More Vegas Internal Settings

jabloomf1230 wrote on 3/14/2008, 10:24 AM
As has been discussed, if you hold down the Shift key while selecting Preferences from the Menu, you will get an additional tab labeled Internal. This tab allows you to change a wide variety of hidden Vegas settings, in addition to the recently discussed default Maximum Preview Ram setting of 1024 MB. Here are a few other settings that I've noticed and/or tinkered with so far:

1) There is a setting to change the minimum preview frame usage. It defaults to 5 frames. If you set it to 1, it gives a smoother preview, in some circumstances, especially when using Preview settings like Best/All. The results are not profound nor are they consistent. This needs more investigation.

2) There are two settings related to Inverse Telecine. One enables it. The other uses it. Enable defaults to FALSE. Use defaults to TRUE. Has anybody played with these settings and in what context? I assume that they allow conversion of 29.97 FPS to 24 FPS, but why are there two settings?

3) There is a setting for the maximum amount of RAM that Vegas will use (not including the Preview RAM). The default is 384 MB (which offhand seems low for a program like Vegas). When you look at RAM usage by program, using a utility such as Process Explorer (which is just a glorified version of the Windows Task Manager), you see that Vegas tends to use right up to this maximum value, but also has additional memory swapped out to the Windows Page File. By raising this value to 512 MB, more RAM was used and the Vegas Page File usage dropped by the same amount. This may or may not influence real world performance, because it is possible that only transiently used DLLs (Vegas libraries) are swapped out to the Page File. Again more testing is needed.

Has anyone else played with these or other settings?

Comments

rmack350 wrote on 3/14/2008, 11:42 AM
I assume most of these internal prefs are there for Sony to use in QA testing. In the end, they decided on a stable configuration and that's what they shipped. I imagine they settled on settings that provided a happy average.

I don't think they're under any obligation to provide support to someone who changes a hidden pref. But then, I haven't heard anyone raving about the support from SCS.

The setting you talk about in number 3 is named "Memory needed by Vegas (MB)". That doesn't really equate to "the maximum amount of RAM that Vegas will use" but it does seem like it gets subtracted from the total amount available to Vegas. So if you set your internal Ram Preview Limit to 2048 then Vegas will offer you up to (2048 minus 384) for RAM preview on the Video tab. So perhaps the Memory Needed pref is just there to govern how much memory you can allot to RAM preview.

Rob

NickHope wrote on 3/14/2008, 12:22 PM
I always set "Use new event delete ripple logic" to "FALSE" (near the bottom of the list).

They introduced this at version 6 or 7 (I forget which), but I prefer it the old way. I like ripple editing but I hate having the gap after an event close up when I delete the event.
Chienworks wrote on 3/14/2008, 12:36 PM
I'll also point out that it's easy to hose your Vegas installation if you tweak some of these settings wrong. Play with these at your own risk. Don't expect Sony to help you out if you do. If you really mess things up the best way to get back to functionality is to do a "cold-reset" of Vegas. Quit all instances, then hold down Shift+Ctrl when starting it up again. This will reset everything back to the original installation settings. Beware though, this even resets all the changes you wanted to keep.
busterkeaton wrote on 3/14/2008, 1:11 PM
With the caveats mentioned above, I wonder if QUAD folks would do better with more render threads. The internal preferences let you use up to 8 render threads.
pmooney wrote on 3/15/2008, 9:05 AM
Dear Buskerkeaton,

Your suggestion inspired me to go in and change the internal settings for rendering threads. I set the max to 8 and the minimum to 4. (I have a dual quad core Xeon in a Dell Precision 690, 4gigs of ram).

My render time on John Cline's rendertest improved dramatically, from 2:10 on the default settings to 1:19 with the new settings. My cpu finally showed near 100% usage on the render, while it was only using about half that under the default.

It's wonderful having so many people on this forum asking great questions and trying to get the max out of their set-ups.

Happy Saturday!

Patrick
NickHope wrote on 3/15/2008, 9:44 AM
Wow! I wonder if there's a downside.
DJPadre wrote on 3/15/2008, 10:29 AM
has anyone found any significant performance improvement by accessing the Ram directly as opposed to running the render through pagefile?
blink3times wrote on 3/15/2008, 1:30 PM
has anyone found any significant performance improvement by accessing the Ram directly as opposed to running the render through pagefile?

I run 8gig ram with no page file... although I have not timed anything... it DOES seem faster than with pagefile.
riredale wrote on 3/15/2008, 2:52 PM
The single change I've made in the Internal tab for years is to reduce the magnetic "snappiness" of adjacent clips on the timeline. Open the Internal menu, and type the word "snap" in the search box at the bottom. Select the "Snap Width" item that comes up and change the default 7 to something less aggressive. I much prefer "3."
Darren Powell wrote on 3/15/2008, 4:51 PM
Very interesting ... any news regarding modifying settings to improve render / RAM / PF usage appreciated ...

D.
TheHappyFriar wrote on 3/15/2008, 4:56 PM
on my AMD Phenom 9600 the # of render threads above 4 doesn't seem to make a difference. The CPU is always using 4 cores 100% & both my render tests ended @ the exact same time.
Kennymusicman wrote on 3/15/2008, 5:16 PM
You will only get a benefit if your system can handle more than 4 thread. Hence a dual quadcore (having 8 threads) should see a benefit. If your processor is running at 100% - then you won't gain anything either - afterall, it's already running at max.
Darren Powell wrote on 3/15/2008, 6:18 PM
Wow ... I just rendered out one of my 10 minute chunks from my film project with 8 threads selected ... my BadAxe Quadcore screamed through the render with lots of audio tracks and FX / Colour Corrected m2t files ... not only did it do it quickly (I haven't done a comparison to 4 threads yet but it was noticeably quicker ... ) but it also did the render without a 'sorry I ran out of memory error' ... this is definitely different ... even the pagefile usage in Task Manager seems different ... it's not climbing up and up and up and then falling over ... it has little peaks and valleys like a ripple all the way along ... but it stays very consistent ... I've never seen that before and the render is just now arriving at 100% without a problem ...

mmm ... I hope I'm not imagining things ...

Darren Powell
Sydney Australia
Darren Powell wrote on 3/15/2008, 6:27 PM
Nah ... definitely an improvement on render speed ... but I just got a 'sorry - reason for error can't be determined' while rendering out a different 10 minute chunk ... maybe I'll restart the machine and try again later.

D.
TheHappyFriar wrote on 3/15/2008, 7:27 PM
ou will only get a benefit if your system can handle more than 4 thread

I misread his statement. I thought he had a duel core. :D
NickHope wrote on 3/16/2008, 5:37 AM
>> on my AMD Phenom 9600 the # of render threads above 4 doesn't seem to make a difference. The CPU is always using 4 cores 100% & both my render tests ended @ the exact same time. <<

Likewise my Q6600.

Thanks for the snap tip riredale! I've set mine to 5 for now.
KelvinWorks wrote on 3/16/2008, 12:44 PM
Yes, a Q6600 only can run 4 threads as it only has 4 processors. The Q6600 does not support Hyper Thread technology (one processor-two threads). The Dell Precision 690 has two quad core processors so it will have a total of 8 processors capable of processing 8 threads.
rmack350 wrote on 3/16/2008, 1:07 PM
It all got covered in other responses. Never mind.

Rob
jabloomf1230 wrote on 3/16/2008, 4:26 PM
Had anyone had a chance to see how this affects preview smoothness? The default is 5.