ProRes and Weapon

[Elsie N]

No worries.


Just a a tiny bit of red history.

.RED And .R3D is the same. The 3 acts as an inverted/reversed capital E

David, are you sure? I thought .RED was the codec devised for the Redray.

 

[David Battistella]

Perhaps that is how they were differentiating the R3D from the REDRAY encoded file. I had high hopes for REDRAY, alas I think it will be giving way to h.265. Better to focus on cameras it think.

 

[Elsie N]

It just seemed that codec worked so efficiently that it would beat everything being offered at the present. I would love a camera that outputted R3D and .RED simultaneously and have that codec released into the wild for viewing purposes. .RED would make Prorez obsolete, IMO.

 

[Randall Weatherington]

It would be really interesting to see if RED submitted the .RED player codec to SMPTE, ATSC, et all to see if any part of it or any variation on it was used when they were establishing compression and broadcast standards.

 

[David Battistella]

It just seemed that codec worked so efficiently that it would beat everything being offered at the present. I would love a camera that outputted R3D and .RED simultaneously and have that codec released into the wild for viewing purposes. .RED would make Prorez obsolete, IMO.

You are totally entitled to that opinion but I can't say I agree. i doubt very much that Prores could be easily unseated. Prores is viable on the entire mac, and now moreso windows platform and has now been firmly entrenched as a production standard on many a-list productions, not to mention the entire trickle down through the whole TV-commercial, broadcast industry. And it's getting better and even RED has licensed it for use in their weapon cameras because of industry demand.

maybe if r3d had been licensed to SONY or ARRI that could have flipped the trend.

 

[Elsie N]

You are totally entitled to that opinion but I can't say I agree. i doubt very much that Prores could be easily unseated. Prores is viable on the entire mac, and now moreso windows platform and has now been firmly entrenched as a production standard on many a-list productions, not to mention the entire trickle down through the whole TV-commercial, broadcast industry. And it's getting better and even RED has licensed it for use in their weapon cameras because of industry demand.

maybe if r3d had been licensed to SONY or ARRI that could have flipped the trend.

I'm not saying it would happen overnight, but once productions, tv stations etc. learned they could get a .RED stream straight from the camera and play it back @ what? 10, 20, 30 ?bits, (couldn't remember what the stream entailed but I remember it was small even compared to H264 and I'm sure it is smaller than Prorez.)... I think those with RED cameras as their acquisition tool would happily embrace that form if there was no additional cost to them.

R3Ds are still the standard for post, but for shoot and playback, .RED is the berries. That would make the RED cameras the defacto ENG camera, IMO.

But it may just be impossible for the camera to do a .RED stream as it may require a ton of encoding that the camera may be unable to handle.

 

[Mathieu Ghekiere]

Do we already know in the meanwhile if Weapon will have 4K Prores or only 2K Prores? I missed confirmation on that?

Regards,

 

[Jacek Zakowicz]

I'm pretty sure that it compresses the debayered monitoring signal so yes- 1080p at this point. If RED outputs UHD or 4K on the HDMI port it will go up to that...

 

[Tarquin Cardona]

I am in total agreement with David on this. Pro Rez will be extremely important ideally at 4KUHD minimum for both Magnesiuum and Carbon - esp as Magnesium buyer more likely to be content or narrative oriented and more likely to benefit. It is pointless to discuss industry making some shift to R3D in any time to be meaningful in the near term.

 

[Elsie N]

I am in total agreement with David on this. Pro Rez will be extremely important ideally at 4KUHD minimum for both Magnesiuum and Carbon - esp as Magnesium buyer more likely to be content or narrative oriented and more likely to benefit. It is pointless to discuss industry making some shift to R3D in any time to be meaningful in the near term.

The discussion was about .RED & Pro rez... not R3D.

 

[Tarquin Cardona]

You are totally entitled to that opinion but I can't say I agree. i doubt very much that Prores could be easily unseated. Prores is viable on the entire mac, and now moreso windows platform and has now been firmly entrenched as a production standard on many a-list productions, not to mention the entire trickle down through the whole TV-commercial, broadcast industry. And it's getting better and even RED has licensed it for use in their weapon cameras because of industry demand.

maybe if r3d had been licensed to SONY or ARRI that could have flipped the trend.

I took my r3d comment from this. In any case - point is importance of good quality pro rez to give these new cameras better odds of changing hearts and minds. That is all.

 

[Björn Benckert]

You all know Prores XQ can hold 16 bit....

So the speculations here is what is the better option between:

1. r3d raw compressed capture (at lets say 8k or 6k) then downscale to UHD or even 2k.

or.

2. Dragon sensor material (probably going in a different path, not trough r3d) down sampled before compression and then saved into a 16 bit proress file.


Not many of us has done proper testing between the two paths. But according to my testings with other material than RED. I think option one will be great for a lot of situations but option 2 will for sure not only be more simple to handle for many but also more importantly render far better end results. The difference is where and how heavy compression kicks in if it happens before or after downsampling.

So I think this new weapon feature (2k now and what sound like soon 4k) is not so much about RED trying only to please the quicktime editors requests. I think it´s just as much or more a question of noise levels, compression artifacts and reaching the best possible end result. Both the 6k and 8k sensor is capturing so much info so downscaling before compression starts to make sense for UHD and 2k finishing (basically you get a cleaner image if you downscale before compression and the compression does not bite as hard). Still having the r3d option is usefull for a lot of things, rescaling above 4k resolution, highspeed etc.

But this is ony speculation ofcourse as not many of us has seen a 16 bit proress XQ file straight from the weapon/dragon sensor. If it´s done right it should absolutly not be compared to shit 10bit ProressHQ and especially not ProressHQ made from a compressed r3d.... straight from the source, downsampled and feed into a XQ should be something completely different from what you seen coming out of your Red cameras so far.

but again this is 11 pages of speculation about something most of us only heard very little about. Probably 16bit, probably UHD, probably straight from the sensor without precompresison... but if all that kicks in...

 

[Elsie N]

I took my r3d comment from this. In any case - point is importance of good quality pro rez to give these new cameras better odds of changing hearts and minds. That is all.

Hah! It's been so long since I visited this thread I forgot how it started. You are correct.

 

[shashbugu]

2k Prores Scaled from 6k Weapon, can easily be scaled to UHD. But if you already have the 6k r3d raw material its very quick to export to Prores 16bit XQ 4k UHD. Whats not to love.

I think 95% of Weapon users will be very surprised and content with scaled Prores 2k. But as its been mentioned before by Jarred 4k prores is being worked on.

R3d's are great, but trust me scaled 2k prores even at 422HQ will work very well for many TV shows, documentaries and run and gun shoots

 

[Björn Benckert]

2k Prores Scaled from 6k Weapon, can easily be scaled to UHD. But if you already have the 6k r3d raw material its very quick to export to Prores 16bit XQ 4k UHD. Whats not to love.

I think 95% of Weapon users will be very surprised and content with scaled Prores 2k. But as its been mentioned before by Jarred 4k prores is being worked on.

R3d's are great, but trust me scaled 2k prores even at 422HQ will work very well for many TV shows, documentaries and run and gun shoots

But the path: r3d to proress 16 bit XQ 4k UHD is less than 1/4 of the quality of a XQ UHD made without the r3d path and usually far less than that... no exact numbers but if you record 1:7 redcode and save a 8k file then a XQ UHD file from the sensor would be about (Im speculating) like 21 times less noise. So if it works that way then it´s not about "its good enough" it will actually have qualities an r3d just can not provide.

 

[David Battistella]

But the path: r3d to proress 16 bit XQ 4k UHD is less than 1/4 of the quality of a XQ UHD made without the r3d path and usually far less than that... no exact numbers but if you record 1:7 redcode and save a 8k file then a XQ UHD file from the sensor would be about (Im speculating) like 21 times less noise. So if it works that way then it´s not about "its good enough" it will actually have qualities an r3d just can not provide.

If you are correct on that then we may be seeing the beginning of the end of R3D, unless we see a new R3D format with a better and faster compression algorithm within the codec (which is possible) and could allow for a realtime r3d file. Call it R3D 2 if you like.

 

[Björn Benckert]

If you are correct on that then we may be seeing the beginning of the end of R3D, unless we see a new R3D format with a better and faster compression algorithm within the codec (which is possible) and could allow for a realtime r3d file. Call it R3D 2 if you like.

Maybe but. The point is with the high oversampling that 6k and 8k gives even for 4k finnishing. It will be far from allways that the best is to go with compression before downsampling. Offcourse if 8k 1:1 r3d came possible it would only be suffering in terms of large data load. for above 4k work, pulling stills, zoom in/pan in scan etc r3d is for sure very handy but the benefits mentioned with down scaling before compression has other benefits. Alot of people have seen that in lowlight situation etc their HD recorders acctually can capture and create a cleaner HD master than r3d to HD proress... and the HD output is 8bit... and offcouse only HD, and only coming from at most a 6k source... then think 8k and 4k 16bit proress with a good downsampling and no compression before that... the lowlight end of the spectrum will for sure benefit from the way weapon is built... if it´s built as I think. If it makes proress from a downscaled r3d then, yes offcourse it´s no better (only time saving) than creating the proress in post/off camera. But I hope / think red has done more than that.

And soon we will know, when is the weapon start shipping?

 

[shashbugu]

But the path: r3d to proress 16 bit XQ 4k UHD is less than 1/4 of the quality of a XQ UHD made without the r3d path and usually far less than that... no exact numbers but if you record 1:7 redcode and save a 8k file then a XQ UHD file from the sensor would be about (Im speculating) like 21 times less noise. So if it works that way then it´s not about "its good enough" it will actually have qualities an r3d just can not provide.

I'm no scientist Bjorn, but raw is raw, i doubt that you'll be able to manipulate a Prores clip like you would an R3D. You really think the quality would be that bad? We have seen 2k Alexa intermediates blown up to 4k DCP's. it looks pretty good. I believe a Weapon 6K or 8K image should be better. What am I missing here?

 

[Michael Tiemann]

You all know Prores XQ can hold 16 bit....

So the speculations here is what is the better option between:

1. r3d raw compressed capture (at lets say 8k or 6k) then downscale to UHD or even 2k.

or.

2. Dragon sensor material (probably going in a different path, not trough r3d) down sampled before compression and then saved into a 16 bit proress file.[...]

I have a really difficult time decoding what the issue was until I did some digging and found that REDCODE encodes linear 16-bit data to a 12-bit value. The documents I've found validating the 12-bit limit of REDCODE all predate the DRAGON sensor. In any case, I think Björn's point is that dropping from 16 bits to 12 bits and then compressing and then downscaling is likely to be worse (by a lot?) than staying at 16 bits through the chain.

I've done my own pixel-peeping at a slightly different question, which is whether it is better to shoot a smaller format (UHD) with lower compression (7:1) or a larger format (6K) with higher compression (16:1) and then down-sample, and the answer is "it depends". But what I don't understand is if we keep the bit depth constant and we keep the overall data rates consistent, are we still talking about whether its best to compress first and then downscale or downscale and then compress? If so, I don't really know the answer. I can believe that scaling 16 bit data to fit a 12 bit container is going to be lossy, especially in the shadows.

 

[Tarquin Cardona]

I think a lot of this tech talk gets a little wide of the mark. The real ask I believe is something akin to why people love other systems so much. They want the camera to be easier to use and to deliver pleasing good quality pro rez etc out of the gate. In others words the camera actually does some of the work and does not leave it all on the user. For some folks the intricacies of how many K at what ratio and blackshade and temp etc is great. For others it feels like the Ikea build has extended to the look itself and its all on the customer to put it together properly. For me that is the epicenter of the issue. You don't want to miss a moment or performance or what have you because the camera is needier than the talent. The idea that with Weapon Red will really step up and make it easier to get the potential from the sensor and system is the way forward. To me thats what makes the Weapon worth the money.