Last edited:
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
billzilla2000
Veteran Member
All Bushs Fault
Senior Member
- Messages
- 1,157
- Solutions
- 1
- Reaction score
- 783
Thanks for clarifying.
PVCdroid
Senior Member
I hope not. My concern is that it would compromise the cameras and built in functionality in some ways. Sony has pretty much mastered their raw compression and it is very good across all their cameras. The only way to show any deleterious effects of their compression is to turn Raw Digger upside down and inside out showing star trails at 300%. It's all ridiculous. Let them use Canikon if lossless raw is required. Sony doesn't need the OCD types anyway. If someone were to show me that the D810 vs A7r shows artifacts from the Sony file from the exact same scene and same exposure, ISO, aperture, etc. with raw files provided, I might reconsider. It has never been done as far as I know.
Last edited:
Depends on how many people engage in the milky way shot fad currently in vogue.I think it can matter, but I think it can be overblown at the same time (and I started this thread.) Clearly there are cases (star trails, high contrast night scenes, even the occasional high contrast day scene) where the Sony RAW scheme shows artifacts where Nikon's lossless scheme doesn't. The question is, why can't Sony offer a lossless mode to, if nothing else, appease those of us who are troubled by its absence?
I think the bigger problem might be the drop to 12-bit precision with continuous, and bulb, shooting (and a couple other modes, but in situations where anything better is just technically impossible--like silent shutter mode. That's understandable due to HW limitations.)
AFAIK Jim Kasson was the first to bring these limitations to light, on his blog, and Rishi added them as potential shortcomings to his A7RII preview after I messaged him, but I think many people are still unaware. Time to start a new thread
naththo
Leading Member
Yes definitely impossible to recover the red channel blownout on red flowers. Red roses are the one is real challenge to get red colour right spot on. Cos since that CMOS has two GG and 1 R and B. Red is the weakest one than the rest. Not much can do about. Even the sharpness in red might be at weakest is it?
SirSeth
Veteran Member
Is this a knee jerk reaction to previous complaining or trolling? Do you mean to discourage people from wanting lossless RAW? Why wouldn't choice and improvement be desirable for those who do want it? It's not like cameras with lossless RAW don't have compression options.
Sony has a great sensor that you can pull a lot out of even with lossy RAW. If more can be pulled out of lossless RAW, then there is no reason to show proof of need. There are plenty of people, myself included, that just want lossless because we believe it would be a benefit some of the time. If you don't believe it would actually add anything for any occasion or shooting scenario, then maybe you've got a point and I'm all ears as to why.
Cheers, Seth
I don't think a request for exemplar images is unreasonable. There are not many examples extant. As far as I know, with the exception of the simulation images that I've posted on my blog, none show artifacts related to the tone curve, and there are only a few that show artifacts from the delta modulation. The photographic community would be better off if we had a richer set of compression affected images to look at.
One deleterious effect of the current lack of variety of such images is the recurring mis-identification of anomalies in a7X images as compression artifacts, which has the effect of creating an impression that images with such artifacts are common.
Just for the record, my position on the Sony cRAW algorithm is as follows. While I would like lossless compression as an option, and would choose it if it were available (I use 14-bit lossless on my D810, which has six precision/compression options), I have never seen an artifact in a cRAW image that I haven't tried to create (and hardly any when I have tried my darnedest), and I will continue to buy and use a7x cameras even if Sony doesn't provide lossless compression.
Jim
While I deeply appreciate the compliments, I'd like to come to dpr's defense (not Amazon, since they are simply a review aggregator) on this one. I am an electrical engineer with 50 years of experience, and I worked as a color scientist for 6 years. I have programming ability and a powerful suite of Matlab (the lingua franca of color science) classes that I've written for image analysis and synthesis at my disposal. I am retired, and have time to work on this. If dpr could find someone with my background and abilities who wanted to work for a living, they would have to pay them a considerable amount of money to compete with what they could earn working in, say, the Santa Clara Valley.I guess it speaks volumes about the quality of the self-professed "professional reviewers" when a blogger (no offence to Jim since he is doing a fantastic job) actually produces better technical output than a large company like dpreview/amazon ever could.AFAIK Jim Kasson was the first to bring these limitations to light, on his blog, and Rishi added them as potential shortcomings to his A7RII preview after I messaged him, but I think many people are still unaware. Time to start a new thread
Another thing that should be mentioned here is that I don't usually produce comprehensive reviews of cameras. I concentrate on the technical side, which is where my strengths lie. If I tried to do complete reviews for every camera I worked with, I'd review far fewer cameras. I also don't organize my site the way that dpr does. I wish I had the time, inclination, and energy to do so, but I'd rather make images and test cameras than organize data bases.
Jim
--
http://blog.kasson.com
Last edited:
Love your work Jim, and much appreciated!While I deeply appreciate the compliments, I'd like to come to dpr's defense (not Amazon, since they are simply a review aggregator) on this one. I am an electrical engineer with 50 years of experience, and I worked as a color scientist for 6 years. I have programming ability and a powerful suite of Matlab (the lingua franca of color science) classes that I've written for image analysis and synthesis at my disposal. I am retired, and have time to work on this. If dpr could find someone with my background and abilities who wanted to work for a living, they would have to pay them a considerable amount of money to compete with what they could earn working in, say, the Santa Clara Valley.I guess it speaks volumes about the quality of the self-professed "professional reviewers" when a blogger (no offence to Jim since he is doing a fantastic job) actually produces better technical output than a large company like dpreview/amazon ever could.AFAIK Jim Kasson was the first to bring these limitations to light, on his blog, and Rishi added them as potential shortcomings to his A7RII preview after I messaged him, but I think many people are still unaware. Time to start a new thread
Another thing that should be mentioned here is that I don't usually produce comprehensive reviews of cameras. I concentrate on the technical side, which is where my strengths lie. If I tried to do complete reviews for every camera I worked with, I'd review far fewer cameras. I also don't organize my site the way that dpr does. I wish I had the time, inclination, and energy to do so, but I'd rather make images and test cameras than organize data bases.
Jim
--
http://blog.kasson.com
My husband was going go into electrical engineering, but he switched to pure math as an undergrad and then did math in graduate school as well. He has a background in applied mathematics, biomath and image analysis on the medical side, programs in Matlab and C# (and a few others), and he is pretty interested in this stuff, too. And you are right, they wouldn't be able to pay him enough!
As an aside, he just wrote us a Windows program with UI in Visual Studio to check folders for duplicate RAW + JPEG, and to only delete the JPEG file if a RAW file with the same name exists in that folder. This is to clean up some of our folders.
Meanwhile I work as a web designer by trade, so I made our little website, do the post-processing since I have years of Photoshop experience, and maintain the content management system of the backend.
It's fun that so many disciplines go into and can help photography as a hobby!
--
http://www.lightfinity.net
Last edited:
SirSeth
Veteran Member
billzilla2000
Veteran Member
I am 100% in that camp, Jim. That said, if you want to see some examples, I can produce same - I've gotten rather handy at sniffing out the compression artifacts, but as usual, it requires extensive reworking of the image (to a practically non-usable state) to reproduce them so that they're easily seen.
The point of the simulations that I did was to look at the changes that one compansion (compression/expansion) operation made to an image. You can't do that by comparing a D810 image to an a7R image even if the raw processing (except for compression) were the same and the read noise were zero because of the stochastic nature of photon capture.
But you can do it with a simulator.
Loading…
blog.kasson.com
Can you see the Sony raw compression artifacts? - the last word
In the previous few days’ posts, I’ve established that: The Sony raw compression algorithm generates artifacts that can be made visible by subtracting an uncompressed image chosen to stress the algorithm from one that has been compressed and decompressed and subjecting the result to...
blog.kasson.com
Jim
That would be useful. I think that once people see what the artifacts look like, they are less likely to mis-identify other things as compression artifacts.
I'm assuming the artifacts that you're seeing are delta mod related. I have not seen tone curve artifacts except with my simulator.
Jim
billzilla2000
Veteran Member
ttran88
Senior Member
Everyone and their mother is made sure to have been told of Sony and lossy Raw's. LoL
I've seen talk about shadow posterization, but that doesn't make much sense to me, since the slope of the tone curve is 1 if you believe the a7x is a 13-bit device (or 2 if you think it's a 14-bit one) in the deep shadows, and by the time it changes to 2 (or 4), there's already plenty of photon noise.
Thanks.
Jim
Keyboard shortcuts
- f
- Forum