XHA1 / Sony Vegas / fps

[Marcus Schuler]

Hi,

I wonder if other Vegas 7.0 Users face the same problem and hopefully somebody has a solution to the problem.

After capturing my m2t files with Vegas I mostly use HDV-Split to separate the file into single scenes. This makes it easier to get an overview over the whole project...When I drag - let´s say five or more - separated clips into my time line and play them - I only get 15 or 16 frames per second.

Before the XHA1 I had the Sony HC1. The separated Clips recorded with Sony's HC1 played perfectly with 25 fps. BUT: If I leave the m2t in a whole (not separating the file into single clips or scenes) the material taken with Canons XHA1 is just fine with 25 fps!

I use a Intel System, Core 2 Duo E 6600, over clocked: 2.7 with 3 GB RAM.

[Gert Kracht]

Hi Marcus,

I don't have that problem. But I don't split up the M2T files.
And there is a reason for that. I'm not sure, but splitting up the M2T files could mean a loss in quality.

The program has to decode and encode the files. It all depends on how it does that. Right now I'm figuring out what programs do with the Mpeg2 files.

Please read this article and it may explain a bit of what I'm pointing at.

http://www.dvxuser.com/articles/colorspace/

For me it's not clear what all software packages do with the content and it would be nice to figure that out. In the mean time I myself try not to 'harm' the content by splitting it up.

But....I may be wrong and I'm very curious about other opinions.
(I'm setting this line on Instant email notification)

[Jon McGuffin]

I would very much like to hear somebody chime in on this subject. Frankly, I can't imagine that loading up a 60 minute .m2t file directly on the Vegas timeline and cutting it up to pieces and putting it together would lead to quality degredation. Isn't this the SOLE purpose of an NLE, so cut and mix in clips?

I shot 63 minutes of video the other day that I'm going to widdle down to about 4 minutes and will probably cut 60 clips and put them together. I can't imagine there being quality issues simply by making the cuts..

Jon

[Seth Bloombaum]

Originally Posted by Jon McGuffin

I would very much like to hear somebody chime in on this subject. Frankly, I can't imagine that loading up a 60 minute .m2t file directly on the Vegas timeline and cutting it up to pieces and putting it together would lead to quality degredation...

Quite right.

Use the "s" key to split the files. Drag multiple instances of a single clip to the timeline. Make subclips in the trimmer. All the displayed clips or events you see are references to the original media file and have the same quality as the original.

Render something, anything - now you may take a quality hit, depending on how you've set it up. Render DV to DV, HDV to Cineform, Cineform to Cineform - negligible. Render HDV to HDV - some quality loss. Render anything to DVD quality MPEG2, you only have what 8Mb MPEG2 can give you. Render DVD MPEG2 to same, usually a significant quality hit.

Repeat rendering - quality losses are cumulative.

[Jon McGuffin]

Originally Posted by Seth Bloombaum

Quite right.

Use the "s" key to split the files. Drag multiple instances of a single clip to the timeline. Make subclips in the trimmer. All the displayed clips or events you see are references to the original media file and have the same quality as the original.

Render something, anything - now you may take a quality hit, depending on how you've set it up. Render DV to DV, HDV to Cineform, Cineform to Cineform - negligible. Render HDV to HDV - some quality loss. Render anything to DVD quality MPEG2, you only have what 8Mb MPEG2 can give you. Render DVD MPEG2 to same, usually a significant quality hit.

Repeat rendering - quality losses are cumulative.

Okay, obviously there is a quality loss in the rendering process but what I gather the user above was questioning was due to the GOP nature of HDV Mpeg-2 stream, when you make a cut (by hitting the S key) is there a loss of quality due to the fact the cut may take place between the Group of Frames?

Jon

[Gert Kracht]

That's exactly what I'm pointing at Jon.
And the same thing could happen when you split up the signal.
But maybe there is someone who can tell us more about the programs. That would be fun to talk about.

@Seth: Cineform has better results? Because it converts the M2T to it's own AVI format ? But what happens when you capture Mpeg2 HDV, make a movie of that with Cineform and then export it to HDV again?

[Jon McGuffin]

The quality conversion from raw .m2t -> Cineform should be nearly flawless.
There are a couple of main benefits to the Cineform Codec and I believe they are:

#1) Cineform codec actually convers the colorspace of HDV from 4:2:0 into 4:2:2 as well which makes it more ideal for any compositing work that needs to be done.

#2) Cineform .avi file is NOT a GOP based compression format and therefore frame precision cuts are better.

It's #2 though that I'm not 100% sure I understand though. Not so much that Cineform is better, but if editing GOP is such a disaster, why is it ever advocated as a solution at all?

In other words, if I import a 63min .m2t HDV file, and the footage needs to be cut down significantly, the more I make cuts, the more I'm hurting the overall quality of the export? Can't imagine that really....

[Gert Kracht]

[QUOTE=In other words, if I import a 63min .m2t HDV file, and the footage needs to be cut down significantly, the more I make cuts, the more I'm hurting the overall quality of the export? Can't imagine that really....[/QUOTE]

Not the overall quality, but the more you cut the original video, decompress the it, to edit it, and then compress it back to it's original state you will cause a sertain loss in quality.

Cineform uses it's own form of compression and that has my interest right now. I'm reading as much info as I can get right now.

Google on 'GOP structure' and you will find many explanations about it on the internet.

[Marcus Schuler]

Just an other result:

Now I have tried Neo HDV (Try Out Version) to convert a m2t-File and split it.
Result: frame rate is even worse than "just splitting" with HDV-Split. In preview mode (HDV 1440 x1080, 25.000 fps, Interlace) I reach just 11 fps. In comparison: putting the original m2t-File (splitted with Vegas or left in a whole) in my time line I get the full 25 fps.

Probably I´m too stupid for Vegas.
What´s wrong? Can someone help?

Here is the link to the original m2t-File directly from the xha1:

http://www.medienserver2.de/xha1.m2t

Marcus

[Seth Bloombaum]

Generally speaking, if you're not getting full frame rate previews with an m2t it is due to the processor bogging down on the decode.

If preview frame rate is low with Cineform, it is usually due to disk throughput issues. Usually. There are lots of threads regarding the advantages of multiple hard disks in video editing, no matter what NLE you're using it is very helpful to store your video on a dedicated drive (not your system drive) for best preview performance.

The bottom line is that High Definition video and HDV in particular require a PC that has sufficient processor and disk throughput to support full frame rate playback.

John McG. wrote:
...if editing GOP is such a disaster, why is it ever advocated as a solution at all?

Seems to me that this discussion has been theoretical. I've had good experience cutting native m2t on the timeline. I don't see quality degradation from cutting in the middle of the GOP.

Gert K. wrote:
...But what happens when you capture Mpeg2 HDV, make a movie of that with Cineform and then export it to HDV again?

I've never gone back out to HDV, myself, never saw the need to. Usually, I'm delivering on standard def DVD or HD WMV (so far). Some people like to archive back to HDV tape; I prefer to archive on hard drive. I have several drives that move in and out of external cases as needed, cost is as low as $0.25 US per gigabyte, but more frequently $0.35-0.40.

I would guess that HDV-Cineform-HDV would be equivalent to HDV-HDV in quality. I've never seen any perceivable loss with Cineform.

[Marcus Schuler]

Hi,

Seth Bloombaum wrote:
If preview frame rate is low with Cineform, it is usually due to disk throughput issues. Usually. There are lots of threads regarding the advantages of multiple hard disks in video editing, no matter what NLE you're using it is very helpful to store your video on a dedicated drive (not your system drive) for best preview performance.

The bottom line is that High Definition video and HDV in particular require a PC that has sufficient processor and disk throughput to support full frame rate playback.

I have a Core 2 Duo System, 2.4 MHz, over clocked to 2.7 MHz with 3 GB of RAM, I use one hard disk for system data and programmes (only Vegas 7.0 and Adobe Audition) and the I have four - in numbers 4 - other Disks for Audio and Video - all Samsung SATA 400 GB / 16 MB RAM. All of them are not older than six months.

I don´t think, that neither my computer is really the problem nor the configuration of it. May be I´m running Vegas with the wrong preferences? So far I never had problems with Vegas. I use it for more than a year. And then I switched to the Canon XHA1 a week ago..... I think, the problem is, that the XHA1 produces other m2t-files than I used to edit with Vegas before.....

[Jon McGuffin]

That's odd... Your system is comparable but likely faster than mine and I can *easily* play the .m2t files from my HDR FX1 on the timeline at their full 29.97fps native speed. It's only when I color correct, add transitions etc that the framerate drops and then, it's not that signficant never really falling below 15fps.

Thanks Seth for the comments above. I too have not had the need to encode back to HDV and use the same export methods you do. I keep the original .m2t for archiving purposes on the hard drive, export to SD DVD or WMV HD.

Jon

[Seth Bloombaum]

Marcus, I agree with Jon - you should be getting full framerate previews with HDV on that machine. I'm using essentially the same config/processor without overclocking.

Perhaps we can narrow in on any problems:

How is native HDV on the timeline? You should be getting rock-solid full framerate of a single track of HDV m2t, using an HDV project template.

I don't have experience with the XHA1, others on this forum are using it. Are you capturing with Vegas' video capture?

[Marcus Schuler]

Hi Seth,

thanks for your reply.

Yes, I´m getting full frame rate with a m2t file which is captured in Vegas. There is only one minor problem: if I seperate the file into clips there is a small tiny audio break between each clip - only for one or two frames. This is due to problems with the m2t Stream of Vegas in conjunction with Canons XHA1. The problem is well documented in the german Vegas forum.

But when I capture with Neo HDV or HDVSplit the performance - even in preview mode - is disastraous poor. I only get 15 or 16 fps. And I´m wondering how this comes from.

Marcus

[Jon McGuffin]

Originally Posted by Marcus Schuler

Hi Seth,

thanks for your reply.

Yes, I´m getting full frame rate with a m2t file which is captured in Vegas. There is only one minor problem: if I seperate the file into clips there is a small tiny audio break between each clip - only for one or two frames. This is due to problems with the m2t Stream of Vegas in conjunction with Canons XHA1. The problem is well documented in the german Vegas forum.

But when I capture with Neo HDV or HDVSplit the performance - even in preview mode - is disastraous poor. I only get 15 or 16 fps. And I´m wondering how this comes from.

Marcus

Hey Marcus,

That sounds like a pretty significant problem and I've not heard that mentioned around here. Would have figured that the standard "HDV" format would be the same regardless of the camera capturing but evidently not. My dad shoots on an A1 and uses Vegas and I haven't heard any complaints or troubles from him but then again, he's kind of a newbie and wouldn't surprise me a ton if he would just think this is normal.

Hopefully Vegas gets us a fix pronto.