Highlight Roll-Off: GH5, GH5s, A7III and X-H1 and ???

[ajay]

12 minutes ago, jonpais said:

Apparently, dynamic range is directly related to bit depth.

Is it? That's not what he states in his video:

 

[Robert Collins]

7 minutes ago, jonpais said:

Apparently, dynamic range is directly related to bit depth.

Yes, I think that is self evident. Take a shot jpeg (8 bit) and raw (14 bit) and you can recover 1 to 2 more stops in raw. Also bit rate will effect DR - I can recover more highlights in a jpeg still than in the more heavily compressed video. These factors would certainly help both the GH5 and X-H1 get closer to A7iii DR in video relative to stills than the sensor size difference might imply.

[Don Kotlos]

With linear encoding like in RAW files, the bit depth determines the upper limit of dynamic range.  jpeg or log use non linear encoding which can map higher dynamic range in a limited bit depth. Thats how jpegs can typically have ~11 stops of dynamic range and log profiles >12. 

The problem is that the more you compress the dynamic range of an image into limited bit death, the more you under-sample the color space and the more limited tonalities & increased banding you get. 

Bitrate has nothing to do with dynamic range. 

[jonpais]

9 hours ago, ajay said:

I found this video comparing the GH5 and A7III with regards to dynamic range. It appears the A7III has at least 1.5 stops more DR compared to the GH5. From what I've seen the GH5s might have about .5 stops better than the GH5 but still puts the DR behind the A7III by quite a bit. Of course there are other criteria that makes the GH5 line more appealing in other circumstances, but for myself, shooting outdoors, it appears the A7III might be a better choice for what I do:

 

Am I just tired, or is there a serious flaw with this test? Correct me if I’m wrong, but I don’t think he should be raising the ISO, he should be opening up the iris. Very strange test. 

[Emanuel]

Both methods apply :-)

[jonpais]

As you raise the ISO, dynamic range is significantly decreased, whereas the aperture setting has no influence on DR.

[Mark Romero 2]

Just Curios:

In RAW, according to DPreview, the a7 III is supposed to be (more-or-less) ISO-less, meaning shooting underexposed ISO 100 and pushing one stop in post doesn't give you a noise penalty when compared to shooting at ISO 200.

I wonder if this holds true for the movie codec as well?

I know there is probably a flaw with this thinking. Why have slog when you could just underexpose using a more linear gamma and push it in post since the noise will equal out?

Probably something to do with the dual gain circuits... Or the movie codec just doesn't act like RAW in this way.

[Emanuel]

5 minutes ago, jonpais said:

As you raise the ISO, dynamic range is significantly decreased, whereas the aperture setting has no influence on DR.

Ah... on that one, yes indeed. Bit rate can perform a play part, obviously. That one is pretty inaccurate from the reviewer side. For some reason GH5S dual ISO is paramount.

Just read the whole thread now :-)

[jonpais]

Which other side is there? ?

[Emanuel]

1 hour ago, jonpais said:

So the test is inaccurate?

Did not follow the YT video... Only the comments of you all now : ) Obviously DR is not the same raising the ISO bar. What to say then? Discredit, to say the least, right?

1 hour ago, jonpais said:

Which other side is there? ?

The other side is the reader's side. Pretty valuable BTW ; ) And if you edit your post without giving notice to the other side, well, that's unfair game :D my dear fellow ;-)

[jonpais]

But I was unmistakably referring to the video, not to the preceeding comments.

12 minutes ago, Emanuel said:

Ah... on that one, yes indeed. Bit rate can perform a play part, obviously. That one is pretty inaccurate from the reviewer side. For some reason GH5S dual ISO is paramount.

Just read the whole thread now :-)

We were talking about bit depth, not bit rate...

[Emanuel]

41 minutes ago, jonpais said:

But I was unmistakably referring to the video, not to the preceeding comments.

Yes, got you now : ) This is what happens when we (myself) lose scope. To be out of context leads to this lack of perspective from my side as your reader this time. You wrote you were tired, I misread that : )) Thanks for correcting my entry anyways :-)

41 minutes ago, jonpais said:

We were talking about bit depth, not bit rate...

Both, though natures of a different kind obviously, they are not so unrelated as apparently people tend to think in one way or another...

[mercer]

48 minutes ago, jonpais said:

Am I just tired, or is there a serious flaw with this test? Correct me if I’m wrong, but I don’t think he should be raising the ISO, he should be opening up the iris. Very strange test. 

I think you may be tired. As you said, it’s a dynamic range test... open the iris, close the iris, the DR won’t change at the same ISO at different apertures... only lens sharpness will. So to test DR, you need to change the ISO on both cameras to see which does better.

[Emanuel]

Glenn, ISO will distort the results when directly affects DR per se. Tired or not, he's right on that one : ) hey, I am tired too... it is 3:00 AM over here, Sunday night or Monday morning :)) Same ISO on both devices should be a different story, of course :-)

[mercer]

That’s the point of the test. You test how well the DR holds the further you are from base ISO.

[jonpais]

4 minutes ago, mercer said:

That’s the point of the test. You test how well the DR holds the further you are from base ISO.

Sorry, I must have missed the point of the test. I’m going back to bed now. 

[Emanuel]

@mercer Don't state that too loud to Fuji though! :D

[Robert Collins]

6 minutes ago, mercer said:

That’s the point of the test. You test how well the DR holds the further you are from base ISO.

Well presumably there are 2 different tests. Relative dynamic ranges at 'base iso' and relative DR at higher isos (or low light). They are not the same - for instance Canon sensors have much lower DR than Sony sensors at base iso but they are pretty similar at higher isos.

The 'Op' seems more interested in relative DR at base iso.

[Emanuel]

43 minutes ago, jonpais said:

Sorry, I must have missed the point of the test. I’m going back to bed now. 

Good night man or more likely to address you happy morning there instead, I believe ; ) And please be back to the old good chap we met and I still think we all want to hear that from you : ) I miss that, at least. This forum is not the same without both of you two in good vibes (I guess no need to explicitly mention the name of the other forum rebel as well, right? :-)

[mercer]

4 minutes ago, Robert Collins said:

Well presumably there are 2 different tests. Relative dynamic ranges at 'base iso' and relative DR at higher isos (or low light). They are not the same - for instance Canon sensors have much lower DR than Sony sensors at base iso but they are pretty similar at higher isos.

The 'Op' seems more interested in relative DR at base iso.

Right but I think the test in question from the video was showing the difference in DR from their respective base ISO. So vLog’s base ISO is what... 400? And sLog is 800. So when you go one stop up in ISO with vLog you’re at 800 ISO and one stop up with sLog you’re at 1600 ISO... two stops up with vLog you’re at 1600 ISO and two stops up with sLog you’re at 3200 ISO... etc, etc...