help with huffy and lossless capture - sefy/anybody

[koinhelgo]

HI
my system first:
amd tbird 1300
asus 7700 64ddr deluxe (have tried both bundled driver 6.31 and latest beta version 12.91)
1.5 gigs sdram (with vcache limited to 512 megs)
40 gig hd(partitioned 34/6- this used for capture) 7200 rpm
creative 5.1 soundcard

using virtualdub as primary capture application
and TMPENEnc 12i to convert to mpegs - using standard 1150 bit rate template for pal vcd.

I have been experimenting with various codecs with different results. I want best possible mpeg conversions for full screen playback on pc (and I put to vcd for standalone dvd)
I am mainly doing back ups of vhs tapes - so films generally an hour and a half or so - nicam/decent vcr so usually pic quality is OK).
I get best mpegs using huffy but I can only get about 2 gigs before it dips out with an error - I have enabled greater than 2 gig capture. Huffy also produces far too big files after mpeg conversion roughly double the size of ones captured with divx codec - but these produce rather blocky mpeg conversions though the full divx conversion is good.

Have been using the Panasonic wavelet codec - which only works properly with the asus beta driver. This produces good quality mpeg conversions with tmpgenc although normally up to 8.5 hours for the conversion with noise reduction and block reduction at 35-70
Often the capture fails though with this - the files sizes seem to vary wildly when capturing. I have used the same settings as preferences for all. Most of the time the file gets to around 4 megs and stops with a file error saying incorrect parameter. The compression rate is often very different (in right hand panel display) this is also true for the Huffy codec. I can sometimes get a whole film at just below 4 gigs, but also get a lot of 4 gig capture failures after about 20 mins - this seems to be when copying tapes from long play recordings - is this because of the noise the "lossless" routines have to also encode.

Basically I've found the lossless codecs gives better results for later conversion but how do I get a bigger than 2 to 4 gig file and how do I regulate the compression.

Any help or thoughts re any of the above appreciated - if I haven't given enough details I can not think what else to say......
regards

[hitechjunkie]

How to use Virtualdubs Multisegment Capture (spill system

[koinhelgo]

Yeah thanks fro the pointer to multi-segment
After a couple of reformats (due to dodgy ram and dodgy soundcard seating)and some weeks in between I forgot all about this ....
However still have the problem that multisegment only supports files upto 2050 megs, so I can't capture a whole film and then convert it to divx as well as vcd. also the quality seems to take a hit - I did two long-play films (back to back on tape) and ended up with five segments - Ok I needen't split the mpegs to put on vcd, but when I load back into Virtualdub to make a divx, although all segments are opened, it only encodes the first one. Is it possible to batch encode and/or join in virtual dub like it is with TMPGEnc?

Open to comments/advice anything...

[Just G]

koinhelgo,
You ask alot, but, as I'm verbose, I'll give alot

First, make sure you are capturing segmented AVI's. Using VirtualDub v1.4.7, from the capture screen, pull down Capture, click on Enable Multi-segment capture. Click on Capture again and Capture drives.... Click Add spill drive and click under the heading Path and enter the drive letter of your capture drive. You should have a row that reads something like, Priority +0, Threshold 50Mb, Path D:\. Next, set Try not to create AVI files smaller than to something like 100MB, and set Try not to create AVI files larger than to 1000MB. This will assure that each segment stays below the 2GB AVI file barrier. And click OK.

You should capture at 352x480(576) (NTSC/PAL, respectively) and 29.97(25)fps. You can cap at 352x240(288), but you'll be discarding one whole field per frame. You don't want to do this. You would then add a Deinterlace filter when encoding to DivX or TMPGEnc. I'd suggest Gunnar Thalin's area based deinterlacer home.bip.net/gunnart/video/#deinterlacearea

If you are using your 6GB partition for capture, you won't get much capture time with HuffYUV ~ 20 minutes. I have about 45GB in my spill system and VDub estimates 3hrs using HuffYUV at YUY2 and Predict median (the best settings, BTW.) So, figure 30 minutes ~ 7.5GB, or 1GB ~ 4 minutes.

If you don't have this kind of space, HuffYUV is not the codec for you. Although, it will produce the best VCD MPG's.

DivX, for example, will usually produce smaller AVI's than the resulting VCD MPG's from it. That is, an 80MB DivX might encode to a 120MB VCD MPG. The caveat in using DivX is that even though it has an .avi extension, the file is truly an MPEG-4. MPEG is a lossy compression; DivX is MPEG; therefore, DivX is a lossy compression. You want to provide the VCD MPG encoder, TMPGEnc, in this case, with as much of the original video as possible. This is why HuffYUV is far superior - nothing is lost. If you don't have the capture disk space, though, this is a moot point.

I'm a bit confused in your process, however, after re-reading your second comment. If you cap with VDub and have cap.00.avi, cap.01.avi, ..., cap.05.avi, you would open cap.00.avi in VDub and all additional segments should be automatically appended. (Doesn't work if you drag 'n drop, though.) Once in VDub, you just need to change your video codec to DivX, set your audio bitrate (I assume you're doing MP3 audio, right?) and do a save AVI to encode your DivX, and follow Truman's advice for encoding to TMPGEnc.

Hope this helps.
Just G


<font size=-1>[ This Message was edited by: Just G on 2001-09-20 21:00:00 ]</font>

<font size=-1>[ This Message was edited by: Just G on 2001-09-20 21:07:43 ]</font>

<font size=-1>[ This Message was edited by: Just G on 2001-09-20 21:08:05 ]</font>

[Truman]

Open the first segment in Vdub, add filters if you like and start the frameserver. Open the .vdr in TMPG. Encode.
Vdub now serves all segments, so you will end with one MPEG, no batch encode, no join.
An avi to avi conversion in Vdub should work in the same way, but I'm not sure.

[ironwood321]

Another good (comercial) codec to use is mainconcepts DVcodec.
This software only codec captures in 720x576 or 720x480 and lower vcd resolutions only. You can use the codec on anologue cards its supports RGB24 input but not many YUV formats.

DV codec will consume data (10minutes of data would be captured in a little over 2GB That's about 3.6MB/persec) if 10mins = 2GB 1hr = 12GB. After you have captured you data to your harddisk. You will have to re-encode it using a codec or mpeg encoder of your choice. For mpeg track down the pasasonic premere plugin as this will handle the 4:1:1 to 4:2:2 colour space conversion better than other mpeg compressors. If you video in Pal format the conversion from 4:2:0 to 4:2:2.