Playback becomes choppy

[barahn]

Here's one- I have a project I'm working on and it is fine at first but when I start adding clips or sliding clips around or doing whatever to the project on the time line, all of a sudden it will become choppy on the playback. When I finish the project, it will still be choppy on playback so I exit out of VT3 and bring it back up and when I do, it plays back the entire project perfectly. But if I start moving things around again it gets choppy again. This is very consistent. My question is - What changes? If it will play all the way through with no problem, shouldn't it con't to do so? What could adding and deleting clips or whatever do to change the playback capability? Anyone experience this??

Thanks

[Brian Peterson]

Have you defraged your drives lately?

[barahn]

Yes - all of them

[Jim Capillo]

Have you tried dropping a force render under it all?

[barahn]

What is that and why would I do it?

[Jim_C]

>>>>My question is - What changes? If it will play all the way through with no problem, shouldn't it con't to do so? What could adding and deleting clips or whatever do to change the playback capability?


Because you are doing just that. You are altering the playback of the clip every time you make a change. When you make a change the system starts processing the changes as fast as it can to be ready for playback. It's called background rendering. There is a green light above the play button, if its not green, its not ready to play. If its green and it still plays choppy you need to reset some of your preferences.
This is, as well as force render, covered in the manual.

[Jim Capillo]

Originally posted by barahn
What is that and why would I do it?

I believe the force render is in the effects folder, but don't hold me to that - I've moved mine to the DVE root. Put it below everything on the timeline and drag out to the correct length. It will force the computer to render whatever you have selected.

Jim is right - anytime you move a clip requires a re-render. You're basically changing everything from that point on. On top of the play button is a thin bar - wait 'till it's green and you should have smooth playback. Depending on your system's speed (you didn't specify your equipment) and the complexity of the timeline, it could be anywhere from a blink of the eye to many seconds before it's ready to play smoothly.

[barahn]

Thanks for the info but I understand that very well. I can lay in a 3 second effect and its ready in a flash. I know all about the green play, it happenes so fast all the time I usually never see it off. My system is very fast. Dual 2.4 xeons with a gig of ram on supermicro board. I have a 150 gig Medea Raid RT array that gives me 62 mbs throughput. I have the VT3 setup on a dual boot so that no other programs are running in the background while I am editing. I shouldn't need to lay in a render effect under a 1.5 hour program AND my system is fast enough to put in a long effect and play it back almost imediately. I didn't have this problem with VT2. I could work on a project for 16 hours and play it back when I was finished and not have to go out of VT2 and back in again. Now I have to go out and back in again to get it to play. I have tested and tested different setting in the preferences and nothing I have done has changed the fact that ANY project will play just fine, even the complex ones, if I play in right after the program starts. Thats stupid. At least until someone suggests something that works, I can still get my weekly broadcast out. Still looking for help.

[Jim_C]

Hi Bill,

Oh ok, did you just install the patch? There is a couple people sayig the same thing on yahoo vtnt. Memory Leak kinda thing. Works fine after reboot but after working for a period simple parts won't play even with force render.

hit ctrl/alt/del and watch your memory performance in the task manager. Is it dropping constantly? Not everyone is having the problem, but obviously some are.

That's all I know......

good luck

jim

[barahn]

OK - I'll watch it and will post later

Thanks

[Jim Capillo]

Originally posted by barahn
I have a 150 gig Medea Raid RT array that gives me 62 mbs throughput.

Is that a typo? I have an IDE raid that gives me 79 mbs or so and I get occasional stuttering with that....... not with the SCSI raid, though. That delivers around 180-190mbs.

[barahn]

The medea RT that I have has a maximum throughput of 65 mbps. That is all. I have another 150 gig that I was told by Medea that if I put them together I can get better than 100. That is one thing that sucks about this type of Medea Raid RT. They have other ones now that are in the 160's but they are expensive.

[Jim Capillo]

Originally posted by barahn
They have other ones now that are in the 160's but they are expensive.

I didn't realize that Medea throughput was that slow, especially considering the expense. You may want to look at that as a possible cause - your system is pretty strong..... you should be kicking *** with it. All I can tell you is that I did get occasional stuttering (again, depending on complexity) when I had just an IDE raid, and that was consistently around 79-80 mbs on autoconfig.

Good luck,

[dml]

Jim, I believe the Medea raid array that Bill has is about 4 years old, so it doesn't represent the current capabilities of Medea's product line.

Matt Drabick

[Jim Capillo]

Well Matt, that aside, I think it could be a possible explanation for the problem he is having.

Bill, do you ever drop frames when capturing?

Looking at this logically, it would seem that the drives are the bottleneck with their low throughput. The system is certainly powerful enough - the drives just can't keep up.

Again, my experience with IDE drives confirms this. I had to force render depending on the drives speed capability to playback certain clips/transitions. As a matter of course, I rendered out projects as one clip to avoid these problems on playback...... these were mostly :30's though. I never rendered out anything longer than about 5:00 or so. These problems totally disappeared when I added the SCSI raid (I use the IDE mainly for storage now).

There are a couple of ways you could test this theory..... call your dealer (or a dealer near you) and ask if he has a SCSI raid you could plug into your machine. Drop that project on the drives and see what happens. If you've got the coin, buy a SCSI array (you can get them pretty cheap on eBay), or explore the SATA solution. There have been a few VT users who are reporting huge speed increases over IDE for real cheap money. Ask your dealer.

If you're running the lastest patch, it really does seem logical that the drives are causing the problems (given the autoconfig reports and your system specs). Not too many folks are reporting stuttering anymore.

I don't know about Jim's memory leak suggestion.... I guess it's a possibility, but being in the software, you would think there would be a ton of people reporting it.

Good luck!

[dml]

So, Bill, do you want to borrow a faster Medea RAID array from me and see if that helps?

You have to wonder why Bill's older (hell, ancient by today's standard) Medea seemed to work fine under VT2 but not VT3. Granted, it's just on the margins of acceptable performance for VT2, but it did seem to work well enough for him under VT2. Seems like the switch to VT3 is the real issue here.

When the VTNT (the first Toaster for Windows) first shipped, Medea jumped on the bandwagon and pushed their early Medea arrays as the perfect solution for the VTNT. We had some problems with the very first Medea arrays not being fast enough for the VTNT, but the one Bill has is fast enough, at least for the VTNT and the VT2. Again, the issue seems to be VT3.

Matt Drabick,
DigiTek Systems
Raleigh, NC

[Jim Capillo]

Originally posted by dml
Again, the issue seems to be VT3.

Matt Drabick,
DigiTek Systems
Raleigh, NC

Yeah, I'd agree. I'm no software engineer, so I don't know exactly what has changed between VT[2] & VT[3]. However, being backwards compatable with prior projects does not mean that the drive performance requirements haven't been raised.

It would seem quite possible that drive speed is the culprit here.:(

[Jim_C]

>>I don't know about Jim's memory leak suggestion.... I guess it's a possibility, but being in the software, you would think there would be a ton of people reporting it.


Agree,
The fella having trouble on the other forum has narrowed it down to only when he uses Type 1 DV files, but is system does slow to acrawl when he uses them requiring a reboot.


so dunno about that..

[barahn]

Yes, Matt is right. This drive is 4 years old. He should know - he sold it to me. It has been solid with no problems ever. It has seen alot of usage but still works like the day I got it. I would like to test Medea's theory and put an identical 150 gig drive on the B side, stipe them together and see if I get 100 plus gbps like they said I could. I have access to another one just like this one so I will try it. Otherwise I guess I going with a scsi array. I guess that would fix it because it is annoying. (Matt, i got your call and I really don't need that one you called about, thanks though.)

[barahn]

Ok - Did it.......didn't work. I did what Medea said - put one on A and the other on B. Still has a maximum throughput of 62 mbps. Tried them in a daisey chain, still around 62.(sometimes it reads out 61) So that is not an option but I did want to give it a try. I need to call Medea and tell them it didn't work like that - at least it didn't increase. Like I said before, Medea does make some much faster drives now if you have the dough.