ISO performance XTi/400D Vs 50D (or 40D)

[hector_evans]

I am considering upgrading from the XTi/400D to 50D. The only problem I have with my XTi is ISO performance, as I

like to shoot indoors without a flash. Should I expect at least a 2-stop improvment with the 50D? Of course, there is

also the 40D. How much has ISO performance improved between 40D and 50D anyway? I read that there hasn't been

much improvement there.

 

Thanks.

[nhut-nguyen]

You need to wait to see more reviews, like the ones from www.dcresource.com, www.dpreview.com, or www.imaging-resource.com. IMO they give honest reviews, normal owners tend to be more bias. These review sites also have sample images done with different levels of ISO, you can always compare the samples of the XTi to the 50D.

[PuppyDigs]

At pixel level, noise performance of the 40D and 50D are about the same at ISO 1600 and 3200, although the 50D is more

prone to banding in shadows (don't bother with H1 or H2). But what idiot enjoys their images at pixel level? The greater pixel

density of the 50D needs less rezzing than the 40D or XTi for a given print size, so the noise it does have is less apparent. So the fact

there isn't any improvement is a moot point because your prints will look better than the 40D or XTi at high ISO. So, yeah, a 11x14 50D

print will look a little less noisy simple

because you won't need to "magnify" the pixels as much.

[matthijs]

In my unprofessional opinion the image quality in the 400D's 800 ISO is comparable with the 50D's 6400 ISO.

 

 

The same goes for 400D 400 ISO vs the 50D's 1600 to 3200.

And the 400D's 1600 vs 50D's 12800.

 

 

Gain is two to three stops from what I've seen. (pbase.com has some nice 50D pictures including some high ISO tests.)

 

 

Kind regards, Matthijs. (having a 400D and looking forward to a 50D hpefully somewhere in 2009)

[g dan mitchell]

You are very unlikely to see anything like a 2 stop improvement in low light performance from upgrading from the XTi. More

likely, the performance will actually be quite similar.

[kari v]

400D's ISO 800 is very clean while 50D's ISO 6400 looks like a mess of banding in the shadows. 400D's ISO 1600 is perfectly usable with correct exposure while 50D's ISO 12800 is... erm...

 

Like this:

http://www.pbase.com/dtuason/image/104606291

 

Some 50D H1 samples are somewhat ok but a rather well lit scene, ISO 6400 and lens stopped to f11 or so is not my idea of actual high ISO use.

 

For my intended use 400D --> 50D goes pretty much like Puppy Face said + in-body dial for higher ISO, which is nice. Of course 50D is also a sturdier body with more controls, better AF and frame rate. These alone make a difference in low light success rate.

 

Two to three stop difference? I'd say no. One stop? Yeah, you can dial in ISO 3200 and it looks ok.

 

How big do you print Hector?

[kari v]

I forgot bigger viewfinder.

 

Basically if you want really much better high ISO you should look into Nikon D700 which actually looks rather

usable at ISO 6400. And I'm sure 5D mkII will do well too with extra megapixels to boot. Pricey? Yes, sir.

[hector_evans]

Thanks for your replies!

 

Kari, I usually print at 12 x 18 in. (30 x 45 cm)

 

When you say the 50D has better AF, does this mean faster or more accurate?

[torin_olsen]

Hector, I've found my 50D to be much faster focusing with my 70-200 f/2.8 (at all apertures) than my previous 20D. Not so much with the 100mm macro. The center AF point has more sensitivity for f/2.8 than previous models.

 

"400D's ISO 800 is very clean while 50D's ISO 6400 looks like a mess of banding in the shadows. 400D's ISO 1600 is perfectly usable with correct exposure while 50D's ISO 12800 is... erm..."

 

Well, if you underexpose the image of course you'll get banding, etc. Comparing "correct exposure" ISO 1600 to underexposed ISO 12800 is no comparison at all... Here is an image I took at ISO 12800 (single incandescent light bulb as the source) while just messing around with my 50D. I did apply some very slight noise reduction, but that's about it. I'd hazard to say ISO 12800 is perfectly usable with the correct exposure. =D Check the exif data.<div></div>

[tdigi]

If your only concern is ISO performance consider the lenses your using.

[gary_anthes]

Torin makes an excellent point that seems overlooked/confused sometimes. It's one thing to look at a camera's performance at high ISO when properly exposed, it's something else (but maybe equally important) to consider how forgiving it is when underexposed. Of course, an image with high contrast is likely to be "underexposed" for the deep shadows even when it is correctly overexposed overall. Therefore, the image Torin posted may not be a good test as it has no deep shadow areas.

[gary_anthes]

Oops, I meant "...correctly exposed overall."

[jbauer]

I would agree with Tommy's post above, if shooting in low light is your aim, cranking up the ISO would be my distant second choice to simply having some fast glass with large apertures attached to my camera body. Are you shooting with f/2.8 or better lenses on your current XTi?

[kari v]

"Of course, an image with high contrast is likely to be "underexposed" for the deep shadows even when it is correctly overexposed overall. Therefore, the image Torin posted may not be a good test as it has no deep shadow areas.

 

True. My low light shooting happens mostly in clubs and such where exposure is tricky and even lighting is unheard of. Even ISO 1600 b&w film (35mm) shows surprisingly little grain in good overall light.

 

By "correct exposure" in my previous post I meant correctly exposed main subject. Everything else will of course suffer.

 

Hector: In general xxD cameras feel more responsive and accurate than xxxD models. No matter where you end up with this ISO debate I think you'll be very happy with 50D.

 

30x45cm isn't huge... sure, noise will be visible if you look closely but is it (ISO 1600) really that bad from one meter or so?

[torin_olsen]

Ya, shadows are problematic at higher ISO's. I was just making a point. =D I did my own (very

unscientific) tests after getting the camera at high ISO's because I shoot a lot of performance

photography.

<br>

<br>

I tried to make it as difficult as possible... Underlit, high shadow areas. Judicious use of noise reduction

can help too. Album comments give an explanation of what you're seeing. Each image also has exif

data still intact which is visible with the "more info" on the right side when viewing an image.

<br>

<br>

<A HREF="http://picasaweb.google.com/TorinOlsen/50DUnscientificHighISODemo?authkey=VV6i3G9-

xaU#">Unscientific 50D ISO test</A>

<br>

<br>

The test was done solely to satisfy my curiosity. If some people find it interesting, great!

[david_john_appleton]

IF you wont a real noise up grade what about a second hand 5D i paid 700 pounds for mine from eBay... no arguments please.... the 5D has the best noise and IQ Evan at 100 iso AT THIS PRICE on the planet. i know its old and it wont play god save the queen it just a fantastic image maker . Still use my 20d for telephoto work or when i need faster frame rate hope this helps Dave