VideoHelp Forum




+ Reply to Thread
Results 1 to 8 of 8
  1. Member
    Join Date
    Jan 2004
    Location
    United States
    Search Comp PM
    I'm encoding from uncompressed avi to mpg2 for the purpose of making instructional guitar DVDs.. I have an athlon 1800xp processsor that runs at around 1.54ghz. My NLE is Vegas 5.0.

    I encode from teh timeline in vegas using it's integrated mainconcept encoder, for an 8 minute segment it takes about an hour to encode! Granted there is some color correction as well as a small Picture in picture.

    When I frameserve to CCE 2.67, I can see that when encoding the part without PIP, it goes at about 30-40% of realtime, but when the PIP pops in, it slows to 10-13% of real time...

    Are these numbers accurate for the processor I'm using? Of course the processor is maxed out at 100% while it's encoding, but I wonder if something is wrong or if that sounds about right for a 1.5ghz processor?

    Thanks,
    Mike
    Quote Quote  
  2. Member daamon's Avatar
    Join Date
    Jun 2003
    Location
    Melbourne, Oz
    Search Comp PM
    Hi sdsumike619,

    I used to have the same kind of set up as you have at present - 1800XP and 512 RAM.

    As a rough comparison, it used to take TMPGEnc around 2.5 - 3 hours to encode 30 mins of DV AVI (no PIP or filtering etc.) to MPEG2. I was using a fairly high bitrate (VBR max 9,000 and average 6,000) and had the motion search set to "high".

    So, I'd say that your timings are in the right ball park... If you want it to go faster - faster CPU and up your RAM to 1Gb.
    There is some corner of a foreign field that is forever England: Telstra Stadium, Sydney, 22/11/2003.

    Carpe diem.

    If you're not living on the edge, you're taking up too much room.
    Quote Quote  
  3. Member
    Join Date
    Jan 2004
    Location
    United States
    Search Comp PM
    yeah don't I know it, but it seems that CCE is really fast compared to the mainconcept encoder if you're going directly from an AVI file, somehow, i think this frameserving slows it down, anyone else?
    Quote Quote  
  4. Member daamon's Avatar
    Join Date
    Jun 2003
    Location
    Melbourne, Oz
    Search Comp PM
    I think I'm right in saying (in fatc, I'm 99.9% sure) that frameserving is done via the memory - no temp writing to disk.

    So it'd be more accurate to say it's your memory limiting the speed of the frameserver...
    There is some corner of a foreign field that is forever England: Telstra Stadium, Sydney, 22/11/2003.

    Carpe diem.

    If you're not living on the edge, you're taking up too much room.
    Quote Quote  
  5. Member
    Join Date
    Jan 2004
    Location
    United States
    Search Comp PM
    hmm that is odd because when I look at my memory usage during encoding when frameserving, it's no different than when using the integrated encoder in Vegas... but the CPU is definitely maxed out
    Quote Quote  
  6. The CPU is always maxed out when encoding.

    This tells you that the CPU speed is the bottleneck!

    If the CPU usage were undel 100% then you'd have a different bottleneck to hunt down. The fact that cpu is at 100% says that the cpu is the bottleneck

    Having said that with a P4 3.0 800Fsb w.HT, 1 gig Dual channel ram etc.. Still runs at 100% usage. However using this set up my proicessing time are considerably shorter than when I was running a 2.66 P4 533 FSB setup.

    Upgrade your cpu and shorten your times or do what I used to do, start encoding overnight, let it finish whilst at work and come home ready to author.

    Cheers
    Quote Quote  
  7. Its not the memory and its not the frameserving. Its the fact that Vegas has to first render your project to uncompressed avi. When framesrving or using the built-in encoder, Vegas has to render each frame before encoding to mpeg. This is why you get the slowdown with the PIP bit as rendering this is more complex. If you have time and disk space, render the whole thing to uncompressed avi, then encode to mpeg seperatley. Add the two times together and the final figure should be pretty close to that you get when doing it in one step. Using CCE may be a little faster than Mainconcept, but make sure the settings used (bitrate, motion search precision, number of passes etc) are as equal as possible before you make any performance comparisons.
    Quote Quote  
  8. Member daamon's Avatar
    Join Date
    Jun 2003
    Location
    Melbourne, Oz
    Search Comp PM
    Originally Posted by bugster
    Its not the memory and its not the frameserving.
    That'll teach me not go drinking on a Tuesday night and then try to offer advice and help on a Wednesday morning... Sorry for the misinformation. Still, I suppose we've both learnt something new.
    There is some corner of a foreign field that is forever England: Telstra Stadium, Sydney, 22/11/2003.

    Carpe diem.

    If you're not living on the edge, you're taking up too much room.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!