How could a 1920x1080 4:2:2 workflow (XDCAM EX HD SDI out-> ... -> Sony Vegas 8.0 in) look like?
Is the EX' 4:2:2 HD SDI output RGB or YUV?
Is the EX' 4:2:2 HD SDI output 10 or 8bit?
What are the data rates?
Which Vegas capture codec would I use?
...
Thanks a lot in advance.

I think XDCAM HD and the EX are 4:2:0, not 4:2:2.


-A

I can see that this is going to become a major interest point for others as well as myself. The latest review:

http://www.simonwyndham.co.uk/xdcam-pmw-ex1.html

is very specific that the camera outputs HD-SDI in 4:2:2.

My understanding is that the HD-SDI spec requires 4:2:2.

Problem is, all other info specifies the camera recordings as 4:2:0, which the MPEG compressed file certainly is. Does the codec expand/uprez the 4:2:0 back to a full bandwidth, 4:2:2 signal for playout through the SDI port?

Is it perhaps 4:2:2 in live recording only, outputting the pre-compression signal to the SDI encoder, which converts it to 4:2:2? This is the more common practice.

What the codec is doing for recording internally and what the image processor is doing for SDI output need have no connection at all. The same goes for the HDMI outputs.

Oops, I over looked the "SDI out" part in Michael's post.

Is it perhaps 4:2:2 in live recording only, outputting the pre-compression signal to the SDI encoder, which converts it to 4:2:2? This is the more common practice.
That would make sense.


-A

Along the same lines, would one be able to capture the HD-SDI footage to a laptop?

Along the same lines, would one be able to capture the HD-SDI footage to a laptop?
If you are using something like the AJA Io HD or Motu V3 and a laptop w/an eSATA card connected to sufficiently fast storage I'd assume the answer would be "yes."


-A

How fast of a data rate are we talking?

How fast of storage is needed?

Thanks

How fast of a data rate are we talking?HD-SDI output is approximately 1.5 gigabits per second. It's a raw, uncompressed feed. Ideally you'll need an encoder on the other end of the pipe to compress it into a manageable, workable format.

this may be a solution for uncompressed video to compressed video in a laptop.

http://www.magma.com/products/pciexpress/expressbox1/index.html

rand

Dunno If I'd want to spend on that solution, since it adds a lot of new potential weak links (connections) to the flow. If I have to be tethered to that much hardware, and I'm thinking studio or indie production here, I'd go to the trouble to bring my G5, get the benefits of having FCP onsite, and use my Kona card to do realtime conversion from SDI to Apple Prores 422, or other edit friendly format, or skip it and go native HD-SDI. Just need a boxcar of RAID storage.

For me, the mantra is still KISS. One thing I do like about the HD-SDI workflow is it's just one BNC connector/line. Only objection to being tethered to even that degree, is not using Steadicam type balanced rigs.

Along the same lines, would one be able to capture the HD-SDI footage to a laptop?


You no longer need to lug around a laptop, Convergent Design have a little box that'll record at 50Mb/sec to CF cards from HD-SDI:

http://www.dvinfo.net/conf/showthread.php?t=102312

My understanding is that the HD-SDI spec requires 4:2:2.

Problem is, all other info specifies the camera recordings as 4:2:0, which the MPEG compressed file certainly is. Does the codec expand/uprez the 4:2:0 back to a full bandwidth, 4:2:2 signal for playout through the SDI port?

Is it perhaps 4:2:2 in live recording only, outputting the pre-compression signal to the SDI encoder, which converts it to 4:2:2? This is the more common practice.

That is exactly how it works. Same with the full size F350, and the Canon XLH1, XHG1.

Live HDSDI before recording is uncompressed 4:2:2. After recording, MPEG compression and chroma subsampling have taken place. Playing back pre-recorded material will be upsampled as you indicated to conform to the HDSDI spec.

-gb-

Re the Flash XDR press release:

The reason to go this route would be to gain 4:2:2 HD-SDI capture capability. The next format step down it offers is essentially what the EX records on SxS cards.

So, $5000+CF cards. 26 Min of storage across 2 16Gb cards
(at HD-SDI/160Mb/sec)
2x16GB cards= $400 (High speed CF16Gb=$200ea street)
4 cards x $200 = 52min for $800.
Euro price for two 25min 16Gb SxS-$1930.
2.4lbs. + batttery.

No thanks. Might be attractive to some production houses, but expensive for the indie producer or filmmaker. Might be a good item for rental. I'd sooner do the SxS shuffle, and convert to a lossless codec on capture. Users could add single cards as they can afford, and as price drops. If third parties truly offer compatible SxS, prices will drop drastically.

Have to make the transition someday. Meantime, I could live happily if the stored signal is upconverted to true 4:2:2 for playout from the cam or forthcoming player.

That is exactly how it works. Same with the full size F350, and the Canon XLH1, XHG1.

Live HDSDI before recording is uncompressed 4:2:2. After recording, MPEG compression and chroma subsampling have taken place. Playing back pre-recorded material will be upsampled as you indicated to conform to the HDSDI spec.

-gb-

Good enough for government work!
(favorite quote from my military days)

Thanks for your answers.

I have some questions about the 4:2:2 feed.

It seems the processing power is needed to convert the raw stream into a compressed, easier to edit stream.

Would it be less taxing to actually capture the raw stream and compress it after the fact, not in realtime?

Or would the raw 4:2:2 files be huge and take up too much space?

Thanks

Would it be less taxing to actually capture the raw stream and compress it after the fact, not in realtime?

Or would the raw 4:2:2 files be huge and take up too much space?

There is this trade-off of sorts that occurs between the two extremes. When you have uncompressed material, you need a wider I/O pipeline and more storage than you do CPU performance. When you work with highly compressed material, the opposite is true...more taxing on CPU, less I/O bandwidth and storage size.

With hard drive prices getting so cheap these days, I see no reason not to build a RAID set up and then you are looking at about $1000 for a decent capture card. Working with uncompressed helps maintain image integrity throughout the post process also.

-gb-

Another option would be to use Cineform Prospect HD.

www.cineform.com

You can capture on a single drive without much trouble. It's a pretty good way to edit, especially for Premiere Pro users. I do believe they have other solutions for other NLE systems.

So in order to make best use of the EX I'd have to:

Get CS3 and Prospect 2K for my Dell M90 (that has 4g ram) to get decent results editing HD.

I want to get this right as mistakes will be costly.
Think I may hold off on CS3 for the moment to see how soon they bring out support for the camera directly.

Thanks for the responses o this thread.

4:2:2 SDI sounds great, but seems like it is pretty intensive to get running.

So it looks like to aquire from the SDI port there are two ways to go:

Flash DXR - no computer needed, records on to CF cards, $5,000

Tether to a computer - needs a hardware card, software to go with the hardware card, and a computer.


Is there a piece of hardware that interfaces with a laptop for the second solution?

Thanks

Aidan,

You should only need Prospect HD, not the 2K version.
Also, Prospect HD should be backward-compatible with PPro 2 as I downloaded the trial version with only PP2 installed, and it worked.

Another option would be to use the Black Magic Intensity Pro card along with Black Magic HD Link and a box like this: http://www.dvinfo.net/conf/showthread.php?t=97526&highlight=Intensity

Or you could use the same box with a Black Magic DeckLink HD Extreme.

This solution along with Cineform should work fairly well.

Jim