[David Timar]

Thanks for the reply. I’ve tested some heavy duty computers, and unfortunately, it is Sony Vegas that is the bottleneck in playing back 4K files. It just cannot decode them on a low resolution, like Scratch for example.

Should should change Vegas to:
1. better access to the RED SDK, so you can adjust the decoding process, and not just the image attributes
2. support RED Rocket, and start really using GPU for playbacks

As it is now, Adobe Premiere CS5 wins this round, because it can easily play back multiple 4K R3D files natively, by utilizing GPU and Red Rocket support.

[David Timar]

Hi Dave,

I’d be editing in 8-bit mode, as color grading is not important.

Still, what would you suggest? Who should I contact? Who can give me a definitive answer to the hardware requirements of playing back 3 4K RED videos simultaneously in Sony Vegas (Draft Full or Preview Half mode at least)?

Thanks
David

[David Timar]

Hi Dave,

Thanks for your reply. I’ll post my question between your reply here:

“Ok, not a Red user, but I can help on the math. An R3D file runs either 28MB/s or 36MB/s… yeah, that’s not far from the limit of some 7200RPM drives, but you should be just dandy with one layer per drive. More than that, and you’ll potentially get HDD thrashing… I’ve run into this with Cineform files, which are only about half that size (60GB/hr).

You specified that you’re seeing 120Mb/s on each drive… I assume you don’t mean that (eg, 120 megabits per second), but actually mean 120MB/s (120 megabytes per second). An actual 120Mb/s would be limiting.. your files are a minimum of 224Mb/s.”

My bad. I’m getting around 100MB/s on my Samsung 1Tb 103UJ drives. So I guess hard drives are not the bottleneck then.

“On your setup, I suspect RAM is a limiting factor. Vegas 32-bit can only use 2GB, and that’s not even all that comfortable a chunk for large 2K projects, much less 4K video. Think of it this way… if you’re only working in 8-bit color, you’re going to need about 27MB per frame per layer, plus all the memory used for Vegas, for compositing, etc. At 32-bit color, that’s over 100MB per frame. No matter how you slice it, 4x the memory footprint of regular 1080p HD work.

I’d recommend using 64-bit Vegas and at least 8GB of memory, before you look any further. A Q6600 is also a pretty ancient 4-core machine… I replaced mine with a Q9550 well over a year ago. But as you’ve seen, it’s not the CPU load that’s the issue.”

Curious… I was under the impression, that if you set the Video RAM to be used by Sony Vegas to 0, then all it’s doing, is reading the videos from the wherever they are, without trying to use the specified RAM for caching. So how come RAM matters then?

You’re saying 100MB per frame. Is that what Sony Vegas needs in memory, in order to actually play back a 4K video? How did you calculated that?

Switching to at least 8GB of RAM would be an easy fix, but I’m worried that it’s not enough to solve this problem.

Dave, do you think it’s possible, that I make a R3D file available for you, you download it, copy it 2 more times so you have 3 separate files, then see if you can play it back properly?

Thanks
David

[David Timar]

Hi David,

I hate situations like this.

I bump into a problem, post a topic, keep trying to find out what could be the problem, get some amazing people helping me, and then without changing anything, all of a sudden the problem disappears.

I just tried doing what I’ve been doing again, and the end result didn’t get squished…. Sometimes I hate computers… They have a mind of their own.

Anyways, it looks like it’s working the way it’s supposed to. 🙂

Thanks for your help David.

Cheers.

[David Timar]

Are we talking about the same thing? I’m rendering into MXF. 1920×1080.

[David Timar]

Hi David,

I’m not sure how codec would come into effect here.
If I take an empty project, render any kind of generated media into it and then use the render function of Sony Vegas, there is no option to use any kind of codec. It’s a simple MXF export option.

Unless I’m missing something.

Thanks
David

[David Timar]

Hi David,

Thanks for the reply.
Here is the screenshot of the re-rendered mxf.

image2.jpg

For the sake of a clean test, I started a new project with 1920×1080, 29.97fps, PAR 1.000, progressive settings. I put a generated color bar on a single video track, then rendered a small part of it into MXF, with the following settings:
– HD
– 1920×1080
– 29.97fps
– 16:9
– Progressive
– 35 VBR (HQ)
– Main

The problem still remains. How can something that is a square pixel, in a square pixel project, become distorted while it’s properties still saying it is square pixel?

Thanks
David

[David Timar]

Quantize to Frames is on, never turned it off. Changed the timeline ruler, didn’t help. At this point, I think this is a bug.
I’ll investigate this further.

Thanks
David

[David Timar]

Hi John,

Thanks for the quick reply. 🙂

My problem isn’t audio unsyncing really. The problem is half-frame gaps appearing at the edges of the video events.

This is how an mxf looks like on the timeline, with the cursor standing on a frame.

See how the end of the video event itself is between 2 adjacent frames?

And if I resize the end of the video event to the frame that the cursor was standing on, I get a half-second gap.

And this happens THE INSTANT I move the video part of an event. EVEN if I DO NOT move the actual event ANYWHERE, just move the video one track up or down from it’s original track.

It could be that my mxf files are damaged, because I only noticed this in some of the mxf files on the timeline. Been editing for a good 2-3 years on daily basis, but I haven’t seen this problem before.

David

[David Timar]

Make sure you use JVC HD Pro tapes only. never put sony in a JVC camera. those dropouts can come from old firmware or dirty heads. get a cleaning tape and make sure you have the latest firmware. go to your jvc dealership and get them to update your camera to the latest firmware. be prepared that your camera might crash and would have to be sent for repair, but it should be fine after that. 🙂

firestores are great, but if there are problems generated by the camera, they WILL make it to the firestore files.
Also, make sure that you use something like Total Commander to copy split files from the firestore, onto your harddrive. the firestore splits long movies automatically, but it’s stupid enough so that the only way you get your movie back into a single file, is if you write to Focus and ask their support on how to combine 2 files….

Also, I STRONGLY suggest that get to know MPEG Streamclip and MPEG2Repair. Google them and get them!

Streamclip will allow you to chop up an m2t file without any re-compressing. 🙂

MPEG2Repair, will test and repair your m2t files. This small little things has save my life many times. Our JVC had a period when it was recording some really nasty errors onto the files on the Firestore. You run this little program, and it tells you exactly what’s wrong with it, and tries to fix it. Works 90% of the time. 🙂

If you get errors in your files and this little utility tells you about it, then first reset your firestore completely. if the error still remains, change the firewire cable. if it still comes up, call your dealership and pray to God… 🙂

David