Canon Xsi Dynamic Range

[marcmccalmont]

Does any one have a comment on the Canon Xsi's dynamic range in comparison to other cameras i.e. 5D, G9 etc.?
Thanks
Marc

[Panopeeper]

I have only one Xsi raw file, ISO 400; this is technically comparable to a G9, ISO 400 as well. The G9 is in the league of "prosumer" cameras, far from the 450D.

You have a 5D. If you upload an ISO 400 image with dark, uniformly lit, uniform color, smooth spots, I compare that with the 450D I have (the best is the same scenery with different exposures, so that there are some spots with illumination comparable to the 450D image).

[feppe]

Does any one have a comment on the Canon Xsi's dynamic range in comparison to other cameras i.e. 5D, G9 etc.?
Thanks
Marc
[{POST_SNAPBACK}][/a]

Answer [a href=\"http://www.dpreview.com/reviews/canoneos450d/page21.asp]here[/url]. Open dpreview's 5D or G9 review and look for their DR numbers. I think they changed the methodology a while back so 5D's DR figures might not be 1:1 comparable, but should give an idea on the difference.

Just checked since I'm considering buying the 450D myself to upgrade from 30D: 5D has 0.5 less DR according to dpreview.

[Panopeeper]

Open dpreview's 5D or G9 review and look for their DR numbers

Several of DPReview's claims regarding the dynamic range are so unserious, that their charts can not be used for anything.

For example the claim, that the DR of the 450D is the same at ISO 400 as at ISO 200, and only 0.1 EV less at ISO 800 than at ISO 400 is almost as ridiculous, as the claim, that the DR of the 40D is the same from ISO 400 to ISO 1600, and that ISO 3200 is 1.5 EV less than ISO 1600 (there is no ISO 3200 on the 40D, it is simply the doubling of ISO 1600 pixel values, which results in *exactly* one stop less DR).

[feppe]

Several of DPReview's claims regarding the dynamic range are so unserious, that their charts can not be used for anything.

For example the claim, that the DR of the 450D is the same at ISO 400 as at ISO 200, and only 0.1 EV less at ISO 800 than at ISO 400 is almost as ridiculous, as the claim, that the DR of the 40D is the same from ISO 400 to ISO 1600, and that ISO 3200 is 1.5 EV less than ISO 1600 (there is no ISO 3200 on the 40D, it is simply the doubling of ISO 1600 pixel values, which results in *exactly* one stop less DR).
[a href=\"index.php?act=findpost&pid=206039\"][{POST_SNAPBACK}][/a]

Good points. But since the dpreview method is using real-world instead of theories, maybe this is the result?

Of course it could very well be that their methodology is fatally flawed...

[Panopeeper]

Good points. But since the dpreview method is using real-world instead of theories, maybe this is the result?

Do you mean it is only theory, that ISO 1600 (meaning 2 EV lower exposure) results in lower dynamic range than ISO 400, in the same numerical range of pixel values?

Do you mean it is only theory, that if the exposure is reduced by 1 EV and there is no ISO extra gain associated with the reduced exposure, then the dynamic range is reduced by one stop? Actually, this is equivalent to the very definition of photographic dynamic range.

Re real world: I attached two screen captures; one with Canon 40D, the other with Canon 450D. Both with ISO 400. The shutter times are very different, but still in the range of comparable.

I selected a small patch on both raw images, uniformly lit, smooth areas (as smooth as I found on the only 450D image I have). These patches show the same intensity. The average red raw and R values are almost the same: 450D = 169, 40D: 170; the standard deviation of the 450D is 20,of the 40D is 18, i.e. the 40D shows *slightly* lower noise. Thus the DR of the 40D is slightly larger than that of the 450D.

Note: only one channel can be compared at once, because the values in the other channels do not match between the two images, due to the different scenery and different illumination (and perhaps due to the different color filters as well). However, this has not relevance regarding the validity of this comparison.

[picnic]

Do you mean it is only theory, that ISO 1600 (meaning 2 EV lower exposure) results in lower dynamic range than ISO 400, in the same numerical range of pixel values?

Do you mean it is only theory, that if the exposure is reduced by 1 EV and there is no ISO extra gain associated with the reduced exposure, then the dynamic range is reduced by one stop? Actually, this is equivalent to the very definition of photographic dynamic range.

Re real world: I attached two screen captures; one with Canon 40D, the other with Canon 450D. Both with ISO 400. The shutter times are very different, but still in the range of comparable.

I selected a small patch on both raw images, uniformly lit, smooth areas (as smooth as I found on the only 450D image I have). These patches show the same intensity. The average red raw and R values are almost the same: 450D = 169, 40D: 170; the standard deviation of the 450D is 20,of the 40D is 18, i.e. the 40D shows *slightly* lower noise. Thus the DR of the 40D is slightly larger than that of the 450D.

Note: only one channel can be compared at once, because the values in the other channels do not match between the two images, due to the different scenery and different illumination (and perhaps due to the different color filters as well). However, this has not relevance regarding the validity of this comparison.
[a href=\"index.php?act=findpost&pid=206054\"][{POST_SNAPBACK}][/a]

All interesting as I'm probably going to sell my G9 and buy a 450 to use as my 'small' auxillary camer, esp. with a 28 f/1.8 and 50 f/1.4.  I've never used a Rebel, not even held one so even though I've compared specs with my 5D and think I would be happy with it as my hiking/urban walking camera for non-photographic centric times (but still wanting a very good camera), hearing more about it 'real world' is helpful in making a decision.

Diane

[feppe]

Do you mean it is only theory, that ISO 1600 (meaning 2 EV lower exposure) results in lower dynamic range than ISO 400, in the same numerical range of pixel values?

Do you mean it is only theory, that if the exposure is reduced by 1 EV and there is no ISO extra gain associated with the reduced exposure, then the dynamic range is reduced by one stop? Actually, this is equivalent to the very definition of photographic dynamic range.

Re real world: I attached two screen captures; one with Canon 40D, the other with Canon 450D. Both with ISO 400. The shutter times are very different, but still in the range of comparable
...
[a href=\"index.php?act=findpost&pid=206054\"][{POST_SNAPBACK}][/a]

No need to get defensive. I was merely making the point I learned in chemistry class in college: theory sometimes doesn't work out in the lab, and rarely in real life. And pixel-peeping has that tendency as well - there are quite a few pixel-peeping articles which come to a similar conclusion.

As for your own DR calcs, how would you qualify that "slight" edge of the 40D? Significant? I'm personally mostly concerned about shadow noise in long (30-300 second) night-time exposures.

[The View]

No need to get defensive. I was merely making the point I learned in chemistry class in college: theory sometimes doesn't work out in the lab, and rarely in real life. And pixel-peeping has that tendency as well - there are quite a few pixel-peeping articles which come to a similar conclusion.

As for your own DR calcs, how would you qualify that "slight" edge of the 40D? Significant? I'm personally mostly concerned about shadow noise in long (30-300 second) night-time exposures.
[a href=\"index.php?act=findpost&pid=206117\"][{POST_SNAPBACK}][/a]

Believe me, all those cameras have good noise characteristics.

But you just wrote a point you haven't mentioned before: you want to do night-time shooting from a tripod.

Search the forum. There was actually one thread about someone asking about cameras good for long exposures.

[The View]

Good points. But since the dpreview method is using real-world instead of theories, maybe this is the result?

Of course it could very well be that their methodology is fatally flawed...
[a href=\"index.php?act=findpost&pid=206040\"][{POST_SNAPBACK}][/a]

I think it is. They shoot a few objects in a lightbox.

They count resolution lines.

(Their sample images are snaps like a tourist would make them, and no challenge at all to a camera, not taking any advantage of a camera's possibilities.)

Actual daylight shooting conditions are much more complex than their lab set-ups.

It's like taking a racing car, and just drive it round the block instead of racing it.



Then they do these extreme close-ups that don't show how the cameras process colors, how smooth the tone is, how it can catch different moods of light with it.

[Panopeeper]

No need to get defensive. I was merely making the point I learned in chemistry class in college

I did not want to come accross as being defensive. Your questioning the validity of the claims was partly justified, and I gave the proof. (The issue qwith ISO 3200 is simple arithmetic.)

And pixel-peeping has that tendency as well - there are quite a few pixel-peeping articles which come to a similar conclusion

Pixel peeping alone means nothing. I am analyzing the unadultered raw images, to exclude the effects of different raw processors, different settings, etc. I wonder if anyone cames to a different result on this factual basis.

As for your own DR calcs, how would you qualify that "slight" edge of the 40D? Significant?

Generally, I would say the difference is far less important, than the difference cause by not perfectly exposing (ETTR), or not selecting the most suitable ISO, etc.

I'm personally mostly concerned about shadow noise in long (30-300 second) night-time exposures

This is a special consideration; cameras need to be tested specifically to this purpuse. I can't help much, but if you cam convince some people to make test shots (raw) with appropriate settings, I can evaluate the results.

Several people contributed to a comparison with 5sec blackframe shots at ISO 1600, between Nikon D300, Canon 1DMkIII, 350D, 20D and several 40Ds (I myself have a 20D and a 40D). The results are telling, but

1. that was only 5sec,

2. the 450D was not even on the market that time.

The test shots were made with lens cap (or better: body cap) on, viewfinder sealed. Only for a feeling, how the results were evaluated, see following manylayer files with captures, crops (all from the black frames).

Histograms, raw images

Crops from ACR converted blackframes

[marcmccalmont]

Bottom line is I am disappointed with the G9 files but need a small travel camera.
A lot of my shots are in lower light or into the sun, so the smaller sensor just doesn't cut it.  450 xsi and sigma 18-200 OS is what I'm leaning towards. If the DR is near the 5D I'd be happy.
Marc

[marcmccalmont]

I have only one Xsi raw file, ISO 400; this is technically comparable to a G9, ISO 400 as well. The G9 is in the league of "prosumer" cameras, far from the 450D.

You have a 5D. If you upload an ISO 400 image with dark, uniformly lit, uniform color, smooth spots, I compare that with the 450D I have (the best is the same scenery with different exposures, so that there are some spots with illumination comparable to the 450D image).
[a href=\"index.php?act=findpost&pid=206025\"][{POST_SNAPBACK}][/a]

Can you post a scene from the 450 so I can try to shoot something similar?
Marc