VEGAS PRO 8a - MAJOR PROBLEMS WITH QUAD CORES

Grazie wrote on 1/9/2008, 11:50 PM

Bob, please don't dismiss my attempt outright - I did what I did as the only way I knew and could think of assisting our chum here!!

And you know what? I don't appreciate being turned off this way. I am NOT a programmer and I struggle much when it comes to the more complex and rarefied elements of our activity. Does this then preclude me from attempting to assist? Or should I just wander off thinking - "Ah, that's for those 'others' to sort out!" - But what I can do is plug-in some words into a Forum and see what comes back. I can do that!

Oh, and BTW, why WOULDN'T it tell us anything? If you are going to counter what I attempted, please supply me with the relevant logical reasons why you believe it would be fruitless? Sorry, your word "nothing". This at least this way even I could learn too.

As it happens you are asking a more precise question - HDV in V8.0a from a Canon XH A1 that shot 24p.

But again, my searches, over on the DVInfo site, was a top level attempt to "catch" the sprats as well as the mackerels in my trawling net. And even so, this was a low haul.

And still, I can't understand just why the fora aren't crawling with this issue? Is it just a unique set of settings? How many people, using this camera, in THAT shooting mode edit in VEGASPro8a ON a Quad? If we aren't seeing these hits then maybe not a lot of these circumstances have been created - yet?

Darren, is it possible that one of the QUADS is duff? And yes ProTitler is a really neat entree for SONY into this part of editing. It's a great first showing of what could become a show stopper!

Grazie

Warren Hedges wrote on 1/9/2008, 11:52 PM

Hi Darren

Have you tried systematically changing / removing sticks of RAM and seeing if the problem persists. Sometimes a bad stick of RAM can cause all sorts of grief.

HTH
Warren

P.S.
(bigrock) >> This is BS, there is no problem with Quads and Vegas. It's silly for one user to make this claim against the product when it's just him.

This does not help Darren who is obviously having severe problems when trying to get a job done. Its not a good situation to be in.

Grazie wrote on 1/9/2008, 11:54 PM

Warren, excellent idea. RAM can be a pig.

Apologies for the mixed animal metaphor!! - Just saw it . . Lol!

Grazie

megabit wrote on 1/10/2008, 1:32 AM

The fact is that after I just built my Quad PC and was playing with overclocking it, at some extreme settings (like 3.6GHz) it seemed stable with torture tests (Prime95), but it crashed when rendering in Vegas (7.0 at those days). I soon discovered it's not those testing/torturing apps, but Vegas that makes the best testing application... And gues what: after a couple o weeks, I discovered my RAM sticks have 2.10V printed on them, while I was using default voltages of my mobo BIOS (1.8V). Once I set it at 2.10V, Vegas (or anything else) never crashed again...

farss wrote on 1/10/2008, 1:37 AM

"Oh, and BTW, why WOULDN'T it tell us anything?"

If you're mining data looking for a correlation and find none there's one of two logical conclusions:

1) There is no correlation.
2) The variable that has a correlation isn't in your dataset.

Now if we remove the noise, such as people saying they're not having problems but they're not even editing HDV we do find a small number of people having issues and they are using quad cores.
What we don't know is if they're editing 24p footage from a Canon camera. Why don't we know, mostly because they haven't recorded that piece of data, searching our dataset is useless to us.
So we need to gather more data and the best way I can see to do that is to ask very specific questions, that helps in two ways, getting more data and reducing the noise in our data. Maybe I should have just started another thread.
This is not an easy task and probably very likely a fruitless one but it's an avenue worthy of investigation. The bad news is if this does turn out to be the determining factor then I don't think there's anything that can be done to help. He's hardly likely to go back and reshoot the footage on another camera.
-----------------------------------------------------------------------------------------------
In the end these problems are almost impossible to diagnose, even more so remotely. I've made several suggestions to Darren to try to help narrow his problem down, some of which he seems to have ignored. I've offered to take a day off from what I'm doing to sit down with him and see if hands on I can find something. I've offered to load his entire project onto one of my machines to see if he can get it 'out the door'. I've emailed him asking specific questions but so far no reply. If I seem a little terse it's because I'm frustrated, I've offered as much help as I possibly can, probably way beyond the call of duty.

Re the RAM thing. I suggested that the first time he reported this problem. I believe Darren has already tried that to no avail. It's a very good idea as I know several people who've had exactly that problem but in this case it seems to be not the problem. Heck it could even be his power supply for all I know. Dropping the number of cores would reduce the demand on the power supply.

What I think is the biggest problem here is a tired, frustrated editor whose trying everything while trying to get a job out the door. Bad combination when it comes to thinking clearly. Thinking back over all he's said I'm pretty confident there is something wrong with his machine. What it is though is very difficult to guess remotely and I doubt he's in any mood to be swapping parts around.

My best advice to Darren, tell the client you've got a major problem, give yourself some brathing space, take a very long walk, get a good nights sleep and start afresh. It's amazing what a clear head can find.

Bob.

megabit wrote on 1/10/2008, 1:51 AM

Darren, if I were you I wouldn't disable anything in BIOS - try setting affinities for offending applications only using TaksMan (right click the process->set affinity); this way you can choose any combination of cores to use. Perhaps this will give you some new hints?

Piotr

Grazie wrote on 1/10/2008, 2:18 AM

Thank you Bob. Now I've learnt something! - Cheers! - G

DJPadre wrote on 1/10/2008, 8:35 AM

"Question for ALL the people saying they're not having problems with Quads and Vegas. Are any of you editing 24p HDV from a Canon camera, more specifically the XH A1? Are you doing a longform project, mulitple tracks etc with that footage."

Ive got an A1, but ive been shooting sports and truucks with it mostly so im hitting the 1080i.

If ud like a test unit, drop me a line ill bring my cam around and u can kick it around mate
I havent edited 1080p from teh A1 and i havent instaled my quad as yet. In fact im waiting to do that tonight if i can get away with leaving afew emails til tomrorow... lol

rmack350 wrote on 1/10/2008, 10:36 AM

Well, I'm grasping at straws with DEP/XD/NX, but it kind of fits the description, as it's supposed to dump the errant program immediately and unceremoniously, especially if it tries to write into reserved space.

What motherboard are you using? I'd like to go get the manual, just to satisfy some curiosity. Usually a BIOS has a key combination to get into almost all settings, and another key combo to get into ALL settings.For instance, if the key to enter BIOS is F9, the modifier might be CTRL+ F9.

DEP doesn't particularly fit into the quad core idea except that processors haven't always supported it, nor have motherboards, and Windows didn't support it until SP2. So you might have just hit a perfect storm where Vegas has a code flaw that shows up in your setup, which supports some newer features. Granted, whatever it is, SCS should fix it but they may not be able to tell what it is and even then they may not be able to fix it in the next two weeks, So it'd be really crucial to come up with a workaround that gets you working.

Hopefully you've called Bob. He's in your region and has offered to try things out on another system. If there's any way you could just move the job to a stable system that'd be a great solution.

I had similar problems with Dreamweaver several years ago. There was a longstanding flaw in code that was only noticeable with a big site because the big site slowed things down enough to see a problem. It took a bit of complaining but eventually the Macromedia team got a great engineer in touch with me and he and I work things out within 2-3 weeks. I was extremely happy, but it took a year of unhappy complaining to get the attention.

In the mean time you need to get a working system and finish the project. Could be a loaner, could be a rental. You might even check with your business insurance policy, given the dire nature of the problem. Generally, in the film business you have to learn that if something is broken mid-job it's faster to get a replacement and fix it later.

Rob Mack

rmack350 wrote on 1/10/2008, 10:41 AM

No, it's not BS. He's had a handful of people with quad cores claim to have the same troubles, and disabling cores seems to alleviate a lot of the trouble. Natural to assume it's VP8a on a Quad that's the problem, whether it really is or not.

This is the second thread that Darren has run on the problem. Go read through them and you'll see more people with trouble.

Not helpful to just say "it's BS"

Rob Mack

rmack350 wrote on 1/10/2008, 10:45 AM

He's swapped RAM around. I wasn't clear about whether he tried various combinations of 2 GB but he definitely rotated them around in a 4GB configuration. I'd certainly try the 2 sticks/2 GB config, though.

Rob Mack

rmack350 wrote on 1/10/2008, 10:48 AM

That's good information! Especially with 4 sticks, which I hear is a little more trouble-prone than 2 sticks.

Rob Mack

rmack350 wrote on 1/10/2008, 10:56 AM

I'll say it again. The best bet would be to find some other machine to finish the project on. You just can't troubleshoot this crap when you're under the gun and if another machine can do the job he'll look a lot better to the client if he swaps machines.

At work, we get our setups from an integrator that specializes in edit systems. They can often loan us a demo system to work with when PPro/axio crashes and burns. Overall it costs more, but it's better to have someone to dump the problem on. Cost of doing business.

Rob Mack

rmack350 wrote on 1/10/2008, 11:06 AM

Megabit,

Here's the previous thread where that was covered. http://www.sonycreativesoftware.com/forums/ShowMessage.asp?MessageID=568787

Rob Mack

Darren Powell wrote on 1/11/2008, 2:01 AM

Thanks for all your hard work on my problem. I have certainly exchanged my memory sticks as 2GB and 4GB banks... and have also tried ALL combinations in BOTH available channels ... nothing helps ...

I've just downloaded the latest version of NEO HDV which is dated a week ago 2008 ... just in case there was a problem with the Cineform Codec ... interestingly the latest version of NEO allows me to (FINALLY!!) convert my m2t files into Cineform wrapped .avi's ... that's enough to tell me that there are problems at Cineform (which I've known for some time... the capture in previous versions of NEO HDV have been crap) ... BUT it also allows me to re-work the project over the next few days to remove ALL of my m2t files (which as you've informed me are full of Main Concept error corrections which could be giving Vegas Pro 8a grief) ... and convert them to the latest Cineform codec. The only reason I'm going to do this is because I'm desperate to deliver something to my client and previously I've been able to render a 720p m2t Main Concept file ONLY from a sequence of clips that were encoded with the Cineform codec.

If you can make any sense of this you're obviously a legend.

Thanks again for your help.

I'll keep you posted on developments.

Darren Powell
Sydney Australia

Darren Powell wrote on 1/11/2008, 2:53 AM

Thanks Bob, sorry I haven't responded yet ... I've literally been inundated with lots of helpful people giving me advice! Which is great ... and thankyou very much! Forums are great ... keep up the good work!

The reason I've been unresponsive is that I've been testing the majority of the different options being suggested by Vegas users.

I've discovered some interesting anomolies in the process ... but none that has given me a clear solution.

Many users are having problems with Vegas crashing (in much the same way as my Vegas crashes) if they have JPEG images on their timelines which are CMYK instead of RGB.

I've experienced lots of problems like this in the past including trying to make a company brochure in CorelDraw 3 (years ago!!) which wouldn't process print because it wasn't CMYK! My images in the brochure were RGB! ... Same problem ... reverse the situation!

Anyway ... my Vegas Pro 8a project still refuses to render ... and I don't have ANY stills whch could be causing a CMYK or RGB confusion.

My project is a feature film which includes Cineform codec clips and m2t clips ... the only reason I went to m2t about 4 months ago was because NEO HDV was causing me problems with capture ... ie: I would capture one 1GB clip and the next clip was about 500 little pieces of broken up into tiny little files of crap!

Anyway ... I've received some information from various users saying that there may be problems with Main Concepts m2t codec (built into Vegas of course, like the Cineform codec) which could be causing Vegas Pro 8 some grief.

Ie: the Main Concept mpeg2 codec is based on a lot of error correction which could be causing problems!

So, based on this information I tried to convert all of my m2t captured footage (which was a lot easier to capture than the NEO HDV stuff) back into a Cineform codec.

The only reason I decided to do this was because the footage on my timeline which was exhaustively captured using Cineform's NEO HDV program ... rendered out perfectly from the Vegas 720p HDV render option using the Main Concept codec.

BUT ... since I made this discovery a few days ago ... I've been unable to convert my m2t footage (captured directly into Vegas) back to the Cineform codec using the conversion program in NEO HDV.

BUT (again) Here's the interesting thing ... I downloaded the latest version of Cineform's NEO HDV (HDLInk) program only because I'm at my wit's end ... and guess what ... I CAN FINALLY CONVERT ANY M2T FILE ON MY RAID TO A CINEFORM AVI ... WITH NO PROBLEMS AT ALL ....

Here's what Cineform write about the latest update to their Cineform Codec ... (straight from their website tonight - I'm four days late ... but you get the gist!)

-------------------------------------------------------------

January 7: v3.2.3 (Build 150a)

Fixed à Driver conflict with other third-party applications

-------------------------------------------------------------

It's very interesting that AS OF TONIGHT ... I can finally convert problematic m2t files to the Cineform codec without any issues ....!!!

Maybe this will help me finally render a file which I can use for my feature film project out of Vegas Pro 8 ... it hasn't so far ... but I'm finally hopeful of ridding my project of m2t files (apparently Main Concept has been bought by DIVX????? ... is that true???) if so ... I really think that producers like you and I should be holding Software manufacturers ... and CODEC manufacturers to account for the delays being introduced into ... otherwise seemingly simple digital video productions.

And yes Bob, thanks for your input ...

-------------------------------------------

My best advice to Darren, tell the client you've got a major problem, give yourself some brathing space, take a very long walk, get a good nights sleep and start afresh. It's amazing what a clear head can find.

---------------------------

I've told the client, taken a long walk ... got a good nights sleep ... and cleared my head ...

Now can we get on with telling SONY and CINEFORM and MAIN CONCEPT ... (and GOD FORBID) ... the APPLE MOV people to start talking to each other so I can finish my film

------------------------------

Cheers,

Darren Powell
Sydney Australia