Canon EOS 5DS R Mark II Talk [CR2]

[rrcphoto]

Update 09/04/2017 @ 16:56 EST: Keep in mind that Canon uses APS-H sensors in some test bodies and this can affect pixel count.

If they're still at the APS-H test sensor stage, I doubt we'll see a 5DSII next year! Or are you implying that your source may have got their facts muddled?

canon tests their sensors at APS-H size because it's the largest most economical one they can do in small batches.

a full frame sensor is stitched with three exposures. APS-H takes only one.

 

[rrcphoto]

and like I said, canon's already been actively demo'ing the 120MP DSLR for the past two years and have actually stated "it's coming.. " it's actually in development now since sept 2016, as a 120MP DSLR. Again, that's not even a rumor, it's already been stated by canon. So what there's going to be two more different high resolution 5D's coming out? that seems unlikely.

and you're wrong on the lenses, you're into the world of oversampling. every lens will improve even the worst.

The lens quality issue won't be a problem if the oversampling ratio is 4:1; a 120MP image with 4:1 oversampling would yield a 30MP image, which is well within the resolving power of all Canon L glass. The problems would arise when attempting to use the full 120MP image resolution without oversampling. Only a handful of L series primes stopped down between F4 and F8 would be capable of corner to corner sharpness at that resolution. But as long as Canon provides the means to record both full-size 120MP RAW files and 30MP oversampled RAW files, then photographers get the best of both worlds. The 120MP may not give you corner-to-corner sharpness with any given lens, but it does give you the ability to crop in the center with a factor of 2 and still yield an image of 5D mk.4 quality. Meanwhile, an oversampled image of 30MP would blow away a 5D mk.4 in terms of IQ with greater DR, and fewer false color artifacts.

sRAW would be 30mp. however you're better off tailoring that yourself.

however yes, the oversampling effect would simply blow out any lower MP camera aka a Mark IV.

you're also correct about the crop factor. it's HUGE and I think people aren't getting this. this is the FIRST camera EVER that beats out any crop format for resolving power.

you can do a 2x crop factor on this and get better resolving than with any m43's camera body. (30mp to 20mp)

so you can use a 300mm lens like a 600mm lens like they do in the m43 world

in the right hands and with the right mindset, this will be a pretty awesome camera if it's this.

 

[rrcphoto]

Update 09/04/2017 @ 16:56 EST: Keep in mind that Canon uses APS-H sensors in some test bodies and this can affect pixel count.

yes/no. they would simply scale the masks for production. but that still wouldn't work out. they'd need around a 70MP APS-H to make that all work out to 120MP.

that would put it at 100MP though, which would be a nice number as well.

 

[unfocused]

Update 09/04/2017 @ 16:56 EST: Keep in mind that Canon uses APS-H sensors in some test bodies and this can affect pixel count.

If they're still at the APS-H test sensor stage, I doubt we'll see a 5DSII next year! Or are you implying that your source may have got their facts muddled?

canon tests their sensors at APS-H size because it's the largest most economical one they can do in small batches...

Okay, that makes some sense now. Although wondering where the 60 mp number came from. A 60 mp APS-H sensor would scale out to about 100 mp, which seems like a big leap. If they are testing a 35.5 mp APS-H sensor it would scale out to about 60 mp, full frame. One seems conservative and one seems too extreme.

Time will tell.

 

[Maiaibing]

I'd love a 120mpx camera, but I don't think there's enough demand for it, because it will be a pain to work with such huge files, it'll be slow and you'll have to downsample almost everything to make it worthwhile.

You are confused. There is no need to "downsample" anything - ever. "Downsampling" is not something any photographer "does" with his/her pictures and has nothing to do with how you view your results on screen or in print - its only - and exclusively - used as a technical path to compare different sensors, but has no meaning when viewing your resulting final images.

sure they do. I take 600mp stitched images all the time and downsample them for additional clarity and sharpness.

My sigma merrills are marvelous for doing this.

You're not downsampling anything. You're just stiching and printning (or viewing) at the size you prefer. Nothing different from what you would do with any other camera - no matter what pixel count.

no, I actually downsample.

I take the image at it's 600mp and when I'm ready to save for final printing, i change the resolution DOWN and choose an appropriate sampling algorithm. that's downsampling.

when you print, even when you don't downsample it happens during the print process anyways.

when you view the full image on your screen,it's significantly downsampled.

I give up...

 

[NancyP]

To tell the truth, I would rather have increased dynamic range and less noise than more pixels.

 

[AdjustedInCamera]

I hope the "Identical body to the EOS 5D Mark IV" part is not true. As a high resolution camera that is often used in live view on a tripod, an up-down tilting screen would be really useful.

+1

Personally find it strange that even reviewers highlight the tilty-screens as being good for getting those 'impossible otherwise shots'. Personally, just having the camera on a tripod, and being able to adjust the screen rather than myself is a massive win.

Also, I wouldn't know how to shoot video at any resolution unless I could hold the camera securely which seems, again, to require the screen to be adjusted so I can actually see what it's filming.

 

[vscd]

To tell the truth, I would rather have increased dynamic range and less noise than more pixels.

Then you should switch your 5DSR with a 5D Mark IV

 

[unfocused]

To tell the truth, I would rather have increased dynamic range and less noise than more pixels.

Well, yes, but that's what the rest of Canon's lineup is for. The 5Ds series is intended for maximum resolution. If you want low noise and high ISO there are other options.

On the other hand, as the 5DIV showed, the old relationship between noise and resolution isn't quite as clear these days as it once was (The 5DIV compares very favorably to the 1DXII despite the higher pixel count).

 

[unfocused]

I hope the "Identical body to the EOS 5D Mark IV" part is not true. As a high resolution camera that is often used in live view on a tripod, an up-down tilting screen would be really useful.

+1

The 5Ds needs the economy of scale that comes with using the 5D body in order to remain competitive. I don't see Canon making any changes in the body that would require a different design and thus, higher costs.

 

[Sharlin]

no, I actually downsample.

I take the image at it's 600mp and when I'm ready to save for final printing, i change the resolution DOWN and choose an appropriate sampling algorithm. that's downsampling.

when you print, even when you don't downsample it happens during the print process anyways.

when you view the full image on your screen,it's significantly downsampled.

I give up...

Well, good, because you seem to have no idea what downsampling is. Downsampling means reducing the sample rate of a signal. When the signal is a digital image, your samples are pixels, and downsampling means reducing the resolution of the image. That is, resizing the image to a smaller number of horizontal and vertical pixels. Every time you export a photo in a lower resolution, your software downsamples the photo. Every time you display a high-res photo on a low-res screen, your image viewer downsamples the photo. And so on.

 

[rrcphoto]

no, I actually downsample.

I take the image at it's 600mp and when I'm ready to save for final printing, i change the resolution DOWN and choose an appropriate sampling algorithm. that's downsampling.

when you print, even when you don't downsample it happens during the print process anyways.

when you view the full image on your screen,it's significantly downsampled.

I give up...

Well, good, because you seem to have no idea what downsampling is. Downsampling means reducing the sample rate of a signal. When the signal is a digital image, your samples are pixels, and downsampling means reducing the resolution of the image. That is, resizing the image to a smaller number of horizontal and vertical pixels. Every time you export a photo in a lower resolution, your software downsamples the photo. Every time you display a high-res photo on a low-res screen, your image viewer downsamples the photo. And so on.

i think i get it..

you can call it downsampling (which makes more sense to me at least) or resampling, i think that's his beef. downsampling i guess is supposed to involve something like the AA filter versus changing the image resolution. however in reality, you're doing both.

there. i think i got the terms right this time.. lol.

 

[Sharlin]

i think i get it..

you can call it downsampling (which makes more sense to me at least) or resampling, i think that's his beef. downsampling i guess is supposed to involve something like the AA filter versus changing the image resolution. however in reality, you're doing both.

there. i think i got the terms right this time.. lol.

As far as I can see, downsampling is resampling "down", ie. to a lower sample rate. Upsampling is resampling to a higher sample rate. In both cases you already have a discretized signal, a bunch of samples.

 

[bholliman]

I'm very happy with my 5DsR. My wishlist of improvements for the MkII is pretty short:
[list type=decimal]
[*]tilt touch-screen LCD
[*]f/8 autofocus, similar to the 5DIV
[*]DPAF
[*]on-chip ADC
[/list]

Personally, I hope they don't increase the megapixel count too much (60mp or less), but I'm sure it will get a mp bump since that seems to come with every new iteration.

 

[tron]

Sharpening an image that has been through an AA filter will not restore detail. It will create detail that may, or may not, look similar to the detail that was lost. There's no magic way to get that detail back. Once it has been defocused you're reliant on algorithms that insert artificial sharpness. They do seem to work most of the time, but especially for monochrome work there's no substitute for having a sensor without an AA filter.

Having spent a good decade or so of my life writing image processing algorithms including multiple sharpening/unsharp mask methods I do know a thing or two about this.

In addition, I've shot tens of thousands of images on my 5DSR, landscape, portrait, macro, etc, etc. And I can't say I have found a single image that has been ruined by moire. Maybe I'm just lucky.

You are lucky. My best bee-eater (the ones that were the closest to me) shots had moire. The same with glossy ibises. On the positive side it blends with the feathers and non-photo related people do not notice it. But I know about it and I notice it. It made me think that 7DII may still have its purpose (I had started using 5DsR for birding).

 

[Act444]

FWIW, I've gotten moire before on bird feathers...................................using a 5D3.

I have not used the 5DSR in nature at this point, but elsewhere I haven't had any issues with moire using that camera. I suppose that if you get it with the R, you'll also get it on a filtered camera like the 5D4 (although granted, the R will show it worse - sometimes significantly so).

 

[Talys]

I've had moiré before on fabrics (for example, a navy sweater). It's super annoying and not really correctable in post, at least not in a non-destructive manner. It has happened maybe two or three times in thousands of shots; I can't recall the body.

The fix is easy, if you can re-take it, of course -- just fiddle with the light a bit, or adjust the distance or angle in a minute way, and voila.

One thing I've noticed: there are times when it appears in the JPEG, but not in the RAW file, but if it's in the RAW, it will always also be visible in the JPEG. So if you're splitting the files (storing JPEG on laptop and RAW on camera, to speed up wifi), you only need to check the RAW file if the JPEG has moiré.

 

[privatebydesign]

I've had moiré before on fabrics (for example, a navy sweater). It's super annoying and not really correctable in post, at least not in a non-destructive manner. It has happened maybe two or three times in thousands of shots; I can't recall the body.

The fix is easy, if you can re-take it, of course -- just fiddle with the light a bit, or adjust the distance or angle in a minute way, and voila.

One thing I've noticed: there are times when it appears in the JPEG, but not in the RAW file, but if it's in the RAW, it will always also be visible in the JPEG. So if you're splitting the files (storing JPEG on laptop and RAW on camera, to speed up wifi), you only need to check the RAW file if the JPEG has moiré.

Often times it is non destructively correctable, though not always.

 

[willhuff.net]

Any guesses about dynamic range improvements with this camera?

 

[RTISTIC]

With all this EOS 5DS R MK ll talk. Canon will now have to come out with a new 5DS model to compete against
Nikons D850 in price and features. I own a 5D Mark IV even though I am happy with it. I can't help but feel slighted
you get so much more at the same price point. WHAT IS CANON DOING WRONG?