Z6III — The pieces are there, but the execution is a little sloppy. Panasonic seems to be using the same image sensor now and getting better dynamic range and less jittery deep shadows on video, for instance.
And the FW implementation is more than a little sloppy. Why do I need to go into a menu and suffer a screen blackout for the privilege of attempting focus shift shooting?
It's really not an issue in any practical situation, nobody is pushing shadows to this extreme on regular basis and the DR difference in real photos is negligible.
Z6 III is an amazing camera for what it is. People tend to focus on the tiniest thing just to say something is better while in real aplication it have little to no consequences.
Immortal wrote:
It's really not an issue in any practical situation, nobody is pushing shadows to this extreme on regular basis and the DR difference in real photos is negligible.
Z6 III is an amazing camera for what it is. People tend to focus on the tiniest thing just to say something is better while in real aplication it have little to no consequences.
The Z6 III's noise issues are noticeable even without any shadow pushing:
.
Thom Hogan gets attention by poking holes in anything he can find...
People constantly point to him as an expert, but post like this are more about keeping the clicks going...
OwlsEyes wrote:
Thom Hogan gets attention by poking holes in anything he can find...
People constantly point to him as an expert, but post like this are more about keeping the clicks going...
jmo,
bruce
I'm not sure that applies here, considering what he had to say about the matter in his Z6 III review:
"I see a maximum of about 10.5 stops of dynamic range in my tests of the Z6III. I regard that as more than enough for virtually any use."
snapsy wrote:
The Z6 III's noise issues are noticeable even without any shadow pushing:
Please do your research before you post something over 1 year old. Even on the same channel you got an updated video from for 4 months ago, after the firmware update.
Flickering been significantly reduced and there is almost 0 chance you gonna see anything. That's on video side.
On photo side this was a non issue to begin with.
Again, people overblown the tiniest thing just to point out that somebody else(camera maker) is doing it better while in real life there is almost 0 difference.
This is an non issue in photo and its more than fine on the video side after the firmware update.
Immortal wrote:
Please do your research before you post something over 1 year old. Even on the same channel you got an updated video from for 4 months ago, after the firmware update.
Flickering been significantly reduced and there is almost 0 chance you gonna see anything. That's on video side.
The video I shared you with and the follow-up video after it about the new firmware were created by me - that's my YT channel. The flickering was greatly reduced but the noise difference remains.
The last post in our very own Z6iii thread here, now archived due to inactivity, was about Pany using the Z6iii sensor...in May. Old news, catch up Thom.
snapsy wrote:
The video I shared you with and the follow-up video after it about the new firmware were created by me - that's my YT channel. The flickering was greatly reduced but the noise difference remains.
I dont understand why you posted your old video instead of the new one to begin with, where flickering is pretty much fine even if the noise remained the same.
And again thats on video side, on photo side this was and is a non issue in real world aplication.
Immortal wrote:
I dont understand why you posted your old video instead of the new one to begin with, where flickering is pretty much fine even if the noise remained the same.
And again thats on video side, on photo side this was and is a non issue in real world aplication.
Because the shadows are still jittery, which is what Hogan specifically complained about in his post being discussed in this thread and what my follow-up window demonstrated as well. And the noise regression vs previous Nikon models (and competing cameras) remains, even with the same sensor.
RoamingScott wrote:
The last post in our very own Z6iii thread here, now archived due to inactivity, was about Pany using the Z6iii sensor...in May. Old news, catch up Thom.
The fact they share the same sensor is old news - the fact Panasonic was able to extract better performance out of the same sensor was only recently discovered.
1) The dumbbell scene is incredibly underexposed despite the claim it’s “exposed correctly” to Nikon’s standards. Weird. It is especially underexposed in prep for an ETTR video. The waveform values seem incredibly suppressed. Check out this video, in contrast - he even directly addresses the "exposed to Nikon's standards" fallacy:
I think sticking to overly literal interpretations of "correct exposure" potentially lead to issues, and this has been true for ages, and why, for example, many of us would reject our light meter's assertions of "correct exposure" if left to their out of the box defaults.
2) Even correcting the exposure and, for the sake of argument, let’s say the problem persists - back when it was acknowledged pre-firmware improvements, it looks like a modest noise reduction (5-10%) and CST from source video to rec-709 in Resolve makes it disappear entirely. There’s a YT video demonstrating this:
It’s important to know these things are there so we watch for them at ISO3200 and above, and correct for them, but every camera has its shortcomings and quirks and it’s our job to know the gear and work with them.
jlafferty wrote:
It looks to me, reviewing the snapsy video:
1) The dumbbell scene is incredibly underexposed despite the claim it’s “exposed correctly” to Nikon’s standards. Weird. It is especially underexposed in prep for an ETTR video. The waveform values seem incredibly suppressed.
2) Even correcting the exposure and, for the sake of argument, let’s say the problem persists - back when it was acknowledged pre-firmware improvements, it looks like a modest noise reduction (5-10%) and CST from source video to rec-709 in Resolve makes it disappear entirely. There’s a YT video demonstrating this.
It’s important to know these things are there so we watch for them at ISO3200 and above, and correct for them, but every camera has its shortcomings and quirks and it’s our job to know the gear and work with them....Show more →
N-Log middle gray is at 35% IRE, which is where middle gray in the scene is exposed to. This is demonstrated by both the zebras and histogram from the camera displayed in the video. Yes, there are shadows in the scene - the purpose of log-encoded video is to get access to the full dynamic range out of the sensor, which means there will be shadows. If doing so yields unsatisfactory results then the problem isn't the scene that attempts to access the dynamic range but the sensors inability to deliver on it.
The video you're referring to about the noise remediation steps was probably mine as well. I included various NR techniques to mitigate both the flickering and noise, the later of which comes at the expense of detail.
snapsy wrote:
N-Log middle gray is at 35% IRE, which is where middle gray in the scene is exposed to. This is demonstrated by both the zebras and histogram from the camera displayed in the video. Yes, there are shadows in the scene - the purpose of log-encoded video is to get access to the full dynamic range out of the sensor, which means there will be shadows. If doing so yields unsatisfactory results then the problem isn't the scene that attempts to access the dynamic range but the sensors inability to deliver on it.
The video you're referring to about the noise remediation steps was probably mine as well. I included various NR techniques to mitigate both the flickering and noise, the later of which comes at the expense of detail....Show more →
I've updated my post with links, and they're definitely not yours. I think there's value in uncovering problems, but at the same time, your videos merely demonstrate issues without putting them into a context other than academic demos. People looking to solve the problem for appealing/interesting/real world footage, and not simply "access the full dynamic range of the sensor" seem to be doing it. It's a bit of a silly video, but the opening makes the point really well:
jlafferty wrote:
[I've updated my post with links, and they're definitely not yours. I think there's value in uncovering problems, but at the same time, your videos merely demonstrate issues without putting them into a context other than academic demos. People looking to solve the problem for appealing/interesting/real world footage, and not simply "access the full dynamic range of the sensor" seem to be doing it.
I dedicated a section to an effective remediation strategy of the flickering issue in the very same video I originally demonstrated the issue:
.
.
As to your other point, when did high-dynamic range scenes become unreal-world footage, esp considering the N-Log specification is designed for 14+ stops and the Z6 III's base N-Log ISO is designed to use much of that? I never really understood that argument - it's like showing a car made to do 130 mph losing control when that speed is reached, then someone saying "yeah, but you're not supposed to drive it at 130 mph".
As for my footage not being appealing or interesting, I confess, it's not - my interest is in highlighting issues and showing workarounds. I leave the art of interesting locales and cinematic presentations to others who are much better at it than I.
