Blackmagic Forum

Blackmagic, can we please have BRAW for 4K DCI and 4.6K on the Pocket 6K please? I believe it is possible with some cropped. I don't mind if there is some crop factor. We need it to complement the 4K DCI and 4.6K BRAW on the UMP G2. Thanks,

Hi Ellory,

Since the pixel pitch of the Pocket Cinema Camera 6K is smaller, the 4K DCI crop on the Pocket Cinema Camera 6K will actually be more cropped in than 4K DCI on the Pocket Cinema Camera 4K.

We've maximized the formats on the Pocket Cinema Camera 6K for the greatest FOV and frame rates possible, so wouldn't it be better to shoot at 5.7k which is 17:9 (to match 4K DCI's aspect ratio) and also get the benefit of oversampling when delivering in 4K DCI?

Yes, I think it would Kristian. Grand idea.
Cheers

That was my fallback - 5.7K. The 5.7k bitrate for BRAW just uses more storage than if it were @ 4K DCI, no? Hence my request. Good idea though. Thanks.

Ellory Yu wrote:That was my fallback - 5.7K. The 5.7k bitrate for BRAW just uses more storage than if it were @ 4K DCI, no? Hence my request. Good idea though. Thanks.

5.7K (B)RAW will give you a perfect de-bayered 4k-DCI output (RGB, what you see on your TV, screen or in the cinema). 4K (B)RAW won't give you a perfect 4k-DCI output.

Commen de-bayer algorithms need a factor of 1.4 (5.7K / 1.4 ~ 4K).

When you want smaller file sizes, use a higher compression ratio or shoot ProRes Proxy.

MishaEngel wrote:

Ellory Yu wrote:That was my fallback - 5.7K. The 5.7k bitrate for BRAW just uses more storage than if it were @ 4K DCI, no? Hence my request. Good idea though. Thanks.

5.7K (B)RAW will give you a perfect de-bayered 4k-DCI output (RGB, what you see on your TV, screen or in the cinema). 4K (B)RAW won't give you a perfect 4k-DCI output.

Commen de-bayer algorithms need a factor of 1.4 (5.7K / 1.4 ~ 4K).

When you want smaller file sizes, use a higher compression ratio or shoot ProRes Proxy.

I hear there is no Prpres flavor for those resolutions.

Michael Odhiambo wrote:

MishaEngel wrote:

Ellory Yu wrote:That was my fallback - 5.7K. The 5.7k bitrate for BRAW just uses more storage than if it were @ 4K DCI, no? Hence my request. Good idea though. Thanks.

5.7K (B)RAW will give you a perfect de-bayered 4k-DCI output (RGB, what you see on your TV, screen or in the cinema). 4K (B)RAW won't give you a perfect 4k-DCI output.

Commen de-bayer algorithms need a factor of 1.4 (5.7K / 1.4 ~ 4K).

When you want smaller file sizes, use a higher compression ratio or shoot ProRes Proxy.

I hear there is no Prpres flavor for those resolutions.

ProRes 4k from a 4k bayer sensor is soft(upscaled)
ProRes 4k from a 4 x 1.4(and up) K bayer sensor is tack sharp.
ProRes 6k from a 6k bayer sensor is soft

Since 6k is not a delivery format who cares.

BRAW is always prefered over ProRes.

So sorry for my ignorance but when de-bayering a 5.7k raw image for 4k is the aspect ratio still the same. What I mean by that is: does a 17:9 5.7k de-bayered image produce a tack sharp 17:9 4k image? (clearly I need to study up on post production file management and de-bayering techniques)

I am about to start a long term project abroad and I have both the P4k and P6k cameras, I was planing on making my P4k the Gimbal and B-Roll camera but if possible I would like to make sure the quality matches as well as possible to each other. So any suggestions are welcomed!

Yes is does, both should be the same aspect ratio, 17:9 of the respective windows are 17:9.
Cheers

Any Bayer sensor can't deliver the same resolution its pixel count is suggesting.
The best algorithms hardly reach 80%. So, oversampling is a good idea if you want the best quality, which includes downsampling to your delivery format. If this is better in the respective camera or in post should be tested. But leaving it to post gives you the opportunity to reframe if needed.

Uli Plank wrote:Any Bayer sensor can't deliver the same resolution its pixel count is suggesting.
The best algorithms hardly reach 80%. So, oversampling is a good idea if you want the best quality, which includes downsampling to your delivery format. If this is better in the respective camera or in post should be tested. But leaving it to post gives you the opportunity to reframe if needed.

Thx for the replies Denny and Uli!
So, in the case of the large project I am about to start (shooting a feature doc abroad in Sri Lanka) do you have any "hypothetical" starting points for workflows using both the 6k and 4k pocket cameras if my goal is to deliver a 2k DCP for fests / theatrical.

Again I was thinking of using the 4k for lock-down shots and gimbal work as that would likely need less stabilization in post and the 6k camera for more handheld and "key" shots that I may want to re-corp or futz with latter. But assuming the P4k is also a Bayer sensor, from this thread I would also need to assume that the 4k sensor probably delivers a tack sharp de-bayered image around image around 3k (is that about right) giving me a bit of wiggle room for stabilizing and re-composing in post but not nearly as much as the P6k.

Then, I just noticed @MishaEngel that your breakdown of de-bayered footage is for ProRes so before I go too deep in the weeds with research I am assuming this also relates to BRAW as well as ProRes as I do intend to be shooting BRAW for the entire project. -- thanks all!

Shawn Convey wrote:

Uli Plank wrote:Any Bayer sensor can't deliver the same resolution its pixel count is suggesting.
The best algorithms hardly reach 80%. So, oversampling is a good idea if you want the best quality, which includes downsampling to your delivery format. If this is better in the respective camera or in post should be tested. But leaving it to post gives you the opportunity to reframe if needed.

Thx for the replies Denny and Uli!
So, in the case of the large project I am about to start (shooting a feature doc abroad in Sri Lanka) do you have any "hypothetical" starting points for workflows using both the 6k and 4k pocket cameras if my goal is to deliver a 2k DCP for fests / theatrical.

Again I was thinking of using the 4k for lock-down shots and gimbal work as that would likely need less stabilization in post and the 6k camera for more handheld and "key" shots that I may want to re-corp or futz with latter. But assuming the P4k is also a Bayer sensor, from this thread I would also need to assume that the 4k sensor probably delivers a tack sharp de-bayered image around image around 3k (is that about right) giving me a bit of wiggle room for stabilizing and re-composing in post but not nearly as much as the P6k.

Then, I just noticed @MishaEngel that your breakdown of de-bayered footage is for ProRes so before I go too deep in the weeds with research I am assuming this also relates to BRAW as well as ProRes as I do intend to be shooting BRAW for the entire project. -- thanks all!

Always shoot 4k BRAW with the BMPCC 4k for a 2k RGB output
With the BMPCC 6k you have the option to go al the way down to 2.8k BRAW for 2k RGB output.

When the (B)RAW pixel ratio is 1.4 or higher than the delivery you are good to go.
BRAW is a lighter codec(for a computer) than ProRes at the same quality levels.

Good advice from Misha. I’m shooting 4K or UHD 1.33x anamorphic to deliver 2K or HD widescreen respectively. 2K/HD still rules most clients needs. When they start expecting 4K deliverables, the 6K/8K camera will be in high demand.


Sent from my iPhone using Tapatalk

MishaEngel wrote:Always shoot 4k BRAW with the BMPCC 4k for a 2k RGB output
With the BMPCC 6k you have the option to go al the way down to 2.8k BRAW for 2k RGB output.

When the (B)RAW pixel ratio is 1.4 or higher than the delivery you are good to go.
BRAW is a lighter codec(for a computer) than ProRes at the same quality levels.

Thx! But just to be clear the BMP6k can not shoot 2.8k BRAW as of yet, so I am assuming you are suggesting that IF BMD makes a firmware update for the BMP6k to allow for 2.8k BRAW then that would pair beautifully with the BMP4k (shooting 4k) and the BMP6k (shooting 2.8k or greater) for a 2k delivery.

Or am I still confused (which is quite possible)

Shawn Convey wrote:Thx! But just to be clear the BMP6k can not shoot 2.8k BRAW as of yet, so I am assuming you are suggesting that IF BMD makes a firmware update for the BMP6k to allow for 2.8k BRAW then that would pair beautifully with the BMP4k (shooting 4k) and the BMP6k (shooting 2.8k or greater) for a 2k delivery.

Hi Shawn,

The Pocket Cinema Camera 6K can absolutely shoot 2.8K in Blackmagic RAW up to 120 fps.

Kristian Lam wrote: Hi Shawn, The Pocket Cinema Camera 6K can absolutely shoot 2.8K in Blackmagic RAW up to 120 fps.

Yup, My bad... I guess my brain shut off when I turned on the camera and saw there was no 4k and 1080 options it somehow failed to see that lit up 2.8k option... OK so I'm gunna hit the books and try to make sense of all of this but I guess if I can ask one more "dummy" question...

Question - the 2.8k option on the P6k - after De-Bayering is it a straight 2k image or is there some wiggle room for say stabilization? Obviously at 2.8k I wouldn't imagine re-framing will be an options... and that leads me to a bit of a "sub-question" where I get confused a bit... if 2.8k is perfect for a 2k delivery (after De-Bayer) then why not give user the option for a 4k BRAW experience that would allow for a bit more wiggle room for said stabilization and re-composing in post (again after de-bayer) or am I still thinking about this all wrong?

Thanks everyone for putting up with my low science knowledge

My take is a little bit different on this ...

I offloaded the P4K and am using the P6K exclusively ... also dropped a 1DC and opted not to buy
a C300II/C200. Thank you Blackmagic!

Since 5.7K 4K DCI and 2K DCI are all 17:9 ... I just record in 5.7K and render on a 4K DCI timeline in Davinci
Resolve. But the majority of the time I output 2K DCI ProRes 422... all from a full width sensor capture. Only if
I needed 120FPS would I need a windowed capture ... or anamorphic.

6K UHD and HD are all 16:9 so a similar workflow could be followed ...

Yes the initial data rates and storage requirements are more but I tend to aggressively edit and discard excess footage. One can archive as Prores HQ or BRAW as desired.

The quality of the 6K files is compelling ... and not just to re-compose or stabilize. Sharpness color integrity and depth exceed 4K capture ... even when rendering as a 2K file.

Bob, easy for me to understand that sharpness is greater, but colour and saturation too? The Colour Science 4 is the same, but the sensor is different and BRAW is tuned to each camera. This is intriguing as we all love the best Colour we can produce.

We have a few folks here that own or will own both cameras. I’d appreciate seeing a careful shootout using a variety of codecs on the same subject under the same conditions, both cameras without Colour Correction. Does anyone think they’ll take on that task?


Sent from my iPhone using Tapatalk

My intent was not to compare the two sensors or cameras .... just 5.7 to 4K vs 4K capture.

Just thought that oversampling the 5.7 to 4K would be analogous to rendering HD from 4K where
a greater color depth ... 420 to 422 is claimed. Also Canon and Panasonic suggest that larger sensor
than 4K renders better sharpness and color. Saturation was not part of my calculation ... that is set within
V4 Colour Science and subject to manipulation in post.

I stand corrected.

Thanks,

Bob

When you shoot 5.7k you have 5744h x 3024v = 17,369,856 bayer pixels
8,684,928 Green pixels 4,342,464 Blue pixels and 4,342,464.

4k DCI is 4096 x 2160 = 8,847,360 RGB pixels = 26,542,080 total pixels(8,847,360 Green pixels, 8,847,360 Red pixels and 8,847,360 Blue pixels).

Test have proven that with 5.7k bayer pixel you can get a perfect de-bayer to 4k RGB pixels(4kR, 4kG and 4kB), whit a normal/fast de-bayer algorithm.
Some claim that they only need a factor of 1.25 instead of 1.4 to get a PERFECT de-bayer, till this day I have never seen scientific proof of this claim (theory + test).

Slashcam.de have just tested the BMPCC6k and it gives a PERFECT de-bayer to 4k DCI (as does the 6k Sony Venice).

The RED 5k(5,120 horizontal) and the UMP G2 4.6k(4608 horizontal) give a ALMOST perfect debayer in Resolve.

Yes, rereading Bob’s post about better colour on the 6K, I see he’s downsampling when he is coming to that conclusion. No argument there, Bob. Sorry I had misinterpreted your conclusions about 6K.


Sent from my iPhone using Tapatalk

I will try to be a bit more articulate with my posts ... easily misconstrued.

Thanks,

Bob