It was not entirely intended to sound like that but I do have a very strong opinion in really just this one area alone. Other than that there are no rules you can't break IMHO.
Anyway I have two thoughts on this. One is I think there is a translation issue here with Sony and we won't know until we see what we get.
The second is as you mentioned when to shoot certain things with the 2 choices given . That's a interesting one but will become confusing to a lot of folks but agree with your comments on when to use what.
I'm just a little baffled how Sony got the wrong message. I mean they have trusted Artisians on hand to refer too like Brain Smith for example. He knows exactly what we are after. How can that message get screwed up.
Just a guess but I think Sony is just using the wrong wording here but the right file format.
Tariq Gibran wrote:
Just OT a bit but having recently started using compact flash cards again with another camera, I'm amazed at the speed advantage the better ones hold over their counterpart SD siblings (compare Sandisk Extreme Pro for both formats). Not only that but at some point SD cards got really flimsy and cheap with regard to build quality. I have had Sandisk Extreme Pro SD cards just break at their thinnest point recently near the contacts. By contrast, the big compact flash cards are very robust and stiff. Maybe that Lexar 2000 SD card would be even faster but I know from recently buying a Lexar 1000x SD card they are unfortunately flimsy with regard to build quality as well. I think this must have to do with the extra contacts needed for these faster, higher capacity SD cards. ...Show more →
Yea agree they are starting to feel even thinner these days or flimsy as you say and not making me happy either. I'm using both Sandisk Extreme Pros and the Lexar 1000 as well. I do like CF cards better as they just feel robust. I wish there where more ruggedizied SD cards available or something that does not take your breath away when handling them.
My guess is that there is dedicated hardware to lossy compress the RAW file before it is written, and it may not be flexible enough to implement (efficiently) a lossless compression algorithm. So their choices are limited by hardware and the best they can do is bypass the lossy compression step and write uncompressed RAW.
But all this is just guesses....
-Tim
GMPhotography wrote:
It was not entirely intended to sound like that but I do have a very strong opinion in really just this one area alone. Other than that there are no rules you can't break IMHO.
Anyway I have two thoughts on this. One is I think there is a translation issue here with Sony and we won't know until we see what we get.
The second is as you mentioned when to shoot certain things with the 2 choices given . That's a interesting one but will become confusing to a lot of folks but agree with your comments on when to use what.
I'm just a little baffled how Sony got the wrong message. I mean they have trusted Artisians on hand to refer too like Brain Smith for example. He knows exactly what we are after. How can that message get screwed up.
Just a guess but I think Sony is just using the wrong wording here but the right file format.
GMPhotography wrote:
Yea yea I'm a born optimist alright. Lol
Yeah, this firmware upgrade will give us losslessly compressed RAW, no 12 bits limit with electronic shutter, clean long exposures, ISO 32, working PDAF with all Canon lenses, lighter weight (since the compression is so heavy), improved battery life, and free donuts for all
alundeb wrote:
Yeah, this firmware upgrade will give us losslessly compressed RAW, no 12 bits limit with electronic shutter, clean long exposures, ISO 32, working PDAF with all Canon lenses, lighter weight (since the compression is so heavy), improved battery life, and free donuts for all
Sign me right up. I like those French Style glazed donuts, can I get two please. Lol
GMPhotography wrote:
Anyway I have two thoughts on this. One is I think there is a translation issue here with Sony and we won't know until we see what we get.
The wording is fairly clear, in the press release, and more importantly in the menu screenshot Fred posted. There are exactly two options: compressed oruncompressed. Anyone hoping that "uncompressed" means lossless compressed is going to be disappointed.
GMPhotography wrote:
I'm just a little baffled how Sony got the wrong message. I mean they have trusted Artisians on hand to refer too like Brain Smith for example. He knows exactly what we are after. How can that message get screwed up.
They didn't get the wrong message. The simple truth is that uncompressed is easier to implement and more importantly easier to test. Imagine if they implemented a lossless compression algorithm but got it slightly wrong...
alundeb wrote:
Yeah, this firmware upgrade will give us losslessly compressed RAW, no 12 bits limit with electronic shutter, clean long exposures, ISO 32, working PDAF with all Canon lenses, lighter weight (since the compression is so heavy), improved battery life, and free donuts for all
Matt Grum wrote:
They didn't get the wrong message. The simple truth is that uncompressed is easier to implement and more importantly easier to test. Imagine if they implemented a lossless compression algorithm but got it slightly wrong...
I don't know why or how they could get a lossless compression algorithm wrong when they already have one which they used on the older cameras (such as the A900). I suspect it's a hardware limitation as Tim suggests.
p.12 #10 · Uncompressed 14 bit for A7sii and A7rii
It may not be a hardware limitation so much as different hardware. The a900 had two processors, which are now (I believe) 4 generations old in it. They were 32bit RISC processors, a general purpose processor. With the new Bionz X, and all the generational improvements and instruction sets, it may just be a huge software issue to port (and more importantly, optimize) the code to the new processor. Till now, this work was done with 2 processors. Porting work done symmetrically in the past to a single chip is a ton of work.
Tariq Gibran wrote:
I don't know why or how they could get a lossless compression algorithm wrong when they already have one which they used on the older cameras (such as the A900). I suspect it's a hardware limitation as Tim suggests.
p.12 #11 · Uncompressed 14 bit for A7sii and A7rii
Tariq Gibran wrote:
I don't know why or how they could get a lossless compression algorithm wrong when they already have one which they used on the older cameras (such as the A900). I suspect it's a hardware limitation as Tim suggests.
Did the A900 definitely have a lossless RAW mode? It's hard to find a definitive answer online - I thought there were only two options, normal and compressed, and I've just downloaded a bunch of A900 RAWs from here and they were all ~36MB which is exactly what you'd expect if they were completely uncompressed.
p.12 #12 · Uncompressed 14 bit for A7sii and A7rii
Raw files vary in size depending how much information is in the picture, if they were all 36, then compression is taking place albeit lossless. That has been my understanding of my 900.
p.12 #13 · Uncompressed 14 bit for A7sii and A7rii
Matt Grum wrote:
Did the A900 definitely have a lossless RAW mode? It's hard to find a definitive answer online - I thought there were only two options, normal and compressed, and I've just downloaded a bunch of A900 RAWs from here and they were all ~36MB which is exactly what you'd expect if they were completely uncompressed.
You mean losslessly compressed? It's a 24MP camera, so I imagine uncompressed would be close to 48MB.
p.12 #14 · Uncompressed 14 bit for A7sii and A7rii
mogul wrote:
Raw files vary in size depending how much information is in the picture, if they were all 36, then compression is taking place albeit lossless. That has been my understanding of my 900.
Lossless compression gives rise to varying file sizes depending on content, uncompressed should give you more or less consistent sizes. I was seeing more or less consistent sizes (from 35.9MB to 36.7MB).
hiepphotog wrote:
You mean losslessly compressed? It's a 24MP camera, so I imagine uncompressed would be close to 48MB.
As far as I know the a900 is a 24 megapixel 12-bit camera, so uncompressed RAWs should be 24 * (12 / 8) = 36 MB, which is pretty much exactly what the files I downloaded were.
p.12 #15 · Uncompressed 14 bit for A7sii and A7rii
Matt Grum wrote:
Lossless compression gives rise to varying file sizes depending on content, uncompressed should give you more or less consistent sizes. I was seeing more or less consistent sizes (from 35.9MB to 36.7MB).
As far as I know the a900 is a 24 megapixel 12-bit camera, so uncompressed RAWs should be 24 * (12 / 8) = 36 MB, which is pretty much exactly what the files I downloaded were.
Uncompressed raw files also contain an embedded jpg image, that varies in size with image content and noise. This explains the small variation you see.
That's not value posterization, that's really processing and a color profile issue... Value posterization in a blue sky is quantiation noise (actual value > qantized value error), and that is very rarely visible in cases like this.
-unfortunately it is an issue that is amplified when you apply the profiles needed for the Sony type color filters, AFTER your chosen raw converter applies pre-processing noise reduction. THAT may induce posterization into bands of clearly visibly delimited colors in very smooth gradients.
Could you mail me that exact raw, then I could show you a few things about the "posterization" in it?
p.12 #17 · Uncompressed 14 bit for A7sii and A7rii
alundeb wrote:
Yeah, this firmware upgrade will give us losslessly compressed RAW, no 12 bits limit with electronic shutter, clean long exposures, ISO 32, working PDAF with all Canon lenses, lighter weight (since the compression is so heavy), improved battery life, and free donuts for all
I doubt the "12-bit read in certain situations increased to full 14-bit" part. Mainly because I know that part is actually communicated directly into the sensor (instructions form the processor to the sensor about how to read and transmit the signal, present in all Exmor-typ senor plate API's) to give you faster ADC times, more MP's/ms throughput on the sensor die proper). The same is true for video mode and when you're in LV-mode between each exposure. It facilitates a much faster reacting AF and a much better (but still quite poor...) subject-following processing.
The rest might be true though. I'm already looking forward to the free doughnut!
p.12 #18 · Uncompressed 14 bit for A7sii and A7rii
theSuede wrote:
I doubt the "12-bit read in certain situations increased to full 14-bit" part. Mainly because I know that part is actually communicated directly into the sensor (instructions form the processor to the sensor about how to read and transmit the signal, present in all Exmor-typ senor plate API's) to give you faster ADC times, more MP's/ms throughput on the sensor die proper). The same is true for video mode and when you're in LV-mode between each exposure. It facilitates a much faster reacting AF and a much better (but still quite poor...) subject-following processing.
The rest might be true though. I'm already looking forward to the free doughnut!...Show more →
Any time Joakim! You will get the weight saving worth in donuts, kg per kg, bring a truck!
p.12 #19 · Uncompressed 14 bit for A7sii and A7rii
That's not value posterization, that's really processing and a color profile issue... Value posterization in a blue sky is quantiation noise (actual value > qantized value error), and that is very rarely visible in cases like this.
-unfortunately it is an issue that is amplified when you apply the profiles needed for the Sony type color filters, AFTER your chosen raw converter applies pre-processing noise reduction. THAT may induce posterization into bands of clearly visibly delimited colors in very smooth gradients.
Could you mail me that exact raw, then I could show you a few things about the "posterization" in it?
When the A7R came out there were a lot of discussion about sky posterization. From the RAW files I was able to test, the issue was always related to profile mismanagement. (assign/convert)
p.12 #20 · Uncompressed 14 bit for A7sii and A7rii
Tariq Gibran wrote:
I don't know why or how they could get a lossless compression algorithm wrong when they already have one which they used on the older cameras (such as the A900). I suspect it's a hardware limitation as Tim suggests.
I found the original a900 manual and there is no lossless RAW option, and given the a900 RAWs I've found have been uncompressed I don't think Sony have ever offered lossless compression, so they would have to go and develop it from scratch (or rather develop an implementation from scratch, there are plenty of algorithms out there to choose from) so it's looking even less likely we'll see lossless in this firmware update.
Value posterization in a blue sky is quantiation noise (actual value > qantized value error), and that is very rarely visible in cases like this.