Canon announces the 5D Mark IV (and two lenses)

[DavidEccleston]

It's more than that. In one of the videos they mention it works even at wide open apertures to cancel out the effects of the AA filter... so perhaps some sort of 5DS R effect is included. We won't know if, or how well, it performs until Bryan get's his hands on one.

The 1.74x crop on the 4K video has me confused. Why didn't they throw a few more pixels on the sensor, and make it do a 2x2 downsampling to do 4K w/ less noise. Perhaps they designed it expecting to downsample by 1.74x, but then didn't like the result, and in the end chose to crop it and use "pure pixels" instead. That's the only big flaw I see, from the specs.

[Busted Knuckles]

1st off, modern dslr cameras (even without a mirror) are massively capable kmaging devices.

I am not sure if it means anything that 3rd party "efs" lenses have new life in 4k land. Sigma 35-50 1.8 hmmm come to mind.

Put that on a slider and oof there is alot tool there for the vid folks.

3rd party efs types also live again in the crop of the sensor. Provides a whole new gamut of 7d2 and 5d4 debates for fps, reach, etc etc . If you can stand a rolling shutter and 4k frame grab. 30 fps?!?!? For kiddo football soccer etc etc. Pretty fun debate. Remeber something like 99.9% of omages are viewed on an 8 bit 1920x1080 screen. How many of those "exact moments" are now captured at 30 fps Oh giggle and snort

And isnt it great we can have it!!! Certainly when I got "back in" with the t3i I couldnt imagine we would be here.

[Kayaker72]

Yeah I've seen the 4K grabs put down a little bit. But to me it's an incredible feature. For example, now you can have both video and 8MP still shots of someone blowing out candles on a birthday cake.

[Busted Knuckles]

Yeah I've seen the 4K grabs put down a little bit. But to me it's an incredible feature. For example, now you can have both video and 8MP still shots of someone blowing out candles on a birthday cake.

I will likely rent one sometime and give 4k frame grab, 4k vid, and the diffraction correction (I am certainly thinking John is in the right direction on sharpening, but time will tell), as well as, testing the 35-50 1.8 as a vid lens (rental too).

I am certainly not all giddy about running out and upgrading - if I get GAS at all it is toward a 1dx1. I am much more likely to just shoot the heck out of what I have and wait until there is some sort of serious wear out and then see what is about.

The improved feature set is "handy" but I am quite pleased w/ the images the 5d3 kicks out. Perhaps when the 3d, physics defying, 18 stops of DR, completely ISO invariant, 30 FPS full res, global shutter 8k at 120 FPS vid, everything and the kitchen sink model comes out....

[Kayaker72]

Look forward to your impressions of the 5D4.

I checked this morning and my shutter count is right around 50k. Might need to take up time lapse photography so I can justify a new toy.

Until then I've decided to go through the owners manual, learn and try things that I haven't and keep my money in the bank. My "issues" with the 5D3 really do not amount to much. As you say, the output is great.

[peety3]

The 1.74x crop on the 4K video has me confused. Why didn't they throw a few more pixels on the sensor, and make it do a 2x2 downsampling to do 4K w/ less noise. Perhaps they designed it expecting to downsample by 1.74x, but then didn't like the result, and in the end chose to crop it and use "pure pixels" instead. That's the only big flaw I see, from the specs.

Why does anybody buy a 30mp camera to do 8mp video? OK, I get it that the older cameras didn't offer 4k so you need a newer model, but otherwise why do people ask "Does the 5DsR do video? How about the 1Dx2? etc."

[neuroanatomist]

It's more than that. In one of the videos they mention it works even at wide open apertures to cancel out the effects of the AA filter... so perhaps some sort of 5DS R effect is included. We won't know if, or how well, it performs until Bryan get's his hands on one.

What an AA filter does is blur the image by a fixed amount. The antidote for blur is sharpening. If you know what the amount of blur is, you can apply the appropriate amount of sharpening to correct it. RAW converters with camera-specific modules (LR, DxO, etc.) already do that for you.

When the 'AA-less' versions of cameras (D800, 5Ds) came out, everybody immediately compared them to the same shot taken with the AA version of the camera with no sharpening applied to either image. That's not the appropriate comparison, you should sharpen the AA'd image more, and in fact you *can* sharpen it more than the AA-less image before you start to see artifacts.

[Busted Knuckles]

Hmmm somemore pondering. Wonder if the magic lantern folks can step up the bit depth on 1920x1080. Not sure why thisbis an idea, perhaps closer to a RAW type recording?!?!?!.

Though I once thiught I would fiddle video, quickly realized that I dont have the skill nor patience for it.

[Joel Eade]

I sold my 7D MKII today and pre-ordered the 5D MKIV (body only)

[Kayaker72]

I sold my 7D MKII today and pre-ordered the 5D MKIV (body only)

Congrats. Can't wait to see your impressions.

Most of what I've seen so far is pretty encouraging. I mean, of course some people are jumping on initial results (some of which seem very dubious)and crying foul, but overall the 5DIV seems like exactly what I would expect from Canon, a well done, worth update to an already exceptionally popular camera. The early pictures I have seen are fabulous and even the early test charts at high ISO are encouraging that there might be a 1/3 - 1/2 stop improvement in noise even at high ISO. I've yet to see the 4-5 stop torture test that people seem to think should be a new standard.

But, in terms of early data, this is by far the thing that has caught my eye the most:
http://www.rawdigger.com/howtouse/Ca...xel-technology

Basically, the dual pixel technology, one row is under exposing by 1 stop. Also, they are observing that the entire 14 bit histogram is being used filled by both pixel row readouts and, for the first time, the full 16,383 (their number---14 bit--2¹⁴ = 16,384) readout potential is being used.

For years, I have heard people lament that 14 bit was a misnomer as the entire range was not being used in the actual readouts. Apparently, with the 5DIV, this may have changed and not only do we have full 14 bit readout, but, with two pixel rows offset by a stop, the potential for 15 bit range of data. Or, to quote the article directly "the dual-pixel raw file for this camera contains 15 bits of raw data, if you consider main and auxiliary subframes together."

Perhaps this will not lead to much, but it definitely has my attention.