a couple of cameras to take the same photo to test the high ISO noise of each? I want to do some comparisons to see what sort of improvement some models might offer.

I am particularly wondering what sort of lighting I should work under and what sort of objects I should use to test.

Pic somewhat related.
>> Anonymous
blue
>> NatureGuy !se3A3TwzdY
line up books on your shelf of different colors and font size. Like one green, one blue, one red, etc. Take a photo of them with a tripod using the same iso shutter speed and aperture at the same focal length on as similar quality lenses if you're doing cross brand, or the same lens if they are the same brand. Show the image and the image at 100% crop. ???? profit.
>> Anonymous
>>129667

Thanks. I should be able to give this a go tonight when I'm home from work.

I just bought a 40D and my initial high ISO shots compared to my original Digital Rebel seem to show almost no improvement, so I felt my original test must have been flawed.

I realize the 40D isn't going to do away with noise, but seeing what looks like no improvement in two cameras released 3.5 years apart makes me wonder.
>> Anonymous
>>129682
If you want high ISO performance, get an EOS 5D or shoot in black and white, which looks good through ISO 3200 with ISO Expansion turned on in your Custom Function menu.
>> Anonymous
>>129682
Why don't you just compare the test shots of other camera sites? They've already done the work for you, and most likely did a better job at it.
>> Anonymous
>>129695

I don't think that would allow a good A/B comparison. No photo site is going to compare the original Rebel to the 40D.
>> Anonymous
>>129705
Did you even fucking TRY? dpreview has them both with their standard tests.
>> Anonymous
>>129711
If I remember right, dpreview also says their tests can't be compared cross-test.
>> Anonymous
>>129715
Then why do they have a "compared to..." section in each of their reviews, showing how one camera sizes up to the rest? Why do they work so hard so standardize their testing?
>> Anonymous
>>129717
They shoot it anew, I guess. Different room temperatures, for instance, could skew noise tests pretty badly.

>Note that noise values indicated on the graphs here can not be compared to those in other reviews)

That's taken right from the "Photographic Tests" section of the 40D review.
>> Anonymous
>>129720
That's not such a big deal. Are you telling me that the only way to choose between two cameras based on noise is to buy both cameras and try them out yourself, only to return or sell the one you don't like?
>> Anonymous
>>129720
It's just a disclaimer because it's next to impossible to completely standardize the tests. As far as temperature skewing results, is ONE DEGREE difference in temperature going to so terribly skew the review that you can't use it to judge the improvement (or lack thereof) in noise control? Either way, they've done a far better job standardizing than you're likely going to by just throwing some books together.

And if you're really so concerned about noise levels, you're fucked to begin with because you're blowing the issue way out of proportion. Just take your pictures and don't worry to whatever minuscule difference there is between the cameras.
>> Anonymous
>>129722

OP here. Most of those last few replies weren't mine, but obviously someone in a similar situation.

I'm jus trying to compare a brand new camera (that I already own) to a 3+ year old camera (that I already own) to just get an idea of what I can or cannot get away with.

I own the 40D, I'm probably not going to change it anytime soon, I'm just hoping I can prove to myself that there is indeed some significantly reduction in the appearance of noise between the original Rebel and 40D. It'd be nice to be able to use ISO3200 without having to think much about it.
>> Anonymous
>>129725
the jump should be pretty good in noise reduction I'd assume, but 3200 is pushing it.
>> NatureGuy !se3A3TwzdY
>>129725
3200? LOL

I don't shoot over 400 personally, then again 15x22 prints of animals are kind of demanding.
>> Anonymous
>>129729
lol not really. ive done 20x30 prints at 1600 and if you view it at a normal distance its hard to tell.
>> Anonymous !SDPEsPMnww
>>129725I'm just hoping I can prove to myself that there is indeed some significantly reduction in the appearance of noise between the original Rebel and 40D.

Why do you give a fuck? You already have an awesome camera (I shoot with a 30D) so it really doesn't matter. Just go out there and use the fucking thing.
>> Anonymous
>>129725
Yes, it does control noise significantly better (especially with the new NR shit on) than even the 30D. No, it doesn't perform miracles make shooting at 3200 something you want to do if you can help it.
>> Anonymous
>>129729

Not frequently. I've done a few local concerts, but I don't know how long I'm going to stick with that. Currently trying to see if I can get any sort of regular gig for a few bucks, but who knows. So far, I've actually had great luck shooting the shows at 1600.

>>129730

This makes me feel good. Web versions will still suck, but if I do prints I can probably get away with a little more. I've done a handful of prints just to test the limit of what noise does and does not do, but never over 1600.
>> Anonymous
>>129732

I guess if you use the 30D and have satisfactory results, then I just need to go out there and do it.
>> Anonymous
>>129725
I'm the poster of the stuff about dpreview. I'm not in even a remotely similar situation; noise doesn't really bother me. I'm just pointing out telling someone to go to dpreview isn't the thing to do in this case. Especially since they've changed some things about their testing since the original Digital Rebel came out, and since they shoot processed JPEGs, not raw files, for these tests.
>> Anonymous
Have you ever shot film at ASA 1600? If not, then count your blessings. Until digital photography, 3200 ASA was unheard-of except with extreme pushing or surveillance film. When I push Tri-X to 1600 in diafine, I'm lucky to get the equivalent of 3 megapixels of image data from a 35mm film frame, because the grain is so large (not necessarily good or bad, just what it is.

You're shooting your 10MP EOS-40, complaining that at ISO 3200 your images might not be totally clear. "ISO 3200 without having to worry about it"...

Come on. This is like asking for a car that can drive 180 miles per hour without overheating. Camera and sensor technology has made incredible strides, but don't kid yourself about what you expect or need from your hardware.
>> Anonymous
>>129747
QFT.
>> Anonymous
>>129747This is like asking for a car that can drive 180 miles per hour without overheating.

Only if it's on a treadmill.
>> Anonymous !SDPEsPMnww
>>129750
But will the plane take off?!?!?!
>> Anonymous
>>129747
Tmax 3200?
>> Anonymous
>>129788

From a quick Googling, it appears to be some very high speed Kodak film. From>>129747's comments, I get the feeling it has very large grain and loses a lot of detail.
>> Anonymous
     File :-(, x)
>>129821
Nope, Tmax 3200 has very fine grain.

Camera-Specific Properties:Camera SoftwareAdobe Photoshop CS MacintoshImage-Specific Properties:Image OrientationTop, Left-HandHorizontal Resolution970 dpiVertical Resolution970 dpiImage Created2008:02:10 22:05:07Color Space InformationsRGBImage Width800Image Height588
>> Anonymous
>>129824

Any idea what settings this was shot at?
>> Anonymous
>>129788

TMZ is a very modern tabular-grain film that requires special consideration to development and temperatures. You're right, it does have less grain than pushing or surveillance film, but it's hardly Tech Pan.
>> Anonymous
>>130154
But then Tech Pan is Tech Pan, and you can't make that the standard for acceptable noise when the vast majority of other films don't even approach it.
>> ac !!VPzQAxYPAMA
>>129682
>I realize the 40D isn't going to do away with noise, but seeing what looks like no improvement in two cameras released 3.5 years apart makes me wonder.
The Digital Rebel is a 6MP camera. The 40D is 10MP. Canon sacrificed high ISO noise ability for cramming in more pixels. I.e., there has been significant improvement when you consider noise per micron of sensor size, but in terms of actual image quality, it's stayed about the same. If you really want clearer high-ISO shots, you need to move to a larger sensor size.

In theory, it seems like Canon should be able to make a 6MP APS-C digital SLR at this point that can shoot ISO1600 noiselessly...
>> Anonymous
>>130493
40D's sensor itself is *worse* than 10D and Rebel in terms of per-pixel noise, but improved on-chip noise reduction makes up for that (our eye doesn't notice the slight loss of chroma information, and Canon, unlike Sony, is clever enough not to overdo it)