Some assorted VHS to PC to DVD capturing questions

[Rudyard]

Well I finally got the new PC and started to do some capturing last night from VHS to PC. However I have been staring at different test captures and trying to compare and would just like some advice before I put in the hours (10 x 4 hour tapes)

Some info on equipment: An old JVC mono VCR to transfer OLD (25 years old some tapes) VHS tapes to brand new 3ghz Dual core, 2 gb ram, 2 160gb hdd (capture to 2nd drive) with a leadtek 7800 VIVO nvidia 256mb card. The vcr is connected through composite connections. I am caturing with Winfast PVR (came with card) and ULEAD video studio 9. I have VEGAS but it doesnt recognise the capture drivers or something so I cant capture with that. The end result will be Dual Layer DVD.

1 - CODEC?. I have tried the uncompreesed AVI, which obiouvsly equals enormous file size. I dont mind captruing an hour and then compressing down to MPEG2 DVD compliant file as that will be the end result. I have tried the Panasonic DV codec but should you capture analogue with software compression (using a codec is software compression isnt it?) I dont knwo if i was going crazy but I am sure the Panasonic DV codec on this particular old tape was producing a 'flicker' more than the original and uncompressed AVI.

Also the card says I can capture straight to MPEG2 if I want but I dont know if that is hardware compression or software compression, as surely you would get a better output if you dedicated all resources to a separte encode after it was doing it on the fly wouldnt you?

For VHS I am a little unsure about what setting I should actually capture on. Which format is best to capture to?

2 - Interlace or deinterlace? - When capturing to AVI I have noticed some 'black lines' in the video when things move. Like every 2nd is black (I dont know exactly how to describe) I think it is to do with interlacing as the display order is every odd line isnt it?

However when I render this AVI to an MPEG it goes away and is fine, wehter i render it as frame based (progressive??) or with a field order (interlaced??) it seems to play fine when converted to MPEG2.

So this is fine as long as I know that it will go away when I colmpress, I would hate to have a lot of footage and then have to do it all again.

Should you deinterlace when you capture or after or just leave it interlaced?

3 - Length of captures. I did about an hour in one big go last night and was wondering is there any unwritten law about the length of capture before you may run into problems? I am planning to do about an hour, then render to MPEG2, an hour then render etc etc. Is there something wrong with that?

4 - Capture resolution. Should I capture to 720x576 DVD resolution or capture at VHS (whatever it is 300m200 odd?) resolution and then when I compress to MPEG2 convert it to 720x576 then? I tried 720x576 last night but dont know if thats the propoer way to do it.

Thankyou for any help you can provide and any other tips you can throw at me will be appreciated.

[Cole]

I can only answer one question there, Rudyard, and that is; from what I know, if you are going to make DVDs from your VHS captures, then capture at DVD resolution.

VHS resolution is similar to a VCD's res. so to capture at a smaller picture then convert it to a higher one will decrease the quality of the picture.

[FulciLives]

1 - CODEC?. I have tried the uncompreesed AVI, which obiouvsly equals enormous file size. I dont mind captruing an hour and then compressing down to MPEG2 DVD compliant file as that will be the end result. I have tried the Panasonic DV codec but should you capture analogue with software compression (using a codec is software compression isnt it?) I dont knwo if i was going crazy but I am sure the Panasonic DV codec on this particular old tape was producing a 'flicker' more than the original and uncompressed AVI.

If the size of an uncompressed AVI is too large then try the HuffyUV codec. It is free and uses very little compression that is considered to be a "lossless" form of compression. File size will be about half if not maybe a bit less than uncompressed AVI. You should not use a DV codec because the computer has to do it with software ... you don't have a hardware DV capture device ... so using a DV codec to capture is a no no in this situation.

Also the card says I can capture straight to MPEG2 if I want but I dont know if that is hardware compression or software compression, as surely you would get a better output if you dedicated all resources to a separte encode after it was doing it on the fly wouldnt you?

The capture card you have, as I understand it, does no hardware MPEG encoding on capture so any MPEG capture would be done with software. This is very taxing on a computer when trying to do MPEG-2 DVD spec captures BUT your computer is FAST enough that it should be able to do this type of capture but bear in mind that your card will give better quality if you capture uncompressed AVI or HuffyUV AVI then convert to MPEG-2 DVD spec using a non-realtime software MPEG encoder such as CCE or TMPGEnc Plus etc.

2 - Interlace or deinterlace? - When capturing to AVI I have noticed some 'black lines' in the video when things move. Like every 2nd is black (I dont know exactly how to describe) I think it is to do with interlacing as the display order is every odd line isnt it?

However when I render this AVI to an MPEG it goes away and is fine, wehter i render it as frame based (progressive??) or with a field order (interlaced??) it seems to play fine when converted to MPEG2.

So this is fine as long as I know that it will go away when I colmpress, I would hate to have a lot of footage and then have to do it all again.

Should you deinterlace when you capture or after or just leave it interlaced?

You should almost never de-interlace video footage. There are very rare instances when you might but no you do not want to do this. Interlaced video looks very strange on a computer because a computer monitor is progressive but the DVD format can handle both interlaced and progressive video so it is not a problem AT ALL to leave your video interlaced. Generally speaking you do damage if you de-interlace. The one difference is IVTC which is a process that takes telecined 29.970fps NTSC and converts it back to progressive 23.976fps NTSC. The IVTC process only works with a film source NTSC and even then it can be tricky to do and not always the best thing to do anyways.

3 - Length of captures. I did about an hour in one big go last night and was wondering is there any unwritten law about the length of capture before you may run into problems? I am planning to do about an hour, then render to MPEG2, an hour then render etc etc. Is there something wrong with that?

As long as you have the NTFS file system you are A-OK. If you have WinXP then chances are you have NTFS unless you updated the computer yourself from an older version of Windows then you might still have FAT32 which has a 4GB limit per file size. NTFS does not have that restriction. Based on what you said so far you sound as if you have NTFS so no problem with length of the capture.

4 - Capture resolution. Should I capture to 720x576 DVD resolution or capture at VHS (whatever it is 300m200 odd?) resolution and then when I compress to MPEG2 convert it to 720x576 then? I tried 720x576 last night but dont know if thats the propoer way to do it.

You should capture at Full D1 resolution which is 720x480 NTSC or 720x576 PAL and depending on your capture card you may have to use 704 instead of 720 to ensure a proper aspect ratio. Usually though you can't tell the difference between say 704x576 and 720x576 as the aspect ratio difference is so very minor. So unless you want to research your card and chipset just capture with a width of 720 instead of 704.

- John "FulciLives" Coleman

[Rudyard]

Awsome, thanks so much that is the kind of advice I was desperately in need of. You just saved me many hours of possible wrong doing!

Thanks again for both responses.

[Rudyard]

Sorry, just rereading your answers, if I use huffy Codec isnt that another software compression just like Panasonic DV codec??

Why is Huffy ok to compress on the fly but not others?

[FulciLives]

Sorry, just rereading your answers, if I use huffy Codec isnt that another software compression just like Panasonic DV codec??

Why is Huffy ok to compress on the fly but not others?

HuffyUV was designed with capture in mind.

Another very good codec to use for capture is the PICVideo MJPEG codec which is another codec that was disigned with capture in mind.

Of the two HuffyUV is considered better in that there is less compression loss (virtually none) but PICVideo MJPEG was around first and for a long time it was the only one that could be used due to the speed of computers but now computers (such as your own) are very fast so HuffyUV works well.

HuffyUV tends to compress by around 2x or 2.5x give or take so at 2x compression that is a savings of 50% over uncompressed AVI. PICVideo MJPEG can produce files about the same size as HuffyUV using the top quality settings but the magic of PICVideo MJPEG is that it has settings that use more compression thus using less HDD space. Of course this was more important in the "old" days when computers were slower and had smaller hard drives.

I should point out that some people still prefer to use PICVideo MJPEG over HuffyUV but PICVideo MJPEG costs something like $28.00 US Dollars whereas HuffyUV is free. HuffyUV is more taxing than PICVideo MJPEG but not by much and both HuffyUV and PICVideo MJPEG are much less demanding than using software only MPEG-2 or DV etc.

- John "FulciLives" Coleman

[thecoalman]

I'll add a little to this...

3 - Length of captures. I did about an hour in one big go last night and was wondering is there any unwritten law about the length of capture before you may run into problems?

The only problem you may run into with long captures is audio sync, if you don't have any don't worry about it.

4 - Capture resolution. Should I capture to 720x576 DVD resolution or capture at VHS (whatever it is 300m200 odd?)....

You'll get varying opinons on this but here's my take on it... Discs are cheap... memories from the past are not. Capture at the highest resolution possible (or whatever the the max of the card is) using the Huffy codec as suggested above. Encode to MPEG at the same resolution using a bitrate no lower than 6000. Overkill?... Maybe but you can rest assurred that you have not compromised the quality in any way. BTW you never want to upscale a video, scaling down is fine but scaling up produces very poor results.

You can capture directly to MPEG using VS but that will probably not produce as good a result as going from AVI>MPEG. Do some short test burns to compare...

One last thing... Get a RW disc to do some testing. Don't go by what you see on the monitor, use a short clip and just drop it on the first play video in VS. Preferably a clip with a lot of motion. As I mentioned above you may run into audio sync with long captures, do one test with long clip before proceeding to creating your masterpiece...

[Rudyard]

Thanks FulciLives, thats much needed info!

thecoalman,

I will keep an eye out for audio synch, ill reread the sticky and see if there are any tips in there that I can use.

Bitrate - I dont fully understand why you would use certain bitrates over others and constant VS variable.
When making photo albums on DVD, I just used 8000 or 9000 at a constant bitrate figuring that would make the best quality.

However I have mucked around a little with variable bitrate and still get good results but can fit more on a disc.

Is 6000 a good min and let the bitrate vary or should you stay constant? And also if using a constant bitrate why would you use 6000 instead of 9000? My thought process was, Ive captured to AVI and thats the best I can get of the original source so now when I compress to MPEG2 I want to have as much of that captured AVI data in there as possible, hence my using 9000 bitrates.

Could you please explain your thoughts on bitrates?

I will do some testing with a RW, but I always thought you had to 'close' a RW disc to make it work on a dvd player and as such it cant be rewritten again? Am I mistaken?

[FulciLives]

Bitrate - I dont fully understand why you would use certain bitrates over others and constant VS variable.
When making photo albums on DVD, I just used 8000 or 9000 at a constant bitrate figuring that would make the best quality.

However I have mucked around a little with variable bitrate and still get good results but can fit more on a disc.

Is 6000 a good min and let the bitrate vary or should you stay constant? And also if using a constant bitrate why would you use 6000 instead of 9000? My thought process was, Ive captured to AVI and thats the best I can get of the original source so now when I compress to MPEG2 I want to have as much of that captured AVI data in there as possible, hence my using 9000 bitrates.

Could you please explain your thoughts on bitrates?

When encoding for DVD using MPEG-2 the bitrate used depends on the running time. The highest video bitrate you would use is generally 8000kbps. That will normally allow approximately 60 minutes (give or take) of video along with audio. That is as "good as it gets". For instance even if you were only putting 30 minutes instead of 60 minutes the DVD format still restricts you to using a video bitrate of about 8000kbps. If you use compressed audio like AC-3 you can "push" that up to maybe 9000kbps but rule-of-thumb says to keep the MAX bitrate at around 8000kbps.

Now if you want to put more than 60 minutes on a DVD (single layer) then you have to lower the bitrate. Once the bitrate gets below the MAX of approximately 8000kbps you then want to do a 2-pass or multi-pass VBR encode. Here you select a MIN, an AVG and a MAX.

For example let us say you want to put 90 minutes on a DVD with 256kbps AC-3 audio. Using a BITRATE CALCULATOR your AVG bitrate would be approximately 6300kbps (this is "conservative" as it leaves a lot of room over for DVD authoring overhead ... at 90 minutes with 256kbps AC-3 audio you could use an AVG as high as 6500kbps but that's "pushing it" if you ask me). You would then set your MAX to about 8000kbps and the MIN to about 2000kbps. The idea here is that the AVG will be 6300kbps but the encoder will increase the bitrate when needed and lower the bitrate when not needed. Hopefully there are enough "low bitrate" scenes so that the encoder can go higher for the "high bitrate" scenes. The less motion the less bitrate needed. Scenes with a lot of motion need more bitrate etc.

Now I should point out that everyone has a different "threshold" as to when to go from CBR to VBR. In my opion if your bitrate is going to be 7000kbps or higher just go with a CBR encode otherwise do a 2-pass or multi-pass VBR as in my example above.

I will do some testing with a RW, but I always thought you had to 'close' a RW disc to make it work on a dvd player and as such it cant be rewritten again? Am I mistaken?

A DVD-RW or a DVD+RW that is FINALIZED cannot be written to again in so far as adding additional data to the disc without destroying what is there but even after you FINALIZE it you can ERASE it and use it over again as if it is a fresh blank disc. Obviously this makes using a DVD-RW or DVD+RW a good choice for "testing" purposes because you are using it to check the final output and if it is wrong you are not wasting a disc.

- John "FulciLives" Coleman

[edDV]

I agree with FulciLives and will add a few comments from experience.

1. Huffyuv lossless 4:2:2 software compression should be considered your default for a "highest quality" capture with that card. Your CPU can handle the load. It will pull data rates down to ~28-35Mb/s (about the same as DV) and will be accepted by Vegas with uncompressed 4:2:2 D1 project settings. Expect Vegas to blow up the size in temp files as you import and filter. If you don't have the HDD capacity, work and encode in segments.

For VHS this may be overkill unless you expect heavy filtering/editing. Your CPU is beefy enough to MPeg2 compress in software using Ulead's realtime Mainconcept encoder. The amount of compression you can achieve will track CPU power. Try both ways and compare.

DV encoding is usually done in hardware (camcorder pass-through or a transcoder). You may have enough CPU power to pull off software DV encoding. Capture files will be slighly smaller than Huffyuv but Vegas temp files will be considerably smaller in DV project mode depending on how you filter

2-4 nothing to add.

[Rudyard]

Thanks a lot I appreciate your help.
Its good to know about the RW - I was always under the impression once it was finalised, that was it. I now can reuse some discs I thought were dead!

[thecoalman]

I now can reuse some discs I thought were dead!

If they a RW's... that doesn't apply to R's. :P That's why RW's cost more. For discs you are making to keep you don't need RW's.