6D & 5D3 Raw images @ 12,800

Talley wrote in post #16784639
One thing moving from 7D to the 6D was disappointment in the moire'. Looked cheap.

When I shoot video on my 7D or 6D, I use a custom function with f/22 as the f-stop. The shallow DOF promised with DSLR video comes at a great aliasing cost, with sharp lenses. You're better off with a dull, fast lens for video, if you want shallow DOF and low ISOs for line-skipping cameras like these.

Talley wrote in post #16784651
I never seen this option. Looks like the 70D has a nice improvement over earlier 18MP images.

Relative lack of banding noise. This doesn't show up in DxOMark tests, but shows up when you actually look at the images!

I'm just not sure if the 5D3 would give me such a clean output image when I absolutely NEED it.

1.4 1/200 & 25,600 ISO and had NO issues with the SIgma focusing.

I Honestly don't feel the 5D3 could give me such quality in this light.

gabebalazs wrote in post #16784672
This topic actually made me do test videos with my 6D and 70D to see which one has more moire (interesting alliteration )

I think the 70D is a tad better in this test, the 6D is a bit sharper but has more moire.

It makes me wonder if that's something Canon could (but do they want to?) fix via firmware. I mean the 5DIII does not seem to have any stronger AA filter, so I suspect the difference in aliasing and moire are a result of different processing and not something physical inside the camera.
Maybe you guys already know why this is...

So perhaps Canon could help us out with a firmware.

It's a bandwidth issue, more than anything. Reading a full sensor of pixels 60 times a second requires a lot of bandwidth, so cameras with less bandwidth are forced to use line-skipping. Aliasing of panned shots or moving clothing prints could be reduced by reading 1/3 of the lines in a pseudo-random selection, but that would also make still scenes look "busy" at the pixel level, reading something different for each output pixel in successive frames.

Interlace is a possibility, but interlace works best when the display device plays the lines back in the same way they are recorded; that's why some video in the old days looked much more realistic than others; the camera scanned the scene in the same way it was displayed, suppressing potential artifacts. Many years ago I created an animation on my Commodore Amiga that had a ball bouncing on one half of the display, not taking the interlaced display into account (30 frames per second), and the other side, the mirror image of the bouncing ball but drawn at the timing for the interlace fields (60 fields per second). The difference was amazing. The Amiga synced everything to the video, though, so that was the reason the interlace worked well. There are very few things as ugly, though, as interlaced video being played back asynchronously or converted to non-interlaced video, or a different frame rate. IMO, everything should be played back with the timing at which it was recorded, ideally. I hope that everyone in the media industries are saving the originals, because someday our displays will be capable of any arbitrary timing. I could see a computer desktop refreshing asynchronously, depending on content change, and video windows being drawn at 24 fps, 50 fps, and 30i at the same time. I am very sensitive to video jitter, and so much of what I see is distracting.

Well after some looking through I can get close to matching the 5D3 25,600 to the 6D.

In Lightroom it takes a -24 on blacks and a +12 in NR and it's pretty close to the SOOC 6D except for the bottom right extreme corner purple noise gain. Remains a splotch and it takes an adjustment brush over that particular corner of -38 on shadow and additional noise reduction of 25 over this spot.

I can clean it up to match so basically it means on files over 12,800 on the 5D3 it takes just a tad of adjusments which I can set lightroom up to perform automatically upon import.

Talley wrote in post #16784690
the 70D has that feature too? I loved that on the T3i. 3x zoom looks great.

I had no idea about this, until this thread. Video suffers a lot less from "digital zoom" than stills do; the brain gathers detail from video over successive frames with slight changes in subject/pixel registration.

I accidentally shot an HD video of a baby Bald Eagle preening itself with my 100-400 on a tripod with a 2x converter and my Pentax Q (I normally take the TC off of the lens with the Q for stills), and I didn't notice it was still there until afterward. The video came out surprisingly well, considering the fact that the lens is made for much larger pixels or film grain, the tiny sensor, and the 2x III TC. I think the lens' aperture may have been wide open, as well, as I don't always remember to stop down the lens in DOF preview mode while changing cameras.

Buylongterm wrote in post #16784692
Let me get this straight. So, after all of your recent posts about switching to the 5DIII your now having second thoughts?? Please say it ain't so. I think you need to spend more time taking pictures and less time obsessing over things. (but I still love your posts)

If shadow and high ISO noise is your main concern, or saving US$1000, the 6D is the camera. If shooting speed and AF speed is most important, then the 5D3 is the camera.

When I bought my 6D, I intended to use it for nightscapes, museums, indoor family gatherings, etc, and I would use the 7D for bird photography and other action. I found myself, after a while, only using the 7D in bright light, and using the 6D in low light more for birds, because even with a 1.4x TC, which puts the subject over a similar number of pixels on the 6D as the bare 7D does, once the ISO gets above 800 for the 7D (1600 on the 6D with the TC), the noise difference becomes quite substantial. That made me wonder if I shouldn't have bought the 5D3 instead, for the faster shooting. There is a lot to like about the 6D, though, with its lack of banding noise, and I don't know how the 5D3 compares in mirror shock, but the 6D has the lowest mirror-shock of any Canon DSLR I've ever used, even without the "Quiet Shutter" settings. The lens IS has the subject in exactly the same position in the frame before and after exposure blackout.

TeamSpeed wrote in post #16784906
If a 7D2 comes out, the 7D will be the longest running single model next to the 1Ds3, if you make the assumption that the 1DX is its replacement.

I know Canon sees it that way, but I certainly don't. You can't replace a 1.3x camera with a FF camera with bigger pixels, as you lose subject resolution. Sensor sizes serve a purpose combined with pixel density, and bigger is not always better. I find Canon's emphasis on shifting to bigger sensors VERY disturbing. I NEED a fast-shooting, 3x-crop 18MP+ camera with PDAF at f/5.6. TCs are only a stopgap, and ruin AF.

palad1n wrote in post #16784913
I would say that most people buy 5dmkIII for AF + high ISO performance combo primarily... noise differences are negligible.

Simply there are no major obstacles 5dmkIII couldn´t handle thanks to superior AF system and almost as good low capabilities as 6D has.

If you need to pull up ISO 100 shadows, or shoot at ISO 512K, the 6D edge is there. As is the low-light AF; what good is fast AF if it goes to sleep when it gets dark?

Let's be real; each camera has some significant benefits, and neither is a subset of the other, in capabilities.

JBlake wrote in post #16785955
And next, for kicks, compare the Nikon D800 or the D610 to the 6D in this setup; the Nikon's look even worse than the 5D III does. Maybe I am missing something here, but, where is the superior Nikon dynamic range advantage?

Most people talking about the DR of a camera are talking about base ISO, not high ISO. Sure, you measure DR at high ISOs, but DR at high ISOs is moot, because at high ISOs, you can "under-expose" at a lower high ISO and get more DR, in the form of highlight headroom relative to your intended gray.

DxOMark-like DR graphs throughout the range of ISOs are really graphs of read noise (only as a quantity, and not taking character into account), from a practical perspective. As you get to higher ISOs, the range of read noises across sensors of the same size have a smaller range than the read noise at base ISO.

mgk2 wrote in post #16786089
I dunno about some of you guys, but how often do you shoot at ISO12800?

Even if you do, those will only ever gonna end up as snapshot or small print, does the lil difference in shadow which btw only visible when pixel peeping matters in reality?

Some people just love picking bones from an egg....

Most images are never viewed large, even if they look good large. We don't have room in our environments for large prints, or 60" monitors all over the place.

Lots of subject matter is only available in low light, and/or using a slow lens and/or fast shutter speeds. Shooting birds in migration after the trees foliate, on an overcast day, I often have to crop images underexposed by two stops at the 6D's top auto-ISO of 25,600. At the size that a bird photo displays on a web page or in a book, such images can be quite usable. It all depends on how unique or rare a capture is. Would you rather get a shot of bigfoot at ISO 204,800, or a picture of a barn and silo at ISO 100?

Do you realize that ISO 800 color film of not long ago was noisier than ISO 12800 on current FF cameras?

mgk2 wrote in post #16786089
I dunno about some of you guys, but how often do you shoot at ISO12800?

Even if you do, those will only ever gonna end up as snapshot or small print, does the lil difference in shadow which btw only visible when pixel peeping matters in reality?

Actually, if the difference between two cameras is not mainly the quantity of noise, but the character of the noise, thenm the difference in noise performances actually increases with smaller image sizes. 1-dimensional (banding) and low-frequency source noises reduce in visibility at a much lower rate than 2-dimensional random noise; in fact, reduction can make banding visible that was not visible at the original resolution in extreme cases.

Here are your 5D3 vs 6D at ISO 12800 next to each other, using some leveling, then changing brightness/contrast. They both have it, but the 5D3 is definitely worse.

TeamSpeed wrote in post #16797049
Here are your 5D3 vs 6D at ISO 12800 next to each other, using some leveling, then changing brightness/contrast. They both have it, but the 5D3 is definitely worse.

Ya first weekend out with the 5D3 I noticed the 6D higher ISO (6400+) had blacker blacks and smoother color and less shadow noise.

I'll just deal with it though. So far the AF has been stellar and I'm very happy with the 5D3. And Ironically I never shoot 1.4 glass in the day except for recently so the 8000 shutter has been used a bit.

Talley wrote in post #16797238
And Ironically I never shoot 1.4 glass in the day except for recently so the 8000 shutter has been used a bit.

Try dropping the ISO to 50 Talley....