Request 4K 10-bit 4:2:2 internal for the DVX200!

Sent my requests for Color Matrix & Color Correction menu choices like the PX270, 10bit 4:2:2 codecs for 720p & 1080p along with the hope of the 1080p120 mode recording audio so one could film long stretches of 120fps just like any other video mode.

Lets hope Panasonic wants to adjust to an ever evolving landscape.
 
DLD said:
That and the cost of executing this request is something no one outside of Panasonic - and folks with a professional knowledge of electronic engineering - knows.
Click to expand...

Point well taken, and I made sure to acknowledge in my email that I'm aware there might have been some sort of engineering reason why it wasn't included. It's 5-6 months from release, so at this point it may well be too far down the road.

But, if it's possible, it sure would be nice. :)

roxics said:
If Blackmagic can offer 4K ProRes and Raw in a $3K camera, I would hope Panasonic can offer 10bit 4:2:2 H.264 in a $5000 camera. I realize it's more compressed but come on. In fact they should be offering ProRes as an option, at least for 1080p.
Click to expand...

As DLD said, they're offering to completely different feature sets. The lens on the DVX200, if it were one of their X lenses, would be ~14-183mm. My 12-35mm and 35-100mm X lenses combined would cost about $2,700 today. The DVX's lens is nearly twice as long, while only losing 1.2 stops of light at the far end. And it has a servo zoom and internal ND filters. It also uses cheaper batteries. And Cfast 2.0 cards are twice as expensive while holding far fewer minutes of raw or ProRos than the DVX200 will of it's MP4 or AVCHD footage.

Not knocking the Ursa mini... the specs vs price make it look like a really attractive camera (especially if you already have a bunch of EF glass... I think BMD is going to sell a ton of them). But it's like comaparing a Ford truck to a Jeep... it's not quite apples to apples.
 
There are people in this forum who will not let you even compare Braeburns to Granny Smiths.

Come on. What roxics means is that the URSA Mini, in offering 4K ProRes in a $3,000 camera, shows that 4K ProRes cannot cost that much in and of itself, because with just $3,000 Blackmagic includes not only 4K ProRes but also RAW, a 5" 1080p screen, dual phantom XLRs, a 4K APS-C sensor --- on and on. And it makes some amount of profit on top of all that.

You can call out the differences between this camera and that camera, to lead to the conclusion that it might be thousands of dollars for feature X. Or you can kick dust up in the air in a storm in uncertainty and doubt, to lead to the conclusion that, you know, we just have no clue how much feature X might cost. But I don't know why you do! Are you in the service of preventing broken hearts? I mean, seriously, what is your motivation for poo-pooing feature requests? I mean, what bad thing are you anticipating if someone asks for something that they might not get?
 
Last edited:
Do we know if the 10-bit external output is real or the pseudo 10-bit that they added to the AF100a? (Not that I've heard good or bad about it.)

-cp
 
I'm still confused by people here saying that recording 4K 10-bit 4:2:2 internally won't provide better results than 4K 8-bit 4:2:0. What about this image?


Taken from this video:
https://vimeo.com/114978513

To me, this shows a clear advantage to recording in 4K 10-bit 4:2:2, so why not ask for it as part of the camera's recording options?
 
Jaime Valles said:
I'm still confused by people here saying that recording 4K 10-bit 4:2:2 internally won't provide better results than 4K 8-bit 4:2:0...
Click to expand...
A "clean" HDMI out for 422 - or for 444 via HDMI 2.0a with something like C300 MK II - keeps a lot of the data from the sensor that the internal recording will simply discard during the compression process.

I think what Barry and harddrive have been saying is that a typical 8.3/8.8 MPX Bayer mask sensor used for/on 4K doesn't have the full 422 data to begin with. In that math equation, to go native 422 4K, you'd need 16.6/17.2 MPX and, to go 444, you'd need 24.9/26.4 (a la F35 for 2K).

PS. And, if this is not what Barry and harddrive think ... sowwwwwweeeee...
 
Just for comparison. The Sony EA50 HD APS-C camera has 13.6Mpixels for 16/9 video or stills. Plenty for upgrade to 4K !!! Just needs the electronics and plenty of room for a fan !!! Could be a lower cost variant to the FS7 for Sony.
 
DLD said:
I think what Barry and harddrive have been saying is that a typical 8.3/8.8 MPX Bayer mask sensor used for/on 4K doesn't have the full 422 data to begin with. In that math equation, to go native 422 4K, you'd need 16.6/17.2 MPX and, to go 444, you'd need 24.9/26.4 (a la F35 for 2K).
Click to expand...
Oh! I think I see what you're saying. If the DVX200 sensor doesn't have as many pixels as the GH4 sensor, the 4K 10-bit 4:2:2 mode won't be useful, unlike that of the GH4. Is that right? I was thinking the DVX200 sensor would be just as pixel dense as the GH4.
 
Jaime Valles said:
Oh! I think I see what you're saying. If the DVX200 sensor doesn't have as many pixels as the GH4 sensor, the 4K 10-bit 4:2:2 mode won't be useful, unlike that of the GH4. Is that right? I was thinking the DVX200 sensor would be just as pixel dense as the GH4.
Click to expand...
It may have as many pixels as GH4 (or it may not, right?) but it's still likely to use the same center 2.3 crop for 4K. So, for practical intent, 4:2:0 will do the trick.

Now, mind you, the key is not only in the sensor itself but in the particular codec and the processing power of each camera, which discard different amounts of data at different points in their processing paths. Toss the data out right away and the end product won't be very good. Perform the compression at the final output stage and the camera can have a sterling image at as low as ~ 25 Mbps (which is the Netflix 4K streaming bit rate). And the catch is that no one really knows what the heck is inside the camera aside of its engineers. So, quite conceivably, a 40 Mbps from Model A can be superior to 100 Mbps from Model B. The only way to do a performance comparison is to do a ... performance comparison.
 
Jaime Valles said:
I'm still confused by people here saying that recording 4K 10-bit 4:2:2 internally won't provide better results than 4K 8-bit 4:2:0. What about this image?


Taken from this video:
https://vimeo.com/114978513

To me, this shows a clear advantage to recording in 4K 10-bit 4:2:2, so why not ask for it as part of the camera's recording options?
Click to expand...

That isn't comparing apples to apples. H264 internal recording has its own banding issues which is what you are seeing most of here. Plus 4:2:0 untreated will look like what one would expect 4:2:0 to look like. It is after all tossing out every other color sample. Chroma up-sampling however helps to add those missing pixels back in.

With a bayer sensor every single pixel is interpolated. So red pixel #2 is created from an average of red pixel #1 and red pixel #3. Yes the math is a bit more complicated but my point is that math is used to create that missing color pixel. When recorded as 4:2:0 the encoder tosses out red pixel #2. By using chroma up-sampling (AE, Premiere, 5DtoRGB) you once again create red pixel #2 by averaging red pixel #1 and red pixel #3. So you really fill back in that missing color sample with a value very close to the original value. Its not a perfect science of course since there are things like rounding error and bayer sensors are a bit more elegant and complex in how they calculate values but it gets you like 80% there to something that is almost visually impossible to tell apart. Especially if you shoot softer details (adapted lenses, detail turned down, shooting RAW, ect.). What it gives you for sure is a complete elimination to the jagged edges from 4:2:0.

4:2:0 on its own is not a bad thing at all. It is all the other junk going on with H264 compression that hurts the image. This is why I would take a 10 bit ProRes external recording over an internal 10 bit H264 encoding any day.
 
Thomas Smet said:
4:2:0 on its own is not a bad thing at all. It is all the other junk going on with H264 compression that hurts the image. This is why I would take a 10 bit ProRes external recording over an internal 10 bit H264 encoding any day.
Click to expand...

There is also a much more practical issue that would need to be addressed... can a SDXC or microP2 card even handle 10bit 4k 4:2:2? AVC-Ultra 4k is 400mbs. Even if they got a more compressed long-gop codec to use, what would the data rate on it be? The current 4k codec is 100mbs. Could they manage to get a 10bit and/or 4:2:2 flavor at 200mb/s? And if so, just how much processing power would it require, and could they stick it in a camera the size of the DVX200 without seriously increasing the cost or requiring much more expensive media? Because creating a jack of all trade's camera like this at this price point that requires me to buy Cfast 2.0 cards or SSD's sort of defeats the purpose.
 
SD cards are very fast now such as the Lexar 2000x series (260MB/s write speed). Even slower cards such as the Lexar 1000x and the Sandisk 95MB/s cards can handle decent bit rates.
 
Last edited:
I have the Sandisk 95mb/s for my GH4, and you're right, they do handle 200mb/s like a champ. But will they handle more than that? Or, a better question... is Panasonic willing to guarantee it? As it stands, the Sandisk 95mb/s cards I have are technically not supported, as they aren't U-3 cards. Would Panasonic be willing to guarantee non-microP2 cards for anything faster than than 200mb/s at this point in time?
 
Last edited:
Thomas Smet said:
That isn't comparing apples to apples. H264 internal recording has its own banding issues which is what you are seeing most of here.
Click to expand...
I see. It's H.264 compression that's the problem in that example, not 8-bit 4:2:0? So the solution for less banding in internally recorded footage is a higher bitrate (say 200Mbps in 4K)?
 
I watched a video over on Vimeo introducing the DVX200. The rep from Panasonic mentioned the camera supported AVC intra and 4K 4-4-4. Of course we have not actually seen the camera yet, but one can hope. That would be huge. It makes me feel so puny with only AVC-100 with my obsolete HPX-250.
 
BillinSC said:
I watched a video over on Vimeo introducing the DVX200. The rep from Panasonic mentioned the camera supported AVC intra and 4K 4-4-4. Of course we have not actually seen the camera yet, but one can hope. That would be huge. It makes me feel so puny with only AVC-100 with my obsolete HPX-250.
Click to expand...

discussed before. The rep confused the Varicam with the DVX200. No specs like this on this cam.
 
You must log in or register to reply here.
Back
Top