amanieux

"the upcoming A7S III is said to be using H.265 compression in their RAW codec which is an interesting hybrid of the two" it is a clever compromise and i know sony marketing team will be happy to label this "raw" but can we (the users) please stop calling "raw" these hybrid lossy codecs that have file size close to distribution codecs but has some more flexibility for better editing in post. the purpose of real raw (as defined by unprocessed data coming out of the sensor) is : 1/ to postpone all processing in post so it can be computed with more precise/sophisticated algorithms on a computer that can compute in non-real time rather than in-camera hardware that must do limited computations in real time due to restrictions on cost, size, power consumption and thermal constraints 2/ achieve higher raw lossless compression ratios for archiving as very complex/slow lossless compression algorithms can be used 3/ offer paths for image quality improvement and reduce raw lossless archive files when better algorithms comes out in the future

this 1:58 slimraw compression result is lossless and the original was lossless cinemadng, correct ?

sorry i am not a photographer i am a computer scientist, my job is to write algorithms were "1" are always visually 1 and "0" are always visually 0 no matter the angle you look at it

sorry you contradict yourself, you say the definition of raw is that its "data not yet processed" and then you say" lossy raw is still raw" but lossy compression is a "processing" of the original data and not a "minimal processing" it is a "heavy processing" involving very complex algorithms what i don't get is that blackmagic could have minimized this BRAWgate but just adding a BRAW 1:1 consisting only of the uncompressed raw data for users who want maximum quality no matter the file size as i am sure there is no possible patent litigation involved on serializing data in a file, any coder can code that in 5 minutes. it is a step down from losless cinemadng raw because it is the exact same data but in a bigger file size but it is better than having nothing if they really have a patent claim litigation to avoid. they perfectly know their users are pixel peeper, how could they think they could get away with it with just a BRAW powder in the eyes ? and you you are correct it was my mistake : "RAW" is not an acronym so we should rather write "raw" or "Raw" if its in the beginning of a sentence

don't want to split hairs but "lossy compression" can be "visually lossless" or "visually lossy" , "lossless compression" is unambiguous because it means compression which does not change a single bit from the uncompressed data so it is visually indistinguishable because it is identical

i suggest we stop amalgamating "lossy compression" with " visually lossless compression" like you involuntary just did "lossy/lossless data compression" is coming from the area of signal processing where there is no room for ambiguity https://en.wikipedia.org/wiki/Lossy_compression

ok fair enough, you are certainly right if you say that the very first stages of the image pipeline are always applied in camera but you are never going to sell me that lossy compression can be part of the definition of a raw data format

yes yes all flavors of braw in 6.2 are lossy but i guess they can deliver a lossless flavor of BRAW in a future update if they want to (this will shut down the controversy about BM really having a patent issue with cinemadng lossless RAW )

defining "raw" vs the real deal "RAW" as "meaning minimal compression to a bayered image" is very tricky because where do you put the limit of "minimal compression" is 3:1 minimal but 12:1 no longer minimal ? what about 5:1 ? they should have avoided this semantic headache an called their format : BVC for Blackmagic Video Codec and avoided the RAW misappropriation altogether . what pisses me off is not that "BRAW Q0" is closer looking to prores than to RAW, what pisses me off is that they removed the only real "RAW" with firmware 6.2

disagree lossless raw is still raw because it is bit for bit identical to uncompressed raw but lossy raw is no longer raw unless you give another definition of raw that is for me "data straight out of the sensor", altering a single bit of data disqualifies for raw otherwise where do you put the limit of more or less lossy compression and still qualifying for "lossy raw" ? is h264 a very lossy raw ?

any flavors of lossy compression is by definition not RAW, RAW can only be uncompressed or lossless compressed to justify the RAW qualification. q0 has of course less details than the lossless cinemadng RAW they removed in 6.2(as expected from any lossy compression).would i dare to say that BRAW stands for BullshitRAW ? if only q0 was 90% as good as lossless cinemadng RAW it could have been an acceptable compromise for 90% video shooters and we could not care less about this confusing and far fetch marketing concept of "lossy RAW" but it is clearly not the case so as camera conspiracy would say in music "why did you do it ?"

i don"t follow you, how can you have more than 12 steps of dynamic range on a 12 bit luminance channel coming out of a bayer sensor that is linearly sensible to light (not log sensible like our human eye are) ?

as a computer scientist it seems 12 bit of information on a bayer sensor that is linearly sensible to light it can only give 12 level of dynamic range as the luminance =0.596*g+0.299*r+0.114*b (with max values of 4095 (the brighest level of white) for r,g and b, resulting luminance value is 4095 (a 12 bit number). am i correct ?

is bRAW 3:1 lossless ? how effcient is braw compared to other lossy codecs ? for example braw 12:1 (32MB/s) it is about 2x smaller than h265 4:2:0 10 bit 400mbits found in xt3 but has it better image quality ? i know it is hard to compare a distribution codec to a acquisition codec but it is fair to compare them as xt3 is using h265 as an acquisition codec to grade in post and xt3 is the same price as bmpcc4k so who made the best compromise for best gradable footage vs smaller file size (fuji or bm) ? i have a hard time believing the few blackmagic design engineers developing braw over 2 years can come close to the efficiency of open standard codecs such as h265 that come from the work of thousands of engineers over decades .

as gordon laing sums it up correctly; "don't get too excited : no PDAF in 4k, no 1080p with the affordable ef-s lenses" - canon you are still #1 at the crippling game

can you please share where you found this data ? thanks

why do you say the sensor has to be a 3:2 ratio, most displays are 16:9, this 3:2 is coming from film era and no longer makes any sense as people are no longer printing photos on paper, new sensors should be 16:9

thanks, and what about intel 8th gen cpu is it the same or they did improve it over 7th gen ?

you don't even need a dedicated gpu, mainstream intel cpu laptop since generation 7 can decode h265 10 bit in hardware not sure if it can go up to 200mbit/s though (intel cpu generation 6 could not : https://www.cpu-monkey.com/en/compare_cpu-intel_core_i5_7200u-653-vs-intel_core_i7_6500u-588 https://www.anandtech.com/show/10959/intel-launches-7th-generation-kaby-lake-i7-7700k-i5-7600k-i3-7350k/6 )

i am quoting this eoshd article paragraph #7

"extremely good Eye-Detect AF which stays clued to a face like Dual Pixel AF does but not the ears" do you mean that xt3 has eye AF in video mode in 4k that can remain focused on the nearest eye without hunting with a fast lens wide open under f1.8 for a head close up framing so it blurs the ears and the nose but keep the nearest eye in focus even if the person moves back and forth on a chair - it would be the first time i see a camera with such a great video feature.

i have a newbie question : when we have a thin black line(less than 1 pixel width) on a white background like a telephone wire in the sky will raw 4:4:4 have double the resolution in x and double the resolution in y compared to h264 4:2:0 ? (for example i dont remember the old bmpcc in 1080p raw having comparable resolution to 4k h264 4:2:0 : ).

fair enough, we all have different taste for image . is human eyes+brain system more sensitive to resolution or to color ?

i have not looked at alexa mini but here is a difference in detail between a7sII vs a6300 for 4k video , look at the detail on the bricks in the center https://ibb.co/jghu47 agree that diffusion filters are very often used but high resolution + diffusion filter in not the same kind of dreamy like softness in the image as blurriness from lack of resolution