Record uncompressed 14-bit (4:4:4) on your HVX200

We will be announcing release dates sometime after NAB, but the target release is during the second half of 2007, allowing for a couple months of beta testing after NAB. Things are moving a LOT quicker this time around, since we now have the experience from Andromeda.

About data rates, Andromeda currently dumps about 28MBytes/sec of uncompressed 24P data, which is roughly 100GB/hour. With Hydra on the HVX, remember that you need not record always at 14-bit, but assuming you do select 14-bit at 24P, excpect somewhere 60-80MBytes/sec at maximum resolution.

You can use the RAM record capability to test modes without a fast hard disk connected, and then use the xLUT tools to see if you can achieve the quality and look you want at a lower recorded bit depth. In some situations where decisions will be made in post or effects will be applied, 14-bit linear is probably desirable. However, in situations where the data is only to be color corrected you may find a 14-bit to 12 or 10-bit lookup table that will give you the exact same results as if you color corrected 14-bit in post, but gets applied in real-time.

Cheers,
Juan@reel-stream.om
 
pertierr said:
We will be announcing release dates sometime after NAB, but the target release is during the second half of 2007, allowing for a couple months of beta testing after NAB. Things are moving a LOT quicker this time around, since we now have the experience from Andromeda.

About data rates, Andromeda currently dumps about 28MBytes/sec of uncompressed 24P data, which is roughly 100GB/hour. With Hydra on the HVX, remember that you need not record always at 14-bit, but assuming you do select 14-bit at 24P, excpect somewhere 60-80MBytes/sec at maximum resolution.

You can use the RAM record capability to test modes without a fast hard disk connected, and then use the xLUT tools to see if you can achieve the quality and look you want at a lower recorded bit depth. In some situations where decisions will be made in post or effects will be applied, 14-bit linear is probably desirable. However, in situations where the data is only to be color corrected you may find a 14-bit to 12 or 10-bit lookup table that will give you the exact same results as if you color corrected 14-bit in post, but gets applied in real-time.

Cheers,
Juan@reel-stream.om

Thanks for the reply! I guess 10 bit 4:4:4 should be enough no real need for 14 bit. I guess if the Viper only shoots 10bit 4:4:4 it might be enough for most of us.:) If this camera is even close to what is claimed it will be truly amazing!

Okay last question what do you anticipate the data rate for 10 bit 4:4:4 at 2K?
 
Juan, thanks for the clarification of the camera frame rate, not camera mode being the key, which Barry helped get me on track with. I hope you nail a good post process to automate an on-line conform of the Hydra TC based files to the relevant off-line format used with the preferred NLE (FCP?).

The faster than RT dailies/off-line file creation seems manageable. With a dedicated data wrangler/editor and some equipment setup to the side (room or van) at the production location, all this could be churned out during the on set hours, and as Nik pointed out, an on location edit could be assembled during that time using a lappy.

Juan, one more question, will SculptorHD 2.0 provide for any useful form of RT monitoring for the DP/Camera op, similar to say DV Rack? If so, by using gigabit, will it be without the slight display latency caused by the firewire connection?
 
Proposed Feature Film full resolution workflow (sanity check):
  1. Sync camera throughout the day with timecode provided by Audio Engineer. (Sound will be recorded on an external system.)
  2. Record through a 1.83 MAC mini to one of a pair of 500GB RAID 1 solutions. (Two should be enough for a day's shooting, a little over 3 hours.) A P2 card will be required to set the 24P timecode.
  3. Evening: technical PA converts all raw files from the Firewire arrays to the full resolution Quicktime file using a windows workstation as the target repository (see below).
  4. PA Copies original RAW files to USB drive for offsite.
  5. PA verifys all files before deleting originals from the Firewire RAID.
Or if you have a remote location for the workstation (up to 300 feet):

  1. Sync camera throughout the day with timecode provided by Audio Engineer.
  2. Record directly to workstation. A P2 card will be required to get the 24P timecode.
  3. Evening: PA Copies the day's raw files to an external drive for offsite.
  4. PA converts all raw files to the full resolution Quicktime files.
  5. PA verifys all files before deleting raw originals.
The workstation would have a dedicated Adaptec RAID card that support a 16 drive RAID 5 array. Using WD SE-16 500GB drives: 7TB = 25 hours at full rez. Massive, because it has to be...

Add it up (Assuming you have an existing workstation):
  • 31 disks @ $125 each (bulk discount above 20)
  • 15 USB external enclosures @ $45 each
  • Adaptec RAID 31605 $995
  • Server class comp case & PSs: $600
Cost to capture 4:4:4 1080 24P - Priceless! :thumbsup:

(Not really, it will cost $6145 + the cost of the Hydra system) :huh:

Assuming I understand it correctly...:undecided

To be fair, even working with Panasonic's compression will require 15 disks and a RAID card that is $575. After looking into the sound forum and at the products they use, I will contract a pro and rent the equipment.

Bit much for me right now, but I would probably hire someone with a Hydra rig to do overcranking on a few scenes or any greenscreen work. If my pockets were just a little deeper I'd be an owner.

Bob
 
To be honest, doesn't even sound like you need all of that. Which is why I'd like to see what Juan is working on as of right now.

I doubt you're looking at a six thousand dollar setup. Me thinks Gruen is OVERTHINKING the needed capacity to use Hydra.
 
Bob Gruen said:
Proposed Feature Film full resolution workflow (sanity check):
  1. Sync camera throughout the day with timecode provided by Audio Engineer. (Sound will be recorded on an external system.)
  2. Record through a 1.83 MAC mini to one of a pair of 500GB RAID 1 solutions. (Two should be enough for a day's shooting, a little over 3 hours.) A P2 card will be required to set the 24P timecode.
  3. Evening: technical PA converts all raw files from the Firewire arrays to the full resolution Quicktime file using a windows workstation as the target repository (see below).
  4. PA Copies original RAW files to USB drive for offsite.
  5. PA verifys all files before deleting originals from the Firewire RAID.
Or if you have a remote location for the workstation (up to 300 feet):
  1. Sync camera throughout the day with timecode provided by Audio Engineer.
  2. Record directly to workstation. A P2 card will be required to get the 24P timecode.
  3. Evening: PA Copies the day's raw files to an external drive for offsite.
  4. PA converts all raw files to the full resolution Quicktime files.
  5. PA verifys all files before deleting raw originals.
The workstation would have a dedicated Adaptec RAID card that support a 16 drive RAID 5 array. Using WD SE-16 500GB drives: 7TB = 25 hours at full rez. Massive, because it has to be...

Add it up (Assuming you have an existing workstation):
  • 31 disks @ $125 each (bulk discount above 20)
  • 15 USB external enclosures @ $45 each
  • Adaptec RAID 31605 $995
  • Server class comp case & PSs: $600
Cost to capture 4:4:4 1080 24P - Priceless! :thumbsup:

(Not really, it will cost $6145 + the cost of the Hydra system) :huh:

Assuming I understand it correctly...:undecided

To be fair, even working with Panasonic's compression will require 15 disks and a RAID card that is $575. After looking into the sound forum and at the products they use, I will contract a pro and rent the equipment.

Bit much for me right now, but I would probably hire someone with a Hydra rig to do overcranking on a few scenes or any greenscreen work. If my pockets were just a little deeper I'd be an owner.

Bob

This does seem a little overkill. See my post earlier about a simple cheap workflow. Mine isn't as redundant as yours but it is cheaper by far.
 
Even so, if you are shooting a feature length production at this level, $6k or more allocated for these items does not seem unreasonable to me, and they would be useable beyond the production. Before this discussion, I had factored a 6.5TB system to edit and on-line an HVX P2 sourced 1080p24 feature using CineForm Prospect HD 10bit, with redundancy in mind. I would prefer to have an over kill amount of drives, as I always seem to fill them all up. Granted, we are indie budget minded here. Hydra, this is going to be cool. :thumbup:
 
yeah, true Kholi.

Regarding audio, I would also consider recording audio to P2, as opposed to an external audio device. The HVX does not have TC I/O, and I am am wondering Bob, are you basing post audio sync on using the clapper board? Or does your proposed work flow entail an automated post audio conform? This is an interesting workflow aspect you raised.
 
14-bit imagery should be extremely smooth with no banding, no blocking, no gradients, just perfectly smooth images.

I think we need to keep in mind here that this is linear data not video or logarithmic encoded footage. 14bits is actually about the minimum for good image reproduction with linear encoding.
 
About audio sync, you can do it any way you'd like. You can shoot with a P2 card or miniDV tape in the camera and get the audio there. The timecode generated on the uncompressed recordings should already match the timecode and start/stop takes on the P2 or tape. Or you can record audio separately film-style and clap-on clap-off. :) If you're doing it this way and your crew has worked with film, it shouldn't be any different or any more complicated.

Some points about the posted workflow:

Bob Gruen said:
1.Sync camera throughout the day with timecode provided by Audio Engineer. (Sound will be recorded on an external system.)

If you're recording audio externally, then this really isn't any different than a traditional film workfow...clap-on clap-off! :)

2.Record through a 1.83 MAC mini to one of a pair of 500GB RAID 1 solutions. (Two should be enough for a day's shooting, a little over 3 hours.) A P2 card will be required to set the 24P timecode.

Where does the P2 card come in? The uncompressed footage will have its own timecode. Another option is a SMPTE timecode output from Sculptor into your audio, or vice-versa(timecode slate, etc).

3.Evening: technical PA converts all raw files from the Firewire arrays to the full resolution Quicktime file using a windows workstation as the target repository (see below).

With Hydra(actally SculptorHD 2.0) your raw files -are- your dailies, so there is no need to process to view what you got. You can take a look at it, set IN/OUT render points, and render only the material you want for the actual master.

If you wish to process everything that's a specific workflow decision, but it is akin to scanning all the cans from a motion picture shoot at 4K, not just the parts that are going to be worked on digitally. Ouch!

3.PA Copies original RAW files to USB drive for offsite.

Or just takes the drives with him/her, like a can. :) It seems like the money you'll spend paying a person to wait for gigabytes of data to be moved to a USB drive can be better spent getting a second record drive that can be taken off site.

I'm just saying this because this is for the most part how our Andromeda users work. Many use 7200RPM firewire drives, and when it's full, you just swap it out like you would a P2 card.

4.PA verifys all files before deleting originals from the Firewire RAID.

Again, there seems to be an extra step in the firewire drive -> USB drive -> windows repository chain. Why not just firewire drive -> windows repository?

We posted a document a while ago showing some sample workflows for Andromeda:
http://www.reel-stream.com/SampleWorkflows.pdf

We'll have a Hydra version online eventually.

Cheers!
Juan@reel-stream.com
 
Juan, I'm sure you thought about this but do you have enough staff and space to support the sea of incoming HVXes that will soon be coming in to be moded? Don't underestimate please. I'd hate to wait months like I did last year for my M2.
 
wow, some of this info is a little overwhelming. ive never had a production big enough to warrant some of this stuff. since hydra may very well boost our beloved hvx into a different caliber of camera, i think this should be its own forum to focus on specific issues that may come up with this obviously more labor intensive workflow.
 
No Sync sound via gigabit in the same stream may be a dealbreaker for me.

First of all the camera is already amazing in terms of quality.

Second, the mod is likely going to be very expensive. I wouldn't doubt 4000$ or more. I think it sounds amazing and certianly worth the money, but for that tag, for a mod, no less, I think you should have sound in the stream. If it's less, then maybe not.

I think in-camera sound recording is one of the main advantages to the digital format in general. It's a shame to lose that no matter what the improvement is.

I've also heard horror stories about andromeda. Crashes, data loss, etc. The tether doesn't bother me at all, and previewing the video stream for focus is something I do anyway. But to have to setup the sound for each clip would get very old very fast. I seriously hope Juan doesn't repeat the same mistake.
 
arrestthisman said:
No Sync sound via gigabit in the same stream may be a dealbreaker for me.

First of all the camera is already amazing in terms of quality.

Second, the mod is likely going to be very expensive. I wouldn't doubt 4000$ or more. I think it sounds amazing and certianly worth the money, but for that tag, for a mod, no less, I think you should have sound in the stream. If it's less, then maybe not.

I think in-camera sound recording is one of the main advantages to the digital format in general. It's a shame to lose that no matter what the improvement is.

I've also heard horror stories about andromeda. Crashes, data loss, etc. The tether doesn't bother me at all, and previewing the video stream for focus is something I do anyway. But to have to setup the sound for each clip would get very old very fast. I seriously hope Juan doesn't repeat the same mistake.

Shoot sound to P2 card. I don't understand what is so difficult about the that workflow.
computers crash sometimes I know how you feel. That is why you have to backup! The product isn't out yet but if you are backing up everyday I can't see you really having a huge problem with data loss.
 
it's not difficult, but it's an extra step.

One of the notions behind the HVX and moreso, digital in general, that I find appealing is the automation of it all. The less technical work there is, the more time to work on the scene with the actors. Don't get me wrong.

I LOVE the technical stuff. But that's my own guilty pleasure. On the set, I need things in a timely manner, and immediate dailes are nice in the digital age. Another step in the chain takes more time. It just does. Period. And with a gigabit connection there's no way they don't have the bandwidth for sound.

Not to mention, one of the nice things about the tethered, direct to disk system, is controlling the feed. If you are only recording one stream of audio, it would be nice to have the video file only one channel of audio. What if you want 8! Add an Mbox or something to your laptop and get another few channels for the scene.

If they allowed you to do this, it would make the whole process better.

Think of it. One engineer at a laptop doing video AND sound, while you're at the camera. Very streamlined. less crew = betterness and happy landlords.

I'll say this though, if they couldn't pull off audio from the camera, but allowed an Mbox-type solution to record audio streams into the video stream in the software for sync sound, I would jump at it. That would be just fine if not better than in-camera audio!

So Juan, how about it~!?!??!?!
 
Back
Top