30 D NOise?? Used??

[mpphoto12]

s it worth buying a 30 D used to upgrade from an XT? how is the noise comparisoion more importantly? thanks

[IAMB]

I picked up a 30D to upgrade from my XT right after the 40D came out in order to take advantage of lower pricing, so it wasn't used, but the noise and focus accuracy difference was very noticeable (especially when shooting indoor sports like basketball). For the price you can snag a 30D at right now if you're looking to move up from your XT, you can't really go wrong.

[Daniel Browning]

The 30D and Rebel XT have almost the exact same sensor. Most measurements show the noise to be the exact same as well (3.6 electrons per pixel). The 30D does have a slightly higher full well (51,400 photoelectrons vs. 43,000), but results in less than a third-stop difference in dynamic range.

[Colin]

OTOH, the feel, build, and frame rate of the 30D are significantly better than the XT.


I've been letting my friend borrow the XT for over a year now, i think, since I got the 30D.

[Jon Ruyle]

30D does have a slightly higher full well (51,400 photoelectrons vs. 43,000)


Does this mean each pixel in the 30D can count 51,400 electrons? That is way more than I would have guessed... but I never knew how to find out. (How the heck do you know all of this? [:)])

[Daniel Browning]

Does this mean each pixel in the 30D can count 51,400 electrons?


Yep. Photons convert to electrons which are converted to ADU.






(How the heck do you know all of this? /emoticons/emotion-1.gif)
<div style="clear: both;"]</div>





Same way I learned anything: I read it on the Internet, so it must be true! [;)]


Actually, that's only half of it: after I read it on the Internet, I test it for myself to see if it's true. In the case of the 30D, I happen to own the 20D, which is the same sensor (even more so than the XT), so I've been able to measure the read noise myself.


I've always had an interest in the technical aspect of cameras, so I've been reading things on the web about sensors then trying it myself to see if I can repeat the experiment and get the same results. Emil Martinec's paper on Noise, Dynamic Range, and Bit Depth ("http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/) has probably been the single most enlightening treatment of the subject.

[HiFiGuy1]

Thanks, Daniel. I bookmarked your link to read later.

[Jon Ruyle]

Very interesting, Daniel. I've read about half of it now.


I always thought photon noise (and thus, what they call "unit gain") was all that mattered, and that read noise was negligible. According to the data presented, however, read noise matters quite a bit, and varies more from camera to camera than "unit gain" (which determines photon noise, it seems).

[Daniel Browning]

Yes, it is interesting. Cameras have been pretty close to the theoretical maximum sensitivity (QE) for a while, so in the last 5 years sit has only improved by one stop or so, and some of that has just been due to allowing suboptimal color separation. The big advancements in low light photography have been and will continue to be in the reduction of read noise. (Larger sensors and wider apertures could also be counted in the same breath, I guess.)

[mpphoto12]

So does anyone know how the noise is compared to an XT? is it worth it used? how aboyt new?

[Jon Ruyle]

Hmm? What? Were we talking about an XT? [:)]


I dunno if this site can be trusted, but they list tons of stats:


http://www.dxomark.com/index.php/eng/Image-Quality-Database/Canon ("http://www.dxomark.com/index.php/eng/Image-Quality-Database/Canon)


The 30D has a low-light iso score of 736. The XT has a score of 637. This number is the highest ISO possible before the signal noise ratio goes above some threshhold. (So, ISO 736 on the D30 is like ISO 637 on the rebel). That's a fraction of a stop. I doubt anyone would notice the difference in normal use.


These guys also have an "overall stat" which I don't trust at all because I couldn't find out what it meant. But it is supposed to take into account noise, dynamic range, color depth, and resolution and who knows what else. On this stat, the XT actually scores ever so slightly higher than the 30D (59.9 to 59.5... they say 15 points is a single stop, whatever that means when you're taking all that stuff together into account, but anyway, no difference at all in practice).


Of course the 30D has other advantages over the rebel (if you care about low light, you might care that the 30D can focus in lower light), but I wouldn't switch for low iso performance.