only lossless is lossless. Both HE and HE* are loosy, though to differing degrees.i’ll use the mode that is mathematically lossless and provides the best speed
(i'm _hoping_ that means HE*)
If you would like to post, you'll need to register. Note that if you have a BCG store account, you'll need a new, separate account here (we keep the two sites separate for security purposes).
only lossless is lossless. Both HE and HE* are loosy, though to differing degrees.i’ll use the mode that is mathematically lossless and provides the best speed
(i'm _hoping_ that means HE*)
It seems Nikon has licensed intoPIX TicoRAW technology which they claim will result in a mathematically lossless reduction in file size of one-third with reduced power consumption.
only lossless is lossless. Both HE and HE* are loosy, though to differing degrees.
Retains same level of high image quality means that - image quality - and not retains the same bytes, just compressed. I also think that both HE variants are two levels of lossy compression, but a better one than what Nikon had before.Rich, that may be. But as far as I can tell there are conflicting indications and nobody has any credible references that would definitively indicate either way.
If you have references that seem definitive, I'd be very interested to see them.
View attachment 28170
Retains same level of high image quality means that - image quality - and not retains the same bytes, just compressed. I also think that both HE variants are two levels of lossy compression, but a better one than what Nikon had before.
Is this drama only in High Efficiency setting ?
Or dose it effect the full raw file.............
from NIkon USA website. There is lossless and then compressed files. as above nearly-lossless. But what happens at high ISO. That is my point. At lower ISO lossless, HE, and HE* may all be great. At higher ISO there may be a difference. A 1 bit difference is a difference. So by your own admission, John HE and HE* may be (slightly) different.Rich, that may be. But as far as I can tell there are conflicting indications and nobody has any credible references that would definitively indicate either way.
If you have references that seem definitive, I'd be very interested to see them.
VVVV from Nikon's literature. note the intoPIX reference:
View attachment 28170
TicoRAW definitely *offers* a mathematically lossless mode, the question is, did Nikon implement it? vvvvvv
View attachment 28171
It’s just speculation about the High Efficiency* setting.
from NIkon USA website. There is lossless and then compressed files. as above nearly-lossless. But what happens at high ISO. That is my point. At lower ISO lossless, HE, and HE* may all be great. At higher ISO there may be a difference. A 1 bit difference is a difference. So by your own admission, John HE and HE* may be (slightly) different.
View attachment 28173
Let me ask you a few simple question. If either HE or HE* was lossless, won't Nikon claim it is lossless? Not that HE* is not visual as good as lossless but it is NOT lossless.Remember “compressed” != “lossy”. We know all the modes are compressed. We know at least one is lossless and at least one is lossy. IMO nobody has pointed out anything that is definitive about HE*. Your guess is that it is lossy and my hope is that it is lossless.
I will say there are reasons it would make sense that HE* is lossless. The TicoRAW provide higher compress at faster rates. This could enable higher lossless frame rates. There are rumors of coming improvements. If they could deliver 30fps lossless RAW, wouldn’t that be worth while? In fact, if the Tico RAW lossless is better in all ways, why would you not add it?
Another little thing. Note all the folks doing demos of the buffer used HE*. If it is lossless it would totally make sense the engineers would have told them to use it instead of lossless compressed because it is the same.
Anyhoo, hopefully someday we’ll know for sure one way or another.
HE* is not lossless but it is very close.
loved the interview. here's a link to the book: The Real DealInterview Joe McNally with GoW
You're kind of arguing against something that you've been perfectly happy with until now. Raw files have always had some noise reduction applied to RAW files whether you know it or not. Raw files already have a color science and white balance applied to them both by the camera and any software you use to process it. No offense intended but it seems like you're getting upset over a very trivial potential change. As far as the different noise amount in the shadows in your example, if the shadows are raised to a point that is pleasing to you regardless of the noise, its likely an adjustment you would have made anyway. If its not pleasing then you would likely lower the shadows, or add contrast to lower them, thus negating the in camera "boost" anyway.It is bad. Because if various parts of the image already have diferrent gain baked in, my +1 exposure will result in real +2 in some cases, +1 in other, so the noise increase is different across the different parts of the image.
Note we already have Active D-Lighting, and there are many cases where it does a good job. But it does it *on the JPEG output*, not on the RAW file such that I can't reverse it.
It's AI based - that's exactly what bothers me. If I wanted automatic, I'd just shoot with the iPhone all day. I want to stay in control of how the various parts of the image are processed, and be able to decide how much noise I want. The camera deciding I don't want noise - how the heck does it know what I want?
If Nikon really did this, I'd be royally pissed off. I always disable all in-camera noise processing and apply it - if and to what extent I want - on the computer, on the areas that I want, on a big screen. There are places where "auto" is good and has its purposes, but not in the RAW output of a - supposedly - professional camera.
Edit: Sorry for the rant, but I would be really, really surprised if Nikon is compensating stacked sensor DR issues by faking image data. Yes, noise reduction is faking raw data. It will be visible clearly in photons2photos graphs, so we'll know as soon as raw files are available.
loved the interview. here's a link to the book: The Real Deal
You're kind of arguing against something that you've been perfectly happy with until now. Raw files have always had some noise reduction applied to RAW files whether you know it or not. Raw files already have a color science and white balance applied to them both by the camera and any software you use to process it. No offense intended but it seems like you're getting upset over a very trivial potential change. As far as the different noise amount in the shadows in your example, if the shadows are raised to a point that is pleasing to you regardless of the noise, its likely an adjustment you would have made anyway. If its not pleasing then you would likely lower the shadows, or add contrast to lower them, thus negating the in camera "boost" anyway.
It's already being done with noise/ISO on a global basis by Nikon, Sony, and Canon. This shows the D5, A7riv, and R6.Maybe? But note what bothers is not a general noise reduction/white balance - that if of course a given. What bothers me is the news of the camera applying different changes to different parts of the image, like Active D-lighting is doing to the JPEG output. As far as I know, this was not the case until now. If you tell me this was already the case, I'd be surprised but willing to learn more.
And lowering the shadows after the camera has raised them will, AFAIK, not result in the exact original values.
It's already being done with noise/ISO on a global basis by Nikon, Sony, and Canon. This shows the D5, A7riv, and R6.
https://photonstophotos.net/Charts/PDR.htm#Canon EOS R6,Nikon D5,Sony ILCE-7RM4
Every time you see a stair step change in the graph, NR is being applied on a global basis at a RAW level.
Nikon is changing their approach with the Z9 to differentiate between a subject or area with detail and a detail-less background. The same AI that identifies the subject can also be used as part of processing.
It's laughable for now because depth mapping is still in its infancy, but in VFX industry some formats like deepEXR are used to fake about every optical phenomenon and layering (what's in front of what) with great success, and I've seen some phones capable of generating pretty convincing depth maps recently. So sooner or later, fake DoF will be indistinguishable from its optical counterpart for normal image viewing.Because this steps into the "image post-processing" area (for me). And all the "AI-based image processing" that I see in phones (e.g. fake bokeh) is laughable today.
from NIkon USA website. There is lossless and then compressed files. as above nearly-lossless. But what happens at high ISO. That is my point. At lower ISO lossless, HE, and HE* may all be great. At higher ISO there may be a difference. A 1 bit difference is a difference. So by your own admission, John HE and HE* may be (slightly) different.
View attachment 28173
And as I said a few times already, that I know and that I'm fine with. What bothers me is:
Because this steps into the "image post-processing" area (for me). And all the "AI-based image processing" that I see in phones (e.g. fake bokeh) is laughable today.
I agree. This has been a controversial point since Nikon first introduced it and then equivocated on what it really is.
We probably won't really know until enough Z9's are out here, in the wild, with production firmware and finalized raw converters, so some credible expert(s) can properly evaluate.
If it is truly lossless, then no problema. If lossy, then it is possibly scene/exposure dependent, and will be important to understand the limitations.
Can't wait to see the outcome.
It's already being done with noise/ISO on a global basis by Nikon, Sony, and Canon. This shows the D5, A7riv, and R6.
https://photonstophotos.net/Charts/PDR.htm#Canon EOS R6,Nikon D5,Sony ILCE-7RM4
Every time you see a stair step change in the graph, NR is being applied on a global basis at a RAW level.
Nikon is changing their approach with the Z9 to differentiate between a subject or area with detail and a detail-less background. The same AI that identifies the subject can also be used as part of processing.
Maybe? But note what bothers is not a general noise reduction/white balance - that if of course a given. What bothers me is the news of the camera applying different changes to different parts of the image, like Active D-lighting is doing to the JPEG output. As far as I know, this was not the case until now. If you tell me this was already the case, I'd be surprised but willing to learn more.
And lowering the shadows after the camera has raised them will, AFAIK, not result in the exact original values.