What’s interesting about this one is that the log profile seems to be what’s making the footage soft. Once he puts it in a standard profile, H.265 looks great!
Gary Irwin wrote:
Another interesting comparison between the Z6III and ZR shooting H.265 NLog, Spoiler: the Z6III is much sharper than the ZR.
ronno wrote:
What’s interesting about this one is that the log profile seems to be what’s making the footage soft. Once he puts it in a standard profile, H.265 looks great!
There's an interesting comment in the comments section of the video ....
Pinned by @RyanTroy willvazquez3218 1 day ago (edited) Thanks for sharing. I haven’t compared yet on using SDR, but in N log, there is a huge drop in the data rate of the ZR compared to the Z63. In both H265 and in Prores HQ. In 265 it’s more than half the data rate on the ZR compared to Z63, and it’s like 30% less with Prores HQ. Of course I’m talking about N log. I’ve been reading another forums that Nikon used less heat dissipating material in the ZR to make it smaller. There is a chance that Nikon purposely thermal throttled the ZR, cause as you know it takes more processing power to compress and write to the card a compressed codec than just writing raw to the card. This may also be why they only have one compression setting for red raw. Because a higher compression setting would require more processing, thus more heat. I know for a fact on the Z8 it overheats much faster Filming H265 compared to raw. I don’t know why N log would be more taxing on the processor than SDR, but there may be a reason. I don’t think that there will be a firmware fix. I think this is on purpose by Nikon. Because how come it’s happening with both of the compressed codecs, like H265 and regular ProRes HQ. I hope they fix it. If you open the files on VLC, you can look at the information and see the stream data rate. And you can see what the average is. Cheers.
Somebody else said they only got 40 minutes of record time on the ZR at 4k60p H.265 NLog before shutting down due to heat. They returned it saying it's useless for wedding photographers.
Since I don't shoot raw, it's looking more and more like I'll just keep my Z6III.
Hi Gary, we mentioned the below comments from willvazquez earlier in the thread.
I’m really wondering now if anyone else here has shot H265 in non-log and can comment on the detail / sharpness results.
Gary Irwin wrote:
There's an interesting comment in the comments section of the video ....
Pinned by @RyanTroy willvazquez3218 1 day ago (edited) Thanks for sharing. I haven’t compared yet on using SDR, but in N log, there is a huge drop in the data rate of the ZR compared to the Z63. In both H265 and in Prores HQ. In 265 it’s more than half the data rate on the ZR compared to Z63, and it’s like 30% less with Prores HQ. Of course I’m talking about N log. I’ve been reading another forums that Nikon used less heat dissipating material in the ZR to make it smaller. There is a chance that Nikon purposely thermal throttled the ZR, cause as you know it takes more processing power to compress and write to the card a compressed codec than just writing raw to the card. This may also be why they only have one compression setting for red raw. Because a higher compression setting would require more processing, thus more heat. I know for a fact on the Z8 it overheats much faster Filming H265 compared to raw. I don’t know why N log would be more taxing on the processor than SDR, but there may be a reason. I don’t think that there will be a firmware fix. I think this is on purpose by Nikon. Because how come it’s happening with both of the compressed codecs, like H265 and regular ProRes HQ. I hope they fix it. If you open the files on VLC, you can look at the information and see the stream data rate. And you can see what the average is. Cheers.
Somebody else said they only got 40 minutes of record time on the ZR at 4k60p H.265 NLog before shutting down due to heat. They returned it saying it's useless for wedding photographers.
Since I don't shoot raw, it's looking more and more like I'll just keep my Z6III....Show more →
Since R3D NE is just ticoraw (N-RAW) with sensor gains and color set differently, and since the ZR supports N-RAW normal quality, there is no reason why it wouldn't be able do a similar color and gain settings but use the higher compression of N-RAW Normal. It could be just marketing / product segmentation reasons why they don't offer this option.
Gary Irwin wrote:
There's an interesting comment in the comments section of the video ....
Pinned by @RyanTroy willvazquez3218 1 day ago (edited) Thanks for sharing. I haven’t compared yet on using SDR, but in N log, there is a huge drop in the data rate of the ZR compared to the Z63. In both H265 and in Prores HQ. In 265 it’s more than half the data rate on the ZR compared to Z63, and it’s like 30% less with Prores HQ. Of course I’m talking about N log. I’ve been reading another forums that Nikon used less heat dissipating material in the ZR to make it smaller. There is a chance that Nikon purposely thermal throttled the ZR, cause as you know it takes more processing power to compress and write to the card a compressed codec than just writing raw to the card. This may also be why they only have one compression setting for red raw. Because a higher compression setting would require more processing, thus more heat. I know for a fact on the Z8 it overheats much faster Filming H265 compared to raw. I don’t know why N log would be more taxing on the processor than SDR, but there may be a reason. I don’t think that there will be a firmware fix. I think this is on purpose by Nikon. Because how come it’s happening with both of the compressed codecs, like H265 and regular ProRes HQ. I hope they fix it. If you open the files on VLC, you can look at the information and see the stream data rate. And you can see what the average is. Cheers.
Somebody else said they only got 40 minutes of record time on the ZR at 4k60p H.265 NLog before shutting down due to heat. They returned it saying it's useless for wedding photographers.
Since I don't shoot raw, it's looking more and more like I'll just keep my Z6III....Show more →
I apologize for the native French video (there is an English track on Youtube also), but it provides a rather convincing explanation for the softness of the H.265 and a way to migitate it.
In short, it is the result of 3 things:
- the H.265 of the ZR is using a long GOP codec that is very sensitive to noise because it averages the noise over several images which results in soft images when noise is present,
- the ZR applies automatically different levels of noise reduction depending on the brightness of the signal, the darker the more noise reduction is applied,
- when applying a LUT in camera for display purposes, the image on the screen appears correctly exposed even when it is several stops under exposed when looking at the Wave form.
The solution would be two fold:
- what can be done now: expose correctly by looking at the wave form. The significantly reduces the noise and therefore mitigates the softness to a large extend (95% according to the video),
- what could be done later: Nikon could could deliver through firmware several improvements including the possibility to control the display when a LUT is applied to give more hints about the level of exposure when shooting log, provide better exposure tools to further reduce the risk of under exposure and a way to control the level of noise reduction that is applied to the H.265 files. These enhancements would both reduce the odds of underexposure and also enable the control of the noise reduction to get something closer to the Z6III.
I would be interested in the comments of those with better knowledge video than I have.
bernardl wrote:
The solution would be two fold:
- what can be done now: expose correctly by looking at the wave form. The significantly reduces the noise and therefore mitigates the softness to a large extend (95% according to the video),
Cheers,
Bernard
This is pretty much the same that I experienced. In broad daylight, the h265 is very comparable to Z6 III. In indoors and low light it is different. On the plus side, as a result of aggressive noise reduction, there is no shadow noise and no flickering in shadows.
Also, the discrepancy between screen and waveforms is a well known issue with Nikon view assist in all current gen cameras. I would argue that even the waveform is not very accurate, if you expose to the max brightness, the highlights will actually clip. One needs to experiment with different exposures to get the hang of it and use Zebras with custom values.
indusphoto wrote:
This is pretty much the same that I experienced. In broad daylight, the h265 is very comparable to Z6 III. In indoors and low light it is different. On the plus side, as a result of aggressive noise reduction, there is no shadow noise and no flickering in shadows.
Also, the discrepancy between screen and waveforms is a well known issue with Nikon view assist in all current gen cameras. I would argue that even the waveform is not very accurate, if you expose to the max brightness, the highlights will actually clip. One needs to experiment with different exposures to get the hang of it and use Zebras with custom values. ...Show more →
Good, but this is not aligned with the generic complains about the softness of H.265. People who have been complaining have been complaining about soft H.265 footage in bright conditions also.
bernardl wrote:
Good, but this is not aligned with the generic complains about the softness of H.265. People who have been complaining have been complaining about soft H.265 footage in bright conditions also.
Cheers,
Bernard
How bright it is (how much ambient light there is) is not the main factor, how much light you give the sensor is the important determining factor in noise. If people shoot N-Log and set the exposure of the main subject according to 800 (while the camera's hardware is really at base ISO) it results in noise. You can mitigate it by giving the sensor more light.
ilkka_nissila wrote:
How bright it is (how much ambient light there is) is not the main factor, how much light you give the sensor is the important determining factor in noise. If people shoot N-Log and set the exposure of the main subject according to 800 (while the camera's hardware is really at base ISO) it results in noise. You can mitigate it by giving the sensor more light.
Exactly my point yes. My point is that the results from those who have been complaining about the soft H.265 is probably the result of them not exposing correctly.
The ZR has the same sensor and processor as the Z6III, which has known (and acceptable) H.265 performance, so why would Nikon change the processing if it's not related to thermal issues?
On the Nikons are people exposing normally with log?
With Sonys we usually expose at +1.5 or even +1.7 with log, and the results look perfect. No noise issues, plenty of headroom.
Also have any of you looked at h.265 non log? One of the above videos suggests that non log is as sharp and detailed as it should be.
Gary Irwin wrote:
The ZR has the same sensor and processor as the Z6III, which has known (and acceptable) H.265 performance, so why would Nikon change the processing if it's not related to thermal issues?
Because they went for a different codec? Have you watched the video I sent a link to and summarized above?
Why would the ZR, that has been thermally optimized compared to the Z6III, have more thermal issues when both are fanless designs?
Let’s not forget that the whole thermal thing is only a theory from folks who still have a hard time accepting the fact that Nikon did without a fan what Canon and Sony were only able to do with a fan. They just went with the basic confirmation bias without attempting to explain the difference vs the Z6III.
Well the difference is apparently in H.265 codecs and under exposure.
ronno wrote:
Thanks - the colors look nice.
Do you like the footage?
Thanks. I am really liking the R3D footage. Especially now that I feel like I'm getting the hang of it, both in terms of exposing in-camera, and processing in Resolve.
Admittedly I still have to try the other codecs
---------------------------------------------
ronno wrote:
On the Nikons are people exposing normally with log?
With Sonys we usually expose at +1.5 or even +1.7 with log, and the results look perfect. No noise issues, plenty of headroom.
Also have any of you looked at h.265 non log? One of the above videos suggests that non log is as sharp and detailed as it should be.
On my FX3, I don't rely on the EV meter as much anymore. I now mostly look at the histogram and zebras, making sure I expose as high as possible before the highlights start to clip. For what it's worth, I think it usually does fall in the +1.7 EV range.
After a bit of back and forth on the ZR I've landed on more or less the same method. Except I'm relying on zebras and waveforms, instead of histograms. I'm still getting my head around it, but feel a lot more confident with it now than I did a few days ago.
Still haven't given h.265 a go, but might do that soon.
Someone posted this comparison video of all the codecs on the ZR today. Pretty interesting and informative:
According to the Bernard video - if people just start exposing way to the right, like we do with the Sonys, at +1.5 or +1.7 when shooting LOG, then the automatic noise reduction should not go nuts and make a mess of the files.
And meanwhile, Nikon needs to allow people to adjust the noise reduction themselves.
(and fix the log display issue while shooting)
Meanwhile I am not sure I am seeing some magic that makes this camera worth getting just for the RED RAW color science alone.
ronno wrote:
According to the Bernard video - if people just start exposing way to the right, like we do with the Sonys, at +1.5 or +1.7 when shooting LOG, then the automatic noise reduction should not go nuts and make a mess of the files.
And meanwhile, Nikon needs to allow people to adjust the noise reduction themselves.
(and fix the log display issue while shooting)
Meanwhile I am not sure I am seeing some magic that makes this camera worth getting just for the RED RAW color science alone.
I watched the video (with auto dubbing on) and it was very interesting. Sounds similar-ish to how CineEI works on Sony cameras. Basically setting the "perceptual" ISO to lower or higher than the base, so you expose for either, the shadows or highlights, as needed. I'm definitely going to try this out.
The RED color science is nice, but yeah I don't think it's reason enough on its own for anyone to make a full switch.
I do think this is a fun complimentary camera to my existing ones and a good excuse for me to try Nikon again after a multiple year hiatus. The fact that I could adapt my Sony lenses so easily definitely made it an easier decision. Would've been a different story if I had to also get a whole new set of glass just to try it out.
