Vdumod+divx5.11=slow

[OMINUS]

well i cant understand this
gknot uses vdmod and it takes about 8hours to make a divx from a dvd movie something which is reasonable
but when i use only vdmod to make divx from a captured avi file it takes 15 hours for the first pass only !!!!what causes this big difference in time encoding?

for the captured avi in vdmod i use only the resizing and the deinterlancing filtes-i resize to 512* and i use divx5.11 in standard mode pro feat off

i cant understand why the time encoding for the captured avi(pegaus mjpeg codec) is so slow opposed to the time for the dvd movies
what causes this?

any idea?

[jggimi]

Your captures are of content that has a 4:3 aspect ratio. I'm guessing the majority of your DVD encodes are widescreen. The captured content has many more pixels per frame. (As you are probably aware, VdbuMod is Gknot's encoder for DivX 5.)

[OMINUS]

true but in 4:3 dvd movies i encode with 9fps but in the avi i encode with 2fps!!with divx5.05 and vdum i was encoding captured avi with 8fps i dont know if divx or vdubmod is the reason for this slow encoding time

[jggimi]

Your question seems to now be "Why is 5.1.1 slower than 5.05?" If that's what your question is, it has been answered many times already in the forum. (Short answer, different technology codec.)

[Angelus]

I think another reason why it takes longer to encode captures than dvd's is because of the large size of captures compared to the size of normal dvd's (it may be a whole read-write issue). It may also have to do with the file structure of the input source (VOB as apposed to AVI). Also with a captured file, you're going from an uncompressed source to a compressed source, whereas dvd is compressed to compressed. That probably makes a big deal difference there too. Just my guess tho.

[jggimi]

AviSynth sends uncompressed video to the encoder.

[dTb]

If your using filters in vdubmod you will incur a performance hit compared to using similar filters within avisynth.

[OMINUS]

so can i create an avs script for my captured avi which can be used from vdubmod?will i see improved encoding time?

[jggimi]

AviSynth scripts (.avs files) can be edited in a text editor, but if opened by an .avi player (or an encoder) they behave like uncompressed video. You're familiar with Gknot, since you use it with .d2v source files. The .avs script created by Gknot (including all filters) is given to VdubMod for encoding in "Fast Recompress" mode. Filters in Vdub(Mod) can only be used when in "Full Processing" mode. Full processing is significantly slower than Fast Recompress.

You can use Gknot as your .avs creator. Gknot supports .d2v, .avi, and .avs as sources you can open. If you capture segmented .avi files in Vdub, you can create a one-line .avs script using SegmentedAviFile() to point to the set of .avi files.

If you currently use VdubMod to edit your captures -- such as setting start and end points, and removing commercials -- you can use VdubMod as an .avs editor. It can add Trim() filters to your script automatically as you delete frames. Merely use "Save" rather than "Save&Encode" in Gknot, use VdubMod to add the Trim() filters, then either encode manually in VdubMod or use the Encoder tab in Gknot.

[Shandra]

Uh... after a 3/4year of own capture experience I always thougt that the captured source isn't that encoder friendly as most DVD material... less grain, less distortian... fewer to zero jumps along the Y-Axis on DVD material + regardless if AVISynth sends uncompressed material (compressed would make no sense, wouldn't it?) it is already mpeged material and my speculations where that with that the encoder can cope better (I never asked why, just accepted it as beeing the cause - so I too well may be wrong on that - but as encodes from MJPeg (yes I bought it from Pegasus) compared to encodes from HUF seemed to be somewhat faster on the long run - but just marginal)... So whereas my DVD encodes ussually were around 20-30fps my captured encodes where always down to 6-24... (as most of the time with those old VHS captures I used a lot of filters) (damned experience last night was that not reading the guide before I used the slowest setting on the new divx codec and was frustrated that a 3pass took me 25hours whereas with the old divx codec I expected something around 9-12 hours.... sigh... RTFM again ;))

[jggimi]

I believe the key factor in reduced encoding speed from captures is aspect ratio. Standard TV signals are 4:3, most DVD encodes are from widescreen. There are many more pixels per frame with 4:3, reducing your encoding fps.

[OMINUS]

well the difference is that in the oldest vd versions i could get about 15fps to encode a capture avi using divx5.11-now with the latest version i get 2fps!!it seems it is true that older is sometimes better :P

[jggimi]

I believe you are comparing apples with oranges, OMNIBUS. Unless I've completely misunderstood, Shandra was describing different encoding speeds with different content, you are comparing the same content type with different encoders.

[Shandra]

jgimmi: I'm doubting that as a cause just because most of my VHS captures are widescreen... and a true 4:3 encode to a cropped widescreen one on captured material never had made much difference on my system on speed... whereas the difference between DVD encode and capture encode is well noticeable... and also I noticed no difference on my 4:3 sourced DVD to the widescreen ones in speed... but I never done any in depth tests and all my encodes vary so much in settings that they are not really comparable (dvd compared to dvd maybe, also captured to captured... but crossover compares... no, too many differences... but then again, even on the few ones where my captured material was so good in quality that I used my dvd settings - they where in general slower... and as most of my latest dvd encodes (old childhood TV series) which are most closely comparable to my good quality source captured material (both 4:3, both needing some noise reduction) the captured where slower - and that I mostly blame on the jitter I always could just minimize on my captures (yeah, tape tracking)... and that is introducing motion on captured source whereas the dvd material has none....
I never had the chance to do a encode from DV material - I guess those could be compared to dvd encodes... but just my speculation....

P.S.: Bad time of year... I am in the "lucky" situation that I could well setup a true comparisson - because I own the first two series of my most favourite BBC series (RoS) on VHS and DVD (third just on DVD) - but I really cannot waste my time on such a setup during this time of year....

[MindlessOath]

after upgrading from a 1.4Ghz athlon to a 2Ghz athlon it has improved my times greatly. it took about 8 hours with 1.4ghz... and with a 2Ghz it only takes about 3,4 hours (6 at max - im guessing now)i have been using autoGK because it is really simple to use. that may be why its going so fast too :P

i also was trying to look for a process to encode all my video's from mpg to two pass divx avi. i need an automated process to do so. i have tried using virtualdub, but it for one takes forever and two... a lot of the mpg's have errors in them, and even the mpg fixing applications dont work to fix the problems. and i tried one program to do the automated process, but it has crappy error correction, so if an error occoured it wouldnt go to the next one etc... vodimey or something i also tried.. same thing. just crap. mabe GK can make something of the sort to convert vids with great error protection. :P