5d Mark iii noise at 1600

[Dr. Benway]

Help!!!

What do you make of these - 5D MKIII @ 1600iso.

Does anyone have definitive settings for the lowest noise with the MKIII?

[Z]

Help!!!

What do you make of these - 5D MKIII @ 1600iso.

Um... that looks absolutely normal in the shadow areas of dark orange light. Give the camera a chance!

[Dr. Benway]

I get your point but I was expecting more from the 5D MKIII.

Here's some more - why the blockiness in the shadow areas?

[wickidwombat]

underexposed images are always going to show higher noise levels in shadow regions and this will be amplified when pushed

the noise is actually very clean and comes up very nice using something like topaz denoise where you can apply different degrees of NR to shadows highlights midtones and colour channels

[awinphoto]

Boy haven't we gotten spoiled... 4 years ago with my 30D 1600 was practically unusable and 800 was pushing the limits... in film 1600 was good for small prints only... now we can shoot 1600 and not think twice...  oh well. 

[bkorcel]

Looks great to me!  How large are you going to print these pictures?  16x20 or larger may not be possible.

I get your point but I was expecting more from the 5D MKIII.

Here's some more - why the blockiness in the shadow areas?

[swampler]

Looks great to me!  How large are you going to print these pictures?  16x20 or larger may not be possible.

I get your point but I was expecting more from the 5D MKIII.

Here's some more - why the blockiness in the shadow areas?

I got good 16 x 20 @ ISO 6400 from my 50D. 1600 on 5D is much better than that!

[Dr. Benway]

You can see the LR metadata.

The first shot - the dog - is that shadow noise acceptable?

The second shot is from a 5D Classic.

I will have some 8x10's printed tomorrow.

[bdunbar79]

I can't take this anymore.  You are underexposing badly.  In digital photography, if you underexpose, you can't really fix it without adding noise.  You can't just add light with LR retroactively; if you didn't collect it, it's not there.  Now let me help you out.  I'd like you to see some success.  You want to slightly overexpose in digital if you are going to be off in your metering.  +1/3 is best, but +2/3 and +1 are salvageable, as long as you don't have blown highlights.  For instance, in landscape photography I meter off the sky, set the shutter speed accordingly, then expose +2/3 typically.  If I overexpose slightly, that's not a problem in LR, underexposing is. 

Your porch photo is difficult because it's a high contrast situation, much like shooting in the woods on a hot sunny day.  In that case, meter off the brightest face, then exposure bracket, 5D Mark III will do a bunch of bracketing, I'd do -1/3, 0, +1/3, +2/3, +1.  Pick the best photo.  When you get better than shoot only on shot with no bracketing.

[briansquibb]

I can't take this anymore.  You are underexposing badly.  In digital photography, if you underexpose, you can't really fix it without adding noise.  You can't just add light with LR retroactively; if you didn't collect it, it's not there.  Now let me help you out.  I'd like you to see some success.  You want to slightly overexpose in digital if you are going to be off in your metering.  +1/3 is best, but +2/3 and +1 are salvageable, as long as you don't have blown highlights.  For instance, in landscape photography I meter off the sky, set the shutter speed accordingly, then expose +2/3 typically.  If I overexpose slightly, that's not a problem in LR, underexposing is. 

Your porch photo is difficult because it's a high contrast situation, much like shooting in the woods on a hot sunny day.  In that case, meter off the brightest face, then exposure bracket, 5D Mark III will do a bunch of bracketing, I'd do -1/3, 0, +1/3, +2/3, +1.  Pick the best photo.  When you get better than shoot only on shot with no bracketing.

I would go even further will the porch picture - large amounts of flash would sort the whole thing out, putting light on the important things (the people) and reducing the ambient highlights

... and reduce the iso to 100/200

[bdunbar79]

I can't take this anymore.  You are underexposing badly.  In digital photography, if you underexpose, you can't really fix it without adding noise.  You can't just add light with LR retroactively; if you didn't collect it, it's not there.  Now let me help you out.  I'd like you to see some success.  You want to slightly overexpose in digital if you are going to be off in your metering.  +1/3 is best, but +2/3 and +1 are salvageable, as long as you don't have blown highlights.  For instance, in landscape photography I meter off the sky, set the shutter speed accordingly, then expose +2/3 typically.  If I overexpose slightly, that's not a problem in LR, underexposing is. 

Your porch photo is difficult because it's a high contrast situation, much like shooting in the woods on a hot sunny day.  In that case, meter off the brightest face, then exposure bracket, 5D Mark III will do a bunch of bracketing, I'd do -1/3, 0, +1/3, +2/3, +1.  Pick the best photo.  When you get better than shoot only on shot with no bracketing.

I would go even further will the porch picture - large amounts of flash would sort the whole thing out, putting light on the important things (the people) and reducing the ambient highlights

... and reduce the iso to 100/200

Yes.  Good point.

[bdunbar79]

I did my own test at ISO 12,800 with no NR and 100 NR setting in Adobe Camera Raw. 

Photo 1 is no NR, photo 2 with NR.  5D Mark III, with 100-400L lens, 100mm, 1/200s, ISO 12,800, f/8.

[impresario7]

My first post on this forum.

I purchased my 5D Mark III in late May and after testing it for a few weeks using mostly L lenses, I thought I had a defective camera body because I was seeing noise at not only 1600 iso but even at lower ISO levels (starting faintly at 800 iso and gradually increasing). My camera retailer (Looking Glass Photo in Berkeley) decided to test the camera before they sent it back to Canon and when they didn't find any problems when making direct comparisons with a 5D Mark II (and saw less noise on the Mark III images at all iso levels, especially visible at higher iso levels), we began to talk about workflow and software. I have been using Lightroom 4.1 to process my files, which appears to be the culprit. Processing with Canon's proprietary Digital Photo Professional made a world of difference. The "haze" of noise (mostly luminance, but also some chroma) was completely gone and the colors were more saturated (perhaps more than I'd prefer, but in many ways, better than what I was seeing in Lightroom).  I wasn't certain about whether Lightroom uses Adobe Camera Raw or a different RAW processing algorithm, but the contrast between the screen view of RAW images is significant. I went back to look at those problematic files and re-processed them in Digital Photo Professional and was struck every time by the lack of noise using Canon's RAW processing engine.

If anyone can suggest how to obtain the same level of noise-free processing in Lightroom, please share your settings. (I hadn't made any changes to Lightroom between using my 5D Mark II and when I started using the Mark III body other than upgrading to 4.1, first 4.1 RC then 4.1 Final.)

Incidentally, I recently tried processing in DxO Optics Pro 7 Elite and the RAW images also look better than in Lightroom 4.1. I haven't been able to make comparisons between DxO and Digital Photo Professional, but I hope I'll be able to go back to using Lightroom because of the Develop module.

Thanks in advance for any feedback.

[wockawocka]

ALso, lightroom sharpens by 25 as a default for 5D3 files. Sharpening any noise in the process.

[SandyP]

Some of you people are HILARIOUS. Taking underexposed photos at high ISO and judging them at near 100% level for noise. Oh boy. SOUNDS FUN.

If you expose an image properly, some areas may indeed still have shadows, obviously. I shoot RAW, I load into LR4 with default settings in LR4, I don't do any extra noise reduction, it looks fantastic. Having shot with a 5D2 since the first few months that camera came out, I can say that some of you are insane. Get out there and do some serious shooting with this. You're just wasting your time doing these ridiculous "tests".