Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Apple + RED

You don't think? or are you just to lazy to google for 5 minutes yourself?

It has become quite clear at this point that you enjoy speculating on topics you know nothing about.
 
It has become quite clear at this point that you enjoy speculating on topics you know nothing about.

How relevant is a RAW compression ratio of more than 6 times these days, considering current flash prices and the need for high quality and flexibility in post?
This is a question

More than 80% of the blockbusters is still shot in uncompressed ARRIRAW (so it seems they don't give a sh*t about compressed RAW).
Just check the high budget($150,000,000+) blockbusters on IMDB

Sony has a patent aggrement with RED so they can do whatever they want.
In February 2013, Red filed for an injunction against Sony, claiming that several of its new CineAlta products, particularly the 4K-capable F65, infringed on patents the company held. They requested that Sony not only be forced to stop selling the cameras, but that they be destroyed as well. Sony filed a countersuit against Red in April 2013, alleging that Red's entire product line infringed on Sony patents. In July 2013, both parties filed jointly for dismissal, and as of July 20, 2013, the case is closed.

Canon has rawlite which has a lower compression ratio than the 6 times mentioned in the RED patent.
Straight from the RED patent: an image processing system configured to compress and store in the memory device the raw image data at a compression ratio of at least six to one and remain substantially visually lossless , and at a rate of at least about 23 frames per second

Canon C500mkII cinema RAW lite 5952 x 3140: 25,00P, 2.1 Gbps 12-bits Pure RAW is 5952 x 3140 x 16bits x 25 fps = 7.476 Gbps
7.476/2.1 ~ 3.56 hence less than 6 times.


Blackmagic has BRAW which doesn't infringe the RED patent but gives the same flexibility as RAW.
That is the reason they made BRAW.

Many who shoot RED and need the best possible quality use the lowest possible compression ratio (we do with 7k 4:1 24..25fps).
Since .R3D compression is based on JPEG 2000 wavelet compression it becomes lossy after a compression ratio higher than 2..3 in most cases.
With a confetti bomb it will be closer to 1 and with a static enviroment it will be higher than 3


The emerging markets (China, India, etc..) don't give a sh*t about US-only patents and they kind of don't like Apple(to put it mildly).
Even FoxNews acknowledges this and off course Trump who started to fight back

Fast flashstorage sell's for less than $200/TB with a read/write speed exceeding 3 GB/s.
https://www.newegg.com/p/1B4-00SK-000S5?Description=micron%209300&cm_re=micron_9300-_-1B4-00SK-000S5-_-Product
$708/3.84TB ~ $185/TB

Apple is kind of late to the compressed RAW party, so I think it will be hard for them to be as relavent with ProResRAW as they where with ProRes.
RedCodeRAW, cDNG, ARRIRAW, X-OCN, CineformRAW,KineRAW, SlimRAW, Canon RAW lite, etc.. where all earlier to the market than ProResRAW, I can't recollect any RAW format for motion picture that is released after ProResRAW.

No speculation whatsoever.
 
Hold on, Jarred is saying camera manufacturers have been welcomed all this time to license RED's IP and implement compressed lossy raw in their cameras ?!

Why Canon for instance hasn't done this for the C-line ? Price? Pride?
(update: I meant why haven't they implemented lossy raw ratios such as 5:1 and greater if they could simply license that from RED)

Canon is a great camera for documentary/run+gun work and with its dual pixel AF can do things a RED can't as of now and you can easily capture 8h+ of footage per day on a doc, compared to narrative that's massive. Rawlite generates so much data so compressed raw is absolutely needed and relevant.
 
It's nice to see Canon coming to the party again [...] and it will be nice to see where ProRes RAW ends up because I think it will gain a lot of traction once whatever agreements are made.

Phil are you saying you expect Canon to implement ProRes Raw and open the door to compressed raw that way ?

Update: I meant lossy when I said compressed. Because it's mathematically lossless, in my mind Canon's rawlite is not really compressed raw.
 
Why Canon for instance hasn't done this for the C-line ? Price? Pride?

Sounds like you haven't looked at their camera lineup recently. A lot has changed for Canon in the last 2 years.

Canon's Cinema RAW Lite has been available from 2017 onward (C200) leading up to their most recently announced and likely all future cameras do in fact record compressed raw internally.

It's true the C-Series didn't launch with this codec, but something "magically changed somewhere" and they can do it now. Even on their upcoming DSLR and potentially Mirrorless bodies.
 
How relevant is a RAW compression ratio of more than 6 times these days, considering current flash prices and the need for high quality and flexibility in post?
This is a question

More than 80% of the blockbusters is still shot in uncompressed ARRIRAW (so it seems they don't give a sh*t about compressed RAW).
Just check the high budget($150,000,000+) blockbusters on IMDB

Sony has a patent aggrement with RED so they can do whatever they want.
In February 2013, Red filed for an injunction against Sony, claiming that several of its new CineAlta products, particularly the 4K-capable F65, infringed on patents the company held. They requested that Sony not only be forced to stop selling the cameras, but that they be destroyed as well. Sony filed a countersuit against Red in April 2013, alleging that Red's entire product line infringed on Sony patents. In July 2013, both parties filed jointly for dismissal, and as of July 20, 2013, the case is closed.

Canon has rawlite which has a lower compression ratio than the 6 times mentioned in the RED patent.
Straight from the RED patent: an image processing system configured to compress and store in the memory device the raw image data at a compression ratio of at least six to one and remain substantially visually lossless , and at a rate of at least about 23 frames per second

Canon C500mkII cinema RAW lite 5952 x 3140: 25,00P, 2.1 Gbps 12-bits Pure RAW is 5952 x 3140 x 16bits x 25 fps = 7.476 Gbps
7.476/2.1 ~ 3.56 hence less than 6 times.


Blackmagic has BRAW which doesn't infringe the RED patent but gives the same flexibility as RAW.
That is the reason they made BRAW.

Many who shoot RED and need the best possible quality use the lowest possible compression ratio (we do with 7k 4:1 24..25fps).
Since .R3D compression is based on JPEG 2000 wavelet compression it becomes lossy after a compression ratio higher than 2..3 in most cases.
With a confetti bomb it will be closer to 1 and with a static enviroment it will be higher than 3


The emerging markets (China, India, etc..) don't give a sh*t about US-only patents and they kind of don't like Apple(to put it mildly).
Even FoxNews acknowledges this and off course Trump who started to fight back

Fast flashstorage sell's for less than $200/TB with a read/write speed exceeding 3 GB/s.
https://www.newegg.com/p/1B4-00SK-000S5?Description=micron%209300&cm_re=micron_9300-_-1B4-00SK-000S5-_-Product
$708/3.84TB ~ $185/TB

Apple is kind of late to the compressed RAW party, so I think it will be hard for them to be as relavent with ProResRAW as they where with ProRes.
RedCodeRAW, cDNG, ARRIRAW, X-OCN, CineformRAW,KineRAW, SlimRAW, Canon RAW lite, etc.. where all earlier to the market than ProResRAW, I can't recollect any RAW format for motion picture that is released after ProResRAW.

No speculation whatsoever.

What points are you trying to make, saying high compression ratios is not in need?

Personally shoot with more than 7:1 compression a lot and for a lot of different reasons. High framerates, Less media used, faster playback on slower machines and drives, easier to move media, possible to play over slow network etc. Its very useful as I see it. My 8k Monstro is the king even if I just do a shitty facebook thing. 8k 16:1 compressed raw from monstro still beats most cameras out there if you ask me, and files size, media backups etc is not even a thing then. Still picture looks stellar.

And for all the motion content out there I can only assume that blockbusters stand for less than 0.1% of whats shot. And also I doubt they use Alexa due to the fact that it shoot uncompressed raw, Last time I asked, they confirmed they use it even though it shoots uncompressed. Simply they choose that camera for a lot of reasons, the no compression factor is not really the one point that makes or breaks the arri cameras.

6k and 8k raw is not something stupid any more, skate kids and everyone else wants it, but files size and data rates still matter. A lot of people still handle material on single spinning disc over USB2 or over slow networks. 10Ge is the bottle neck for most larger post facilities. If the file can not play trough such bandwidth their problems grow. So yes, highly compressed raw will as I see it be very much in demand for a long time forward, especially as the sensors get better and hold more resolution.
 
Speaking of Apple+Red, be interesting how r3d performs on the new 16" with Radeon Pro 5500 8GB.

Also, now with functional keyboard!
 
What points are you trying to make, saying high compression ratios is not in need?

How relevant is a RAW compression ratio of more than 6 times these days, considering current flash prices and the need for high quality and flexibility in post?
This is a question.

Personally shoot with more than 7:1 compression a lot and for a lot of different reasons. High framerates, Less media used, faster playback on slower machines and drives, easier to move media, possible to play over slow network etc. Its very useful as I see it. My 8k Monstro is the king even if I just do a shitty facebook thing. 8k 16:1 compressed raw from monstro still beats most cameras out there if you ask me, and files size, media backups etc is not even a thing then. Still picture looks stellar.

X.265 (HEVC) 4:4:4 12 bits lossless straigth from the camera gives you even more compression and also stellar looks and it is supported by both NVidia and AMD on their newest GPU's. Shooting 8k 16:1 over 5k 6:1 or 4k 4:1 for shitty facebook content will only give you more rolling shutter.


And for all the motion content out there I can only assume that blockbusters stand for less than 0.1% of whats shot. And also I doubt they use Alexa due to the fact that it shoot uncompressed raw, Last time I asked, they confirmed they use it even though it shoots uncompressed. Simply they choose that camera for a lot of reasons, the no compression factor is not really the one point that makes or breaks the arri cameras.

ARRIRAW is lossless and with HDE it stays lossless, it's just one of the many selling points of the ARRI cameras.

6k and 8k raw is not something stupid any more, skate kids and everyone else wants it, but files size and data rates still matter. A lot of people still handle material on single spinning disc over USB2 or over slow networks. 10Ge is the bottle neck for most larger post facilities. If the file can not play trough such bandwidth their problems grow. So yes, highly compressed raw will as I see it be very much in demand for a long time forward, especially as the sensors get better and hold more resolution.

The combination of 2001 USB 2.0 and a 2013 Red 6k is a weird combination knowing that USB 3.0 was introduced in 2011. For linus youtube stuff (shot at 8k 22:1) you just get extra rolling shutter for free and as a viewer in 4k, you won't see the difference in picture quality over youtube when something is shot with a helium at 8k 22:1 or 4k 5:1 (because of the youtube compression) the extra rolling shutter is sometimes visible.

I'm not saying that high compression is irrelevant but asking the following question: How relevant is a RAW compression ratio of more than 6 times these days, considering current flash prices and the need for high quality and flexibility in post?
 
X.265 (HEVC) 4:4:4 12 bits lossless straigth from the camera gives you even more compression and also stellar looks and it is supported by both NVidia and AMD on their newest GPU's. Shooting 8k 16:1 over 5k 6:1 or 4k 4:1 for shitty facebook content will only give you more rolling shutter.

What camera shoots 4:4:4 12-bit HEVC (at all) and in file sizes smaller than red's raw?
 
How relevant is a RAW compression ratio of more than 6 times these days, considering current flash prices and the need for high quality and flexibility in post?
This is a question.



X.265 (HEVC) 4:4:4 12 bits lossless straigth from the camera gives you even more compression and also stellar looks and it is supported by both NVidia and AMD on their newest GPU's. Shooting 8k 16:1 over 5k 6:1 or 4k 4:1 for shitty facebook content will only give you more rolling shutter.




ARRIRAW is lossless and with HDE it stays lossless, it's just one of the many selling points of the ARRI cameras.



The combination of 2001 USB 2.0 and a 2013 Red 6k is a weird combination knowing that USB 3.0 was introduced in 2011. For linus youtube stuff (shot at 8k 22:1) you just get extra rolling shutter for free and as a viewer in 4k, you won't see the difference in picture quality over youtube when something is shot with a helium at 8k 22:1 or 4k 5:1 (because of the youtube compression) the extra rolling shutter is sometimes visible.

I'm not saying that high compression is irrelevant but asking the following question: How relevant is a RAW compression ratio of more than 6 times these days, considering current flash prices and the need for high quality and flexibility in post?


To me shooting FF gives a certain look using a certain lens. cropping into 5k gives a different look with the same lens among other things. So no, more rolling shutter is not the only thing I get when shooting 8k and honestly I never really had rolling shutter problems, not even in 8k, when that happens camera is way to shaky or things moving to fast anyway.

And X.265 is not raw there is a difference.

Arri raw might be a selling point for arri, we work with arri material all the time and yes sure as the camera has such a shittly low resolution you better record uncompressed especially if mastering in 4k. We did some projects using their in camera prores upscale and yes quite disaster for chroma key. But its quite easy to turn that around and saying, ARRI not having compressed raw, is not really their key selling point, is it? Or do you think for one second that if arri added a compressed raw option that they would loose their high end clients?

Sorry ment to write usb3. What I mean is flash price does not really matter much. The hundreds of productions I came across the last couple of years none where having their rushes on SSDs. Pretty much all DIT´s except for TV series used single disks or delivering on small slow raid with less than 3 spinning drives or such to offload their rushes.

I dont care much what happens on youtube. To me capture compressions is relevant to what you intend to do with the matierlal in post, if there is a lot of chroma key, composting, colors that needs to be heavily twisted, low exposure levels, light colortemp far from balanced etc. Then yes low compression is needed. But if things are well exposed and you are not trying to make sunshine out of a rainy day and are happy with what you see in camera and possibly prefer a bit of texture, and dont have much movement in camera then yes compression can be cranked up quite high. 10:1 is normally no issue, I use it alot even when there will be loads of FX done to it.

As to me resolution is better than compression. I much rather have a 8k 8:1 than a 4k 2:1 Cmos capture to play with.
 
Another thread lost. All is lost.


To bring the topic of discussion back around. Anyone at Red got an ETA on a R3D FCPX Plugin update? :)
 
6k RAW (less rolling shutter) being a sweet spot for 4k/UHD rgb-delivery with a compression ratio between 3 and 6.
http://www.reduser.net/forum/showthread.php?179123-6K-Dragon-X-DSMC2-Petition!

Well I got that on monstro but for sure, 99% of what I do I prefer more pixels and the bigger sensor for most of the lenses I use. And what others use I dont care much about I done my testing. And not sure why you link to that thread. Ask those guys there if they would prefer a monstro instead of a 6k dragon and Im sure not a single one of them would turn down such offer if costs where the same. So no I dont see the logic in that comment. And yes, I did not see anyone in that thread that wanted to trade in their monstro for a 6k Dragon...

I used monstro and dragon side by side a lot. Mixing 6k and 8k and also shot both at 6k ana. To me the monstro chip is simply better, less noise, more usable range, more precision in the highlights. I dont know about re-draw time differences, but would not be surprised if thats better on monstro as well at same resolution. But again re-draw time I never seen as a big issue... and soon, very very soon there will be a some what fix for camera move introduced rolling shutter issues for red cams. Just wait think you will be surprised. :)
 
Well I got that on monstro but for sure, 99% of what I do I prefer more pixels and the bigger sensor for most of the lenses I use. And what others use I dont care much about I done my testing. And not sure why you link to that thread. Ask those guys there if they would prefer a monstro instead of a 6k dragon and Im sure not a single one of them would turn down such offer if costs where the same. So no I dont see the logic in that comment. And yes, I did not see anyone in that thread that wanted to trade in their monstro for a 6k Dragon...

I used monstro and dragon side by side a lot. Mixing 6k and 8k and also shot both at 6k ana. To me the monstro chip is simply better, less noise, more usable range, more precision in the highlights. I dont know about re-draw time differences, but would not be surprised if thats better on monstro as well at same resolution. But again re-draw time I never seen as a big issue... and soon, very very soon there will be a some what fix for camera move introduced rolling shutter issues for red cams. Just wait think you will be surprised. :)

Yep Panasonic has a Global Shutter (with a claimed DR of atleast 16 stops according to Andrew Reid, https://www.eoshd.com/2019/09/ibc-show-panasonic-8k-organic-sensor-prototype-camera-with-on-chip-nd-and-16-stop-dynamic-range/)

Monstro 8k and Gemini 5k are both very good sensors and better than Dragon 6k, if only they where 6k.
Still looking for a good 6k sensor camera within the S35/APS-C specifications, currently where are stuck with the Hulk/Helium sensor on 7k.
 
Sounds like you haven't looked at their camera lineup recently. A lot has changed for Canon in the last 2 years.

Hey Phil, it sounds like you might have skimped over my post. Two lines below the part you quoted I mentioned rawlite so I'm very well aware of its existence. :)

My understanding of RED's patent is they hold the rights for lossy compressed raw data (5:1 and greater), NOT for uncompressed or mathematically uncompressed/lossless ratios such as 3:1. Canon’s rawlite mathematically lossless 3:1 implementation seems to have allowed them to dodge that lossy bullet. If they were licensing RED’s IP then why wouldn’t they offer more flexible compression ratios such as 8:1 and greater right now? Their line is used extensively in doc work where higher compression ratios would be a godsend.

It sounds like Apple wants to expand Prores Raw, which to me implies broadening their compression ratio into the lossy 5:1 and beyond which would thus infringe RED’s patent and the reason why they petitioned it.
 
Phil are you saying you expect Canon to implement ProRes Raw and open the door to compressed raw that way ?

I see that I didn't use the right wording here, hence Phil's misinterpretation of what I was saying. I meant lossy when I said compressed. Because it's mathematically lossless, in my mind Canon's rawlite is not really compressed raw.
 
Yep Panasonic has a Global Shutter (with a claimed DR of atleast 16 stops according to Andrew Reid, https://www.eoshd.com/2019/09/ibc-show-panasonic-8k-organic-sensor-prototype-camera-with-on-chip-nd-and-16-stop-dynamic-range/)

Monstro 8k and Gemini 5k are both very good sensors and better than Dragon 6k, if only they where 6k.
Still looking for a good 6k sensor camera within the S35/APS-C specifications, currently where are stuck with the Hulk/Helium sensor on 7k.

In a way, the Red eco system of sensor sizes and resolution is pretty great. 5K is much more efficient in light capture and file sizes, yet delivers good 4K. Monstro can go full 8K or crop in s35 for more efficient needs. And helium is a great tool for high end s35 needs. Because motion capture is still largely super 35, many lenses are better there. It is good to have an 8K option in that realm.

6K or 5.7K for 4K delivery is great, but i prefer the low light and efficiency of 5K for most things, and feel that chasing the perfect downsample is not worth the trade offs for me. The 8K and 5K division of labor works well. Two specialized cameras, rather than one doesn’t do either efficiency nor extra high resolution.

Komodo is rumored to be 6K, so, there will be a simple go anywhere camera with perfect 4K output soon. But when 8K tv’s were released, and the realities of broadcast and internet, chasing the best raw resolution for HD or 4K became less crucial.
 
Last edited:
misinterpretation

Strangely in the audience of many who very much tagged along with the headlines and promoted the process of this endeavor for whatever reasons, I took the time to read the patent closely several times as well as the contesting documents, and was able to digest it more deeply. Which was funnier from my perspective earlier on because clearly people we fueled deeply by some other emotions.

Many are sighting the examples given in the patent in regards to what "at least" and the various compression ratio examples set forth, but the patent itself appears to cover all forms of compressed raw. I would say I could be wrong about that, but it appears to be the case.

The patent covers compressed RAW data and does reference specific frame rates and resolution targets, but he compression itself is not exactly "limited to" 6:1 throughout the patent.


In reference to what I would suspect Apple would want with ProRes RAW and truly what this patent contest was actually about. Remember Apple moves millions of phones, which is rather different than the thousands of cameras that hit our profession. Licensing at the scale is very different when it comes to financial ramifications. The strategy here is common in tech if you've been tracking the journey's of the largest companies on Earth. Apple and Samsung enjoy fencing with each other often for instance for just such things which either result in major payoffs, licensing, etc. So their effort here was likely squarely aimed at created an environment where this patent would be invalidated which would allow them to deploy ProRes RAW freely into their own devices and creating a large install base, then additionally eventually offering licensing to manufacturers to support ProRes RAW in the cameras themselves. This adoption would in turn make a defacto standard that Apple could very much have a great deal of control over in the process despite other forms of RAW likely existing if this patent was overturned. This isn't really speculating as the general path occurred already to rather successful levels where ProRes itself is very much an industry standard globally. I strongly feel they want to repeat that success with a mezzanine codec to a codec used for image acquisition.

In which case strategically speaking, fair game and a worthwhile drop in the bucket to give this a whirl as I've written a few times surrounding this topic.

However the patent is real and has been upheld before.

The fallout from this truly is now Apple in particular knows the path to ProRes RAW in their devices and likely other devices requires licensing, this is what all technology patents are all about. Which all companies are aware of and apparently some have worked out deals that we are aware and likely unaware of.


And to circle back to the validity of compression in motion picture filmmaking and to touch on some other points here. Yes, anything that can help save production dollars is valid.

RED - Implement REDCODE RAW, a variable compressed RAW codec. Owns the patent essentially on Compressed RAW for motion picture use.
ARRI - Focus on ARRIRAW, but introduce a post process of HDE via Codex, essentially a zipped ARRIRAW file
CANON - Found a way to get Cinema RAW Lite going and you'll here about straight Cinema RAW more soon
SONY - Their path is via add-on RAW modules and we don't know the net results of whatever negotiations they had other than that
BMD - Create a new codec that sort of works like RAW, but isn't RAW, which is attractive to their general audience base as it's easier to deal with than CinemaDNG (small, less intensive on a processing level) and at least something they can develop further
PANASONIC - Output a RAW signal to an external recorder like Atomos, which has publicly stated some sort of licensing agreement is going on
APPLE - Create ProRes RAW and try to get it integrated into the marketplace to expand adoption.

Bare in mind I'm close with most of these companies and what Jarred is stating is very true regarding a pathway to a compressed RAW licensed workflow. But if you can work around that, the obvious path to more profit is not dealing with that. And considering R&D and other licensed tech involved with a few of these cameras already, you don't want to always play it that way. And I'll say this, I do think what Canon is doing this round is very admirable because they are clearly licensing a few bits of tech in their cameras now to give people generally what they've been asking for for a long while now.


And to the point of the value of a compressed workflow. Just today I finalized a bid of 5 different workflows for a project filming in 2020. Spanning a rather long filming duration and cost per TB needed for the entirety of the shoot, the two extremes are: $57,832 versus a $289,161 total on expected used storage across the technologies they want to use on and off set for the general compressed versus uncompressed workflow in this case. i.e. The cost of the data footprint, which are estimates I have to put together often when bidding. This is one single camera project, though a decent sized one, when you start taking into account multiple productions or larger multi-cam shoots. Yes, this adds up. It's not uncommon to see much larger number than this on a project either. I'm going to leave the film costs out of this discussion, because they were looking to film 65mm 5-perf and it quickly got out of hand once considering the additional costs. Fine if you can write your own ticket, but when data ranges from 5 figures to 8 figures, you'll talking a whole different ballgame. In this case, chewing into the budget too much.


This general topic has been covered around the web a whole bunch. Now the real thing to watch is where ProRes RAW pops up next and whoever has new compressed RAW codecs coming out.
 
Back
Top