C500: When will Canon give us 12 bit 4k XF-AVC or 120fps XF-AVC?

L

Lkorver

Well-known member
Any chance the C500mk2 could in the future, be updated shoot 4k/12bit in XF-AVC? It's unfortunate that the choice is limited to either 12bit RAW or 10bit XF-AVC. There seems to be some middle ground to work with here. Would this be something they save for the "c300mk3".

And what's the likelihood that the C500mk2 would ever be updated via firmware to shoot 4k @ 120fps (XF-AVC 10bit?). Does it have the horsepower to do so?

I shoot a ton of solo doc work, so CRM has been overkill for what I do with the c200 right now, and a real burden in terms how quickly cards are filling up throughout the day and drives are filling up at night. Give me a camera that can shoot 12bit 4K in XF-AVC and i'll be a happy camper. 120fps 4K would be amazing as well!
 
The bit rates/file sizes for the 12-bit XAVC and Canon Raw Light are about the same.
 
Do you really think you'd see an appreciable difference? Do you really push your grades around hard?
I shoot mostly RAW with our C200 and I can barely tell the difference overall between the 12bit 24p and the 10bit 60p.
 
I’ve settled on lossless 10 bit Cineon log compression workflow via Slimraw for 12 bit uncompressed CDNG off loads from my Bolex. Same format used for film negative scans. There is no difference in DR or grading range between the two. Everything is recoverable. I’ve graded comparison shots side by side in Resolve. For REC709 or HLG REC 2020 HDR color managed workflow the end result is identical.

What good would 12 bit video be when most TV’s still use 8bit+FRC to show 10 bit color values?
 
Last edited:
Compressing raw is more efficient than debayering and then compressing all of the new extrapolated data created by the debayering process. This is why Canon added the 12 bit 444 2K to the C300 MK II which was actually an efficient use of the 4K sensor and had a similar data rate to CRM at 2K. I actually wish Canon would add a LongGOP to the C500 mK II for interviews and events.
 
DLD said:
The bit rates/file sizes for the 12-bit XAVC and Canon Raw Light are about the same..
Click to expand...

This surprises me, but I’m not a math wizard so I’ll take your word for it :)

To my other question, would 96fps or up to 120fps @ 10bit be attainable with the current hardware canon is offering with the c500mk2 via cfast express?
 
Last edited:
Lkorver said:
This surprises me, but I’m not a math wizard so I’ll take your word for it :)

To my other question, would 96fps or up to 120fps @ 10bit be attainable with the current hardware canon is offering with the c500mk2 via cfast express?
Click to expand...

12 bit 4:4:4 in either RGB or YUV records 3 times the uncompressed data of a 12 bit Bayer raw format. R, G, and B pixel values for the full frame resolution vs monochrome luminance values only for raw. Compression determines final file size but it takes 3 x compression for a gamma corrected video file to equal the size of an uncompressed raw file at the same bit depth.
 
Lkorver said:
This surprises me, but I’m not a math wizard so I’ll take your word for it :)

To my other question, would 96fps or up to 120fps @ 10bit be attainable with the current hardware canon is offering with the c500mk2 via cfast express?
Click to expand...

Even if the hardware did support it, probably never happening. If you want wild promises about new features coming in FW, Sony actually plays that game, Canon doesn't. Buy a Sony, they're promising all kinds of new things in 2020 and 2021 for the FX-9.
 
A 12-bit 4k 4:4:4 compressed file, downsampled from the 6k sensor would be grand. And possibly around 1Gbps compared to 2Gbps for the 6k raw (which is a big deal for day-to-day data loads).

I think there's close to zero chance Canon will provide it, but I agree, it'd certainly be nice to have. The general consensus seems to be that Canon footage is noticeably nicer in 12-bit than 10-bit.

I forgot to check it when I went to see the camera. But I wonder if it's possible to shoot 24/25fps in the camera's slow/quick motion mode (which is 10-bit raw, and 1Gbps-realtime in 60fps)? I suspect you'd lost audio sync or something like that, but if the compression ratio remains the same as it is for 60fps, you'd cut the datarate in half. And 10-bit logarithmic raw is supposedly equivalent to 12-bit linear raw. That'd be a handy half-way measure.
 
Lkorver said:
This surprises me, but I’m not a math wizard so I’ll take your word for it :) ...
Click to expand...
I looked it up last night. It's 960 Mbps for the 12-bit and 1 Gbps for the CRL.

CRL isn't a "real" Raw but it plays one on TV. The 6K bit rate/file sizes ought to be twice of 4K. Thus, you need the CFexpress cards.

And, as mentioned above, Canon's MO is to drop the price on a given camera as it ages and release a followup model with improved specs. How it plays out in the EF vs. the RF in-house battle is yet to be seen.
 
Grug said:
A 12-bit 4k 4:4:4 compressed file, downsampled from the 6k sensor would be grand. And possibly around 1Gbps compared to 2Gbps for the 6k raw (which is a big deal for day-to-day data loads)..
Click to expand...

If the c300mk2 can do 4k XF-AVC 10bit @410mb/s, wouldn't the c500mk2 be able to provide the same codec at 410mb/s? It's listed as 810mb/s in the spec sheet for the C500mk2. Does the larger sensor preclude the c500mk2 from matching the c300mk2 offerings?

To that end, the c200 4k options are limited to 8bit XF-AVC (150mb/s) or RAW Light (1GB/s). Would 4k 10bit XF-AVC at 410mb/s be doable with this camera? If so, that would fill a nice gap for me.
 
DLD said:
CRL isn't a "real" Raw but it plays one on TV.
Click to expand...

Isn't the definition of RAW video a signal that is scanned straight from the sensor without DeBayering? Do you not consider CRL RAW because it's compressed? Wouldn't that mean that RED, BM RAW
and all other compressed on-camera versions aren't RAW either then? Or is there something else specifically about CRL that you think makes it not RAW?
 
puredrifting said:
Isn't the definition of RAW video a signal that is scanned straight from the sensor without DeBayering? Do you not consider CRL RAW because it's compressed? Wouldn't that mean that RED, BM RAW
and all other compressed on-camera versions aren't RAW either then? Or is there something else specifically about CRL that you think makes it not RAW?
Click to expand...

It's a secret sauce inside those codecs as defined by lawyers and their engineers. Red Raw is real Raw. Sony X-OCN, BMD Raw and CRL are not deemed "raw under the law". ProRes Raw is probably real as well, which is why there was an attempt by Apple to litigate away Red's patents on Raw (so far unsuccessfully). Not knowing much ... eh, I mean, almost anything about these codecs, I assume that they select and delete the information present in the raw signal by frames in a different - but probably similar - way the H264 does it. Except H264 permanently deletes some info and compresses the rest, while these pseudo Raw codecs with their higher bit rate juggle all of the data.
 
DLD said:
It's a secret sauce inside those codecs as defined by lawyers and their engineers.
Click to expand...

Makes legal and marketing sense, I've just never heard of RAW digital video defined that way, everyones said for years that if its recording RAW sensor data and DeBayering in post, then it meets the definition of RAW from an engineering standpoint. When I shoot RED, I am typically using the 5:1 or 8:1 compression and while the footage looks slightly better than
the C200 CRL I have shot of the same subjects, same lighting, same everything, I always attributed that to the fact that the $30k to $50k RED is probably just a better sensor than the $7,500.00 Canon, not that their RAW recording is "better", although the REDCINE X-Pro software is definitely better. I'm open to being educated on this as I am not nearly as much
of a techno nerd as some of the crew here so your input is appreciated.
 
RAW used to mean: (1) not yet demosaicked and (2) uncompressed.

Now it just means: (1) not yet demosaicked.

---

You can:

(1) take a RAW image
(2) extract the three colors each into their own monochromatic layer (removing the gaps between the pixels, if that helps you imagine it)
(3) then compress each of the three layers with whatever codec you like (TIFF, JPEG, Wavelet, etc.) at whatever aggressiveness you want (lossless, lossy, 3:1, 5:1, 20:1, etc.)

---

Best explanation ever of the most common type of image compression, the Discrete Cosine Transform:

 
To me, the non-Raw "Raw" sounds like reverse engineering .... in reverse. In other words, the executives know it would be too much trouble to litigate, so they do whatever they need in order to disqualify the codec from being legally Raw. Meaning that there's some processing but not enough to be visually lossy.

A hypothesis - let's say one is shooting 24 FPS but the fake Raw is actually shooting 25, then demosaics the 25th frame ... and then completely discards it. There's no damage to the footage (sort of) and there's an H264 frame. Now, you're legally compressing a pure Raw data.

Sounds farfetched? What if it's the 6th, the 12th, the 18th and the 24th frame? And those serve as "control frames" with the rest being compressed Raw?

Like I said, my career in computer languages ended with the COBOL/FORTRAN era but the above sounds feasible, as one of the ways to slide by Red's patents.
 
You must log in or register to reply here.
Back
Top