12 bit RED RAW - 12 bit DPX ?

[Rainer Fritz]

Was just a question today if there will be in future the possibility to get out of the RED RAW Converters 12bit or 16bit DPX Files ?

thx
rainer

 

[Mike McCarthy]

As I understand it, the Cineon and DPX formats are limited to 10 bits, although there are many current variations that aren't covered in the standard, so maybe its possible, but they wouldn't be universally compatible. (For example-Speedgrade can export DPXs with alpha, but AE can't read them)

 

[Miguel "Macgregor" De Olaso]

If you want 16bits, just go with with TIF or OpenEXR.

 

[jake blackstone]

As I understand it, the Cineon and DPX formats are limited to 10 bits, although there are many current variations that aren't covered in the standard, so maybe its possible, but they wouldn't be universally compatible. (For example-Speedgrade can export DPXs with alpha, but AE can't read them)

There are 12 and 16 bit versions of DPX and everybody supports them...

 

[Rainer Fritz]

There are 12 and 16 bit versions of DPX and everybody supports them...

Thats the reason why I'm asking, because I want to bring the full 12 bits of the cam in the CC and then to the DCP. Digital Cinema Standard is JPEG2000 with 12 bits. Why not using them.....
It is not always possible to conform direct RAW in the CC, so I want to use the full color depth from RAW to DPX.

thx
rainer

 

[jake blackstone]

Thats the reason why I'm asking, because I want to bring the full 12 bits of the cam in the CC and then to the DCP. Digital Cinema Standard is JPEG2000 with 12 bits. Why not using them.....
It is not always possible to conform direct RAW in the CC, so I want to use the full color depth from RAW to DPX.

thx
rainer

You can also use 16 bit TIFF for DCI, but you're just trying to reinvent the wheel. Not all bits are created equal. Just because you get 12 bits from the RED camera doesn't mean, that you'll get more precision, than using a 10 bit LOG DPX format. The trick is the proper transcoding. If you don't do it correctly, the data is gone forever and you wouldn't be able to recover unless you'd go back to r3d. And keep this in mind, that at some point in the process, your data will have to be mapped into a LOG curve, as DCI specifies 2.6 gamma and D54 color temperature at 14 ft/l. That is why film normally scanned as 10 bit LOG DPX and then graded with the properly built P3 as a target LUT on a calibrated screen. This way while you're grading you have a WSWG image representation on the screen and at the end you just need to render it out either as a JPEG2000 or 16 bit TIFF.

 

[M Most]

...DCI specifies 2.6 gamma and D54 color temperature at 14 ft/l.

Not to be nitpicky (although I'm about to be anyway), but the white point specified for projection in the DCI specification is actually .3140 in x and .3510 in y, which is about 6300K, not 5400, and better matches the design of Xenon lamp based digital cinema projectors. There is also no such thing as "D54," as 5400K is commonly referred to as D55.

Also, just to be more specific, the DCDM - the common delivery element for digital cinema I think you're referring to - is generally specified to be uncompressed, 12 bit TIFF, encoded in X'Y'Z' color space with 2.6 gamma. When the DCP is made, JPEG2000 compression is performed, the "essence" is wrapped in an MXF wrapper, and, optionally, encryption is applied.

 

[jake blackstone]

Not to be nitpicky (although I'm about to be anyway), but the white point specified for projection in the DCI specification is actually .3140 in x and .3510 in y, which is about 6300K, not 5400, and better matches the design of Xenon lamp based digital cinema projectors. There is also no such thing as "D54," as 5400K is commonly referred to as D55.

Also, just to be more specific, the DCDM - the common delivery element for digital cinema I think you're referring to - is generally specified to be uncompressed, 12 bit TIFF, encoded in X'Y'Z' color space with 2.6 gamma. When the DCP is made, JPEG2000 compression is performed, the "essence" is wrapped in an MXF wrapper, and, optionally, encryption is applied.

Those are not nitpicks and I do appreciate corrections, when they are warranted. So, allow me to explain a source of confusion or misunderstanding. Apparently, there is a D54 white point, but I didn't realize, that it's not as common as D55. D54 is described by CineSpace in the user manual and it's described as white: xy[0.3349, 0.3497](D54). I was under impression, that this described P3 chromaticity, but it is white: xy[0.3140, 0.3510] as you had properly pointed out. So, my bad...So with this out of the way, let's move onto Digital Source Master (DSM). That is what I was talking about. This is from DCi format description errata. "The DCDM gamma-encoded X', Y', and Z' color channels are represented by 12-bit unsigned integer code values. These 12 bits are placed into the most significant bits of 16-bit words, with the remaining 4 bits filled with zeroes" And this is from Filmlight:"Currently the preferred exchange format is 12 bit colour embedded in a 16 bit TIFF file". And that was my understanding as well. That is why I suggested 16 bit TIFF format for DSM. I wasn't talking about DCDM.
So, I hope this clears up any confusion

 

[M Most]

...Digital Source Master (DSM). That is what I was talking about.

The Digital Source Master, in most cases, is represented by 10 bit Cineon Log DPX files that are used for the film recording. The digital cinema version is derived from this by baking in a proper LUT, using 12 bit precision, and ultimately converting to X'Y'Z' color space. So I suppose if one is talking purely about the Digital Cinema deliverable, what you're saying could be true. However...... a number of projects are now being done targeting Digital Cinema as the primary deliverable. This is true of almost all animation projects, but some live action projects are being finished this way as well. In that case, an inverse LUT is being used to derive the Cineon Log files for the film recorder.

You are correct about the 16 bit TIFF file, I was talking about the actual information. Sorry for the confusion, my bad.