Comments

farss wrote on 11/4/2003, 5:28 PM
Do it all the time.
If you have HT they'll both take about the same time so it should be faster. Upside is even if the times are added up you can leave it run while you sleep.
Chienworks wrote on 11/4/2003, 6:55 PM
About the only drawback is that your computer will bog down quickly, so if you want to do something else while you're rendering it will quickly become very sluggish. If you don't want to do anything else then go for it. If you have Windows XP then you can set the priority for various jobs. When i set up multiple renders i set each instance of Vegas to "Below Normal". That way they'll eat up all the processor time available when it's just them running, and they'll gracefully relinquish processor cycles when you want to do something else. I've often got half a dozen renders running this way, and have had over a dozen going at once.

When setting up multiple renders, i would also suggest that you open all the projects and go to File / Render As, set up any rendering parameters you want for each, but don't click Save. Then when you've got them all to that point go back and click the Save button on each. If you've already got a few renders running before you try to open the next it can become a painful process to get the next project open.
musman wrote on 11/5/2003, 3:39 AM
You know, I've only tried this once rendering our 2 avi versions of the same project. They turned out to have flaws and I haven't tried this again. Is this normal? Also, would you run into any problems when doing something like rendering in After Effects or Boris and rendering in vegas?
Grazie wrote on 11/5/2003, 3:48 AM
This is an excellent thread. I'm reading it with more than just a passing interest . . . wot with my new MONSTA coming . . . . I'm looking for some multi- render "recipes" I could potentially use.

Grazie
Maverick wrote on 11/5/2003, 8:11 AM
Further to my initial query I attempted to do two 40 minute projects together.

I started one at 8pm and the other at around 10 pm. Previous attempts at rendering ahve produced times of around 10 hours.

This morning the first one had been going for over 15 hours with 71% completed and over 20 hours more to do. The second had only 35% after 13 hours.

Not good.

Cheers
Mandk wrote on 11/5/2003, 8:38 AM
I did one two hour project rendered by itself with nothing else running on the computer that took 36 hours. (a lot of stuff going on in the video - rerendering it a second time with a few fixes was only about 3 and one-half hours).

Two at once seems like a distant dream to me!
Grazie wrote on 11/5/2003, 11:11 AM
Sorry but, might have missed it, what PC specs are we talking here?

Grazie
Maverick wrote on 11/5/2003, 11:41 AM
The system I used is an ABIT KR7A-RAID ATA 133 with 768 MB memory, Athlon 1700+CPU. Three HDD; 20GB for system & Apps, 60+80GB for video. All three are ATA133 7200 but 1 cannot get HDD0 to DMA mode although the other two are.

Liam_Vegas wrote on 11/5/2003, 12:09 PM
Hmm... that DMA (or lack of it) will be killing the performance.. IF.. your temp files are going to that drive.
shogo wrote on 11/5/2003, 12:29 PM
I do it as well, I have 3ghz p4 HT and 1 Gig ram. Hell I even capture and render projects at the same time with no drop frames and the video that gets rendered never messes up.The most I have done is rendering from to vegas instances (converting full DV into WM9 256k) and capturing another clip given the render's took a little longer but I had to have 20 videos converted to WM9 in a 256k version and a 56k versions over a weekend worked very well.
Maverick wrote on 11/5/2003, 12:53 PM
No.

temp files, prerendered, etc. all on HDD1. Main files on HDD2.

Grazie wrote on 11/5/2003, 1:40 PM
shogo - What are typical render times?

G
johnmeyer wrote on 11/5/2003, 3:57 PM
I did a quick test:

1. Put checkerboard on top track, put some DV video on bottom track. Set them both equal to five seconds.
2. Set transparency on the checkerboard to zero so that the video on the bottom track shows through the black checkerboard.
3. Save the project (.VEG) file.
4. Render using the NTSC DV template.

On my 2.8 GHz P4, this took exactly 20 seconds. (I repeated to make sure).

I then opened a second instance of Vegas, opened the same project, and set both instances to start rendering. I started the first instance rendereing, and then immediately switched to the second instance and started it. The instance that was in the foreground completed in 25 seconds. The instance in the background finished in 38 seconds. I repeated several times, rendering to different disk drives to see if that made a difference (it didn't); and switching to the background instance as soon as the foreground instance finished rendering (no difference).

Total time to render two identical VEG projects using only one instance of Vegas: 40 seconds; using two instances, the total time increases to 63 seconds, almost 50% more time.

Conclusion: If you want to finish your rendering in the fastest possible time, multiple instances are a lousy way to do it. You are much better off using a batch script that renders them one after the other.

Here are some links to various kinds of batch scripts.

http://www.sundancemediagroup.com/help/kb/kb_download.asp?id=137
http://www.sundancemediagroup.com/help/kb/kb_download.asp?id=123
http://www.ayizwe.net/VegasScripts/BatchRenderGUI.js.txt
http://www.ayizwe.net/VegasScripts/DirectoryConverter.js.txt

(P.S. Added after initial post: Please see my later post for a correction to this conclusion).
Chienworks wrote on 11/5/2003, 4:29 PM
John, i've got one tiny quibble. ;)

Where did you get the 63 seconds from? Remember that the two rendering times overlap, not run sequentially. You would only add the 25 and 38 together if you didn't start the second instance rendering until the moment the first stopped. You said you started the second instance "immediately". You didn't say how long it took you to switch from starting the first to starting the second, but let's guess you could do that in, say, 5 seconds. That means that from the time the first started until the second ended would be 43 seconds, not 63. This is barely worse than 40.

I've done lots of renderings that would have taken about 2 hours individually. When running four of them simultaneously the first might end in 5 hours, the second in 6 hours, the third in 7.5 hours, and the fourth in 8.5 hours. This doesn't mean that it took 27 hours to render all four. If it took me a few moments to switch from starting the first to starting the last, then the total time is about 8.5 hours (the length of the last render to finish). This is only slightly more than the time to run each render sequentially.

In fact, it's probably faster for me because running them sequentially would require that i pay attention enough to notice when each render ends so that i can start the next right away. Running them simultaneously means i can go away and do other things, or sleep, for the entire time. Of course, using a batch rendering script would avoid the problem. However, the difference in simultaneous rendering time vs. sequential rendering time is very minor.
Grazie wrote on 11/5/2003, 5:26 PM
Could sombody put up a "Graph" of all this fabulous experimentation - it is too important to be left as just words . . . it would be a valuable "Performance-Check-Graph" - yeah?

G
johnmeyer wrote on 11/5/2003, 5:54 PM
Chienworks,

Color me stupid. You are absolutely correct.

It took less than one second to start the second render, so the total render time for the two was 38-39 seconds. This is slightly less than the 40 seconds for doing them sequentially.

The various batch scripts are still useful if you need to render the same project to several different formats, or if you have dozens of AVI files to render to MPEG, or if you have dozens of VEG files to render to several different formats (there are batch files for each of these cases). However, if you just need to do a few renders, I completely revise my earlier comments; opening several instances of Vegas is the way to go.
JL wrote on 11/5/2003, 6:10 PM
I simultaneously ran 2 rendertests on 2 instances of V4. Result 2:07 and 2:09 respectively, i.e., 2:09 total for both. A single rendertest for my machine is 2:00. [2.4 GHz dual xeon]
Maverick wrote on 11/5/2003, 6:57 PM
>My experience is the same: slightly longer, but the advantage is that you >can start them all on their way and go to lunch.

Or a weekend if my experience is anything to go by. This evening I will reboot first, then render the two projects. Each project, as I said, is around 40 mins but there are quite a few changes in opacity, added media, track motion, etc. so I am expecting them to take a while on my 1700+ system.

I'll report back tomorrow and let you know.

Night all.