Sporgon said:
BeenThere said:
ahsanford said:
And of all the things this market segment of camera needs to hype, 8K timelapse isn't it.
The D810 needs to fuse its own brand line's legacy -- that mythically good sensor -- with the tech overload of the D5/D500 releases.
They 100% can do the latter (4K, immense buffer, anti-flicker, NFC, illuminated keys, etc.). But the former was Sony EXMOR and they (to my knowledge) have never made a sensor that good in-house. It remains to be seen if they can do that themselves or if they have to concede defeat there and buy a 2+ year old A7R II sensor to do it.
- A
I would love to have the A7r2 sensor in a Canon.
I wonder how much of this desire for the Sony sensor in a Canon camera comes from the crushed blacks in Canon CR2 files, something that is actually revisable.
Already fixed in an ACR update many many months ago.
Sporgon said:
In another thread Privatebydesign referred to the sliding black point on a CR2 file as opposed to the fixed point on a NEF.
For years this has no longer been true. Nikons, Sonys all also have an offset which preserves noise levels around the black (SNR=1) floor. Canons just have a higher one, which hasn't really helped them (since it's not added at the analog level).
Sporgon said:
i don't quite know what this means, but I do know that compared with Sony and Nikon raw files the blacks on the Canon are crushed down on conversion and generally require opening out. You then get a tonal quality similar to the other makes. Also to optimise IQ in the Canon it requires a moderate ETTR, even on the 5DIV it seems from my fiddling. If sample images are shot to produce an optimised OOC jpeg image, with no further PP, then the Canon is immediately thrown into a disadvantage compared with Nikon and Sony. As virtually all on line images are shot this way, attempts to compare say D810 files with a 5DIV result in the Canon appearing to be (slightly) inferior. Obviously there will be a greater gap with older Canon cameras. Also it looks to me that despite the a7RII using the BSI Exmor R sensor, Nikon's use of the older 36 mp Exmor sensor makes it the better one.
That's now how we do our DR shootouts. We ETTR our tests, ETTR'ing both the Canon and the Nikon and/or Sony. It wouldn't be fair to ETTR one and not the other, would it?
Your last point about the Nikon's older 36MP Exmor sensor being better - in what way? In base ISO DR, yes, but only because of ISO 64 mode, which effectively extends pixel capacity so the camera can capture more total light. But the a7R II's BSI + dual-gain design allows it to achieve better high ISO performance.
Sporgon said:
Just why the CR2 files have this crushed black on conversion I have no idea. It is not a result of lower DR, and using crushed blacks to hide the poorer bottom end. Going back to the early days of digital by Nikon D200 was much better in dark tones than the 5D, but if the charts from the likes of Photon to Photos are to believed, it had much less DR. The 5D was harsh in comparison and needed the black point lifting in post, as have all the ones since.
If Bill's measurements show the D200 to have lower base ISO DR than the 5D, the D200 probably had lower base ISO DR than the 5D. That's how much I'd trust his data (more so than DXO, actually).
Sporgon said:
This harshness in the dark tones, if uncorrected, does make it appear as if the Canon has less DR than it actually has, and is why (I presume) under exposing on the Canon is pretty disastrous if you intend to lift shadows later, whereas on the Nikon, Sony and Pentax (the ones I have worked files from) it doesn't really matter.
This doesn't change any dynamic range measurements. Canons tend to have more noise in their images, which affects shadows more than bright tones. That's why their shadows aren't as clean, as that's why their measured low ISO DR is lower than peers (though they've caught up quite a bit).
Sporgon said:
As long as I'm not outside the DR I can make even my 5DII images look like those from a D810.
? Well, sure, when the DR of the scene doesn't extend the DR capabilities of your camera, it's total light capture that matters, which ends up being similar between FF cameras, save for the D810 at ISO 64 where SNR rivals some medium format cameras. Whether you'd notice that or not is another matter though.