Deinterlacing VHS or not that is the question...
 

Hello guys,

First thanks to admin and staff for this site.

Is that commun workflow step to deinterlace VHS video after capture....I can be missing something...I am far from being a expert.

I would like to have your tought on this topic.

I need to create archives from VHS-PAL. I read many topics that says that you don't need to deinterlace if you aim is not to create streaming. In fact you will lose some information during the proccess. I guess that I can deinterlace in Adobe premiere pro (the one I have) after the capture.

But then I read this topic, that you can deinterlace, without losing anything, by capturing at higher resolution.

http://www.divx-digest.com/articles/vhs_capture.html

I will appreicate you tough of the subject because my experience says that there is often worlds beetween theory and pratice, and me I don't have much pratice...

Regards

Deinterlacing is always a destructive process. So it's best avoided.

However, some workflows require deinterlacing. For example, somebody that converts VHS tapes for Youtube viewing. That process requires the interlaced VHS video be converted into a digital progressive stream. Quite a few restoration workflows carry the same requirement, due to the way the filters function.


As far as the divx-digest guide is concerned...

Very outdated information. Anything written before 2004 is generally worth ignoring. Things changed around that time, between video capture cards, DVD recorders, and quite a bit of software available. Earlier methods were slipshod at best.

Not to seem condescending to the person who wrote it, but this is quite obvious. If I were to write an article titled "How to Butcher VHS Video", then my article would share a lot of the same content as this guide does!

This is accurate information, regarding interlace. At the time, many people were insisting VCD's 352x240 resolution was "VHS quality" because they were confused about "x240" (digital pixels) versus "240 lines" (VHS resolution). While both numbers were 240, the represented different measurements. The x240 is vertical, while the 240 lines is horizontal (as well as expressed in non-digital terms). 240x480 would have been more accurate than 352x240, as numbers to describe VHS resolution.

This is meaningless jabberwocky. I makes almost no sense. And it's not accurate when it does.

^ This was accurate.

This is absolutely horrible advice. In fact, in 2001, it was far worse than it is now. At least VirtualDub and Premiere both contain semi-decent deinterlacing options in 2011. In 2001, everything was drop-frame or blended, making it look rotten. The irony here is this advice effectively makes the initial "don't capture 352x240" advice null. Whether captured at 352x240, or later deinterlaced in an editor, then result would have looked the same.

^ This was a terrible video capture card.

^ This was inaccurate. You don't improve quality by discarding picture data.

If you want to archive, you capture in Huffyuv AVI (lossless compression) and store the AVI as is. Then from there you can go to be making DVD, DiVx, deinterlace or any other format while always keeping the original interlaced file.

Thanks for your inputs...I had the feeling that deinterlacing was "adding" more trouble than anything else...but I am thankfull that someone with knowledge (like you) put things in the right context with facts.

Hello metaleonid,

your affirmation raise a good question to me. Does it mean that winDV or extase DV capture free softwares capture AVI in a non lossless compression format?

I ask because to be honest I never ask myselft if these programs were doing acqusition in a lossless compression way...

Thanks

DV is a lossy 5:1 data compression.

With NTSC video, it also has a fairly lossy 4:1:1 colorspace compression (worse than DVD).
With PAL, the color is 4:2:0, the same as DVD.

DV25 (mini-DV, consumer DV) is not lossless.
It is inside of an AVI container, but remember that AVI is not a "video format".

If you search the site for "DV", you'll come across a number of articles and forum posts on the topic.

With YouTube, the quick method I use is VirtualDub's built in yadif deinterlacer. For more advanced workflows like PAL<>NTSC conversions or temporal filtering, I have been experimenting with the QTGMC AviSynth script with some pretty impressive results. The downside is getting it setup and the fact that its REALLY SLOW in processing video.

I read a bit about colorimetric theory...but to be honest...right now the only thing I know is that I will capture my PAL-VHS via ACDV-300, so far acquisition worked well with Premiere pro and Extase. ACDV-300 seem to be a expensive capture device that I will probably discard if I was consult before initial purchase. At least I succed to get a JVC s-vhs HR7600EK in wich I put more hope vs "speudo TBC" from ACDV-300.

Thanks for the infos

I don't know anything about estase DV, but AFAIK WinDV captures in DV format which is 5:1 compression and 4:1:1 color space.

Basically I have 2 capture devices: the Philips SAA713xxx based TV Tuner card and Canopus ADVC-300.

I use the 1st one to capture in Huffyuv lossless AVI. I use it mainly to capture from the high quality VHS cassettes that are of high importance to me. The process of capturing and syncing audio/video takes a long time. Fortunately I don't have many of those to deal with.

I use the 2nd one (Canopus) to capture in DV (because it doesn't give any other option) with WinDV. I use it mainly with not so great quality VHS cassettes and that are not high importance to me. I use 3d noise reduction during capturing. The capture quality doesn't look worse than the original. The 4:1:1 color space doesn't make worse what wasn't great and sharp to begin with. And plus, the end result motion has far less noise than the original VHS which is satisfactory to me. This process is very quick.

What I do is not necessarily what you might want to do but perhaps you might want to consider those as options. But one thing. If you want to archive, do not deinterlace.

I use AVISynth script for PAL->NTSC conversion. I took it from someone on internet and then modified it myself. I use slight variations depending if the original footage progressive or interlaced. I use TDeint function. One thing is that it gives various results depending on the original footage even if footage is interlaced. I once converted the interlaced PAL VHS to NTSC DVD. The footage was wedding. The end result DVD was outstanding. Motion were smooth as if the original footage was shot by an NTSC camera. I then converted a music concert from interlaced PAL VHS to NTSC DVD. Used the same AVISynth script. The result were horrible. All the time here and there I would get repeated and dropped frames during PAL->NTSC conversion. I can post the script I use here by the way.

--Leonid

That's ok. It has preview function. Play with settings a bit. Try Strong 3d noise reduction, but either weak or NO 2d noise reduction. Play with 3d noise reduction. Now remember, that the end result is going to be compressed. Thus post processing is not recommended unless your final destination is uncompressed AVI which is still going to be limited to DV compression. That's why I recommend all the cleaning filters be done during capture process. If you have JVC S-VHS deck, make sure it's in Edit mode. Don't put it in auto or soft or sharp modes. You want noise reduction and damage to the original footage to be done not more than once.
In Autogain filter (2nd video tab), try to select fixed rather than auto and reduce Y a bit to make sure you don't have White Peak spread all over your picture when it has bright moments.

And by the way. If you use ADVC-300, use WinDV. Don't use anything else.

And by the way, I personally think ADVC is a good device. Maybe overpriced, but it's good at what it does. And I haven't had any dropped frames with it. If it had an option to capture uncompressed or with lossless compression with 4:2:2 colorspace, then it would probably be the near to perfect device.

Hi metaleonid,

I will try to follow you :) So when you use Huffyuv lossless AVI you got file bigger (5X) vs WinDV?

Can you give me the amount of space for 1 hour of video, just to give me a idea.

If I want to capture my VHS-PAL the best way (by the way I have some good quality tape (all original TAPE recorded in SP mode and played very few time), I have to buy a PAL card capturing device like this one for example

http://www.ebay.com/itm/ATI-ALL-IN-W...item2eb9532d34

I read on this forum that ATI all-in-wonder were appreciated. You can see a significative difference at the end product (DVD ripped)...

Thanks for your patience

Hi again,

ACDV-300
Okay if I resume I should try stong 3d noise, but no weak or no 2d noise reduction.

JVC S-VHS deck
use Edit mode (avoid auto; soft; sharp mode)
Fixed autogain filter (instead of auto)
Reduce Y a little.

Thanks

Hi,

I'm not sure you can follow me using the same hardware I do. The LifeView FlyVideo 300FM which I use is long out of print. And I don't think the newer cards work with this http://auzol.narod.ru/index_e.html software which is the only software that allows to use native Philips reference drivers. Also it might be pain in the ass to install it. I'm not discouraging you, but just if you get it and it doesn't work, don't blame me. :)

I will give you precise figure of space.

The amount of space of 54 minute and 49 seconds of 720x576 PAL DV video with uncompressed LPCM 48kHz/16bit audio is 12,182,06Kb

The amount of space of 54 minute and 48 seconds of 704x576 PAL Huffyuv video with uncompressed LPCM 48kHz/16bit audio is 24,229,884Kb

Personally, I do not like ATI even though I haven't owned one. Why? It's already been discussed.
Here's the link: http://forum.videohelp.com/threads/1...hilips-SAA7130

The guy compares 3 cards: Brooktree AverTV, ATI and Philips based AverMedia DVD EZMaker. Well AverTV is not an option. It sucks. ATI doesn't produce as much details as AverMedia. So then even if it can capture uncompressed/losslessly compressed, what's the point of of capturing if it can't produce as much details as Philips based card.

Also download bitmap files post #3 which also shows that Philips based card gave superior image.

However, maybe the guy had different sharpness settings. I don't know.

Now look at this thread here: http://www.digitalfaq.com/forum/vide...7314-chip.html

I made comparisons between ADVC and FlyVideo. Note that FlyVideo Y/C NTSC separation simply suck. But if you compare S-Video, you will see that both cards are pretty close to each other.

So do some more research. In fact, I am still after yet another capture device. I need to capture from the LaserDiscs and for that I need good Y/C separation for NTSC. As I just mentioned my tuner card is horrible at it.

If you resume, you should try and see strong 3d noise. See if you're happy with it.
As for 2d. Set it either to Weak or to OFF. Do not use any values beyond weak. It will simply blur the image.

JVC - yes, edit mode. Don't forget S-Video out. Don't use composite.

As for Canopus Video 2 tab that depends on your image. If you have a lot of white color, try to set it to FIX and reduce Y. If you start seeing colors other than white, keep reducing Y. It's very tricky thing. You have to use preview option (output to TV) and judge by your eyes.

And by the way, any tuner card might give you audio/video sync problem. No matter what experts say, the drift might occur due to the fact that audio card is a separate entity from the video capture card. Canopus seems to be good at it although I haven't checked thoroughly. But there's technique to fix sync problem.

And since I just finished capturing one PAL VHS tape, I thought I would share the images. This was not simultaneous capture. I screwed up settings on Philips, so I had to recapture for just tuner card.

Anyway, Philips Huffyuv compressed. Reduced white peak gain. No noise reduction. Used S-Video.

Canopus - 2d filter Weak, 3d filter strong. Reduced Y. The rest unchecked. Used composite.

Hi again,

Your pictures talk by itself. Also, as you wisely said, each case is unique, but you knowledge is very valuable for me. In my case I will work with pretty much the same kind of scene. I got a contract to archive PAL-VHS tape of bulgarian philosopher Omraam Mikhael Aivanhov. So if you look on youtube it's always same kind of scene, white clothes caracter with sometime white background or blue background. So I think it's a good new because I will work with very static scene.

Even if they already invested in the ACDV-300, I will do some search for a PAL capturing card that can match Huffyuv, showing them differences between your pictures may convice them to invest in a "philips like" card.
I don't know if it's a stupid question but can I look for good NTSC/PAL card or I'm better to look for card that do just PAL or just NTSC?

I don't know if I read that on this thread or forum (I read so much now that I don't remember where I read it), may look "off-topic", apparently buying I s-VHS jvc deck without remote is a very bad investment? I assume it's because the board deck by itself is not sufficient to take advantage of all features?

In any case merry christmas to all of you!

Peace

Leonid,
Funny and encouraging for me to read what you said about Lordsmurf "But still it is great to hear expert's input."
Yeah there is expert and expert...Einstein call it relativity...lol

I think Lorssmurf is the a kind of "papa smurf":D

http://www.digitalfaq.com/forum/vide...ernatives.html

This can be solved by using a DirectShow capture device with built in audio capture. I have done 3+ hour long captures with zero audio sync problems with them. Nobody ever said Brooktree BT8x8 cards were decent. They were popular because they were cheap, widely available, and had a highly tweakable driver available. The composite input on it was total junk, it used a cheapo notch comb filter!

Composite comb filtering even on modern capture cards is an afterthought. Most of them have 2D line comb filters at best. At least one can rely on an external device to do the job properly if needed. If one wants to see what modern capture cards can do, check out this much more recent comparison thread.

http://forum.videohelp.com/threads/3...on-screenshots

I've never made a claim that I am an expert. Especially when it comes to theory. I have had a lot of experience using various capture devices which some of the experts here haven't dealt with. But it's the same as being the actual expert. I'm still learning.

If you use the built-in audio chip rather than external sound card, then it's less likely you will get sync problem. I would still suggest to re-capture the very last minute of your footage, spot the frame in the original footage and recaptured footage and see how sound has drifted away from that frame. It can be done with VirtualDub.

I used to have Hauppage Win TV or something. The quality was horrible. Not just Composite but S-Video. I compared the Philips capture to BT and the difference can be seen very clearly.


I can't see many of the images on this thread. Most of them don't exist anymore. But I am attaching 5 images. There's a capture card that I purchased in 2004. It was Compro VideoMate Ultra. I think they were one of the first Tuner cards (if not the first) that used 3d comb filter for composite. I don't know if ATI were earlier then them. And they were definitely the 1st tuner card that used 3d comb filter with Philips SAA713x chipset. They used to manufacture the cards with Brooktree chip, then they revolutionary combined 3d comb filter and Philips chipset and shortly they incorporated Conexant 10bit chipset with 3d comb filter.

The card uses NEC D64083 for Y/C separation (same as used in ADVC-300). After Y/C get separated, they are fed into Philips SAA713x as if S-Video is plugged in. Because of that I can't use Philips native drivers and thus I can't control White Peak. I have to decrease Brightness to reduce white peak.

Images were obtained when I plugged my DVD player Compro VideoMate Ultra using composite. Judge for yourself. On Philips settings: Brightness 40%, Sharpness + 5.

Me too :)

Now I realize that what seem to be hard to find is a respectable compatible card with windows 7 that can capture lossless AVI...it's kind of ironic.

It's actually pretty easy to set up. (In fact, it's about to get much easier, because of new Avisynth guides coming to digitalFAQ.com in January.) QTGMC operates at about 10-15fps on an optimized quad-core system, using the multi-threaded version of Avisynth 2.6. It ran at 4fps on a dual core with several NR filters in Avisynth, as well as some processing in VirtualDub 1.9.x.

Extase? I think something's been lost in translation.

Premiere is an excellent non-linear editor, but I would suggest never capturing with it. Frame drops are common. Premiere is simply too much of a resource hog to leave enough system resources open for the capturing process. CPU, RAM and hard drive speeds need to be as unfettered as possible when capturing video. Premiere gobbles up most available RAM and CPU, and then it parks a huge temp file footprint on the designated hard drive. It's simply not the best tool for the task.

The ADVC-300's so-called "TBC" doesn't do anything to improve the image quality. The NR filters inside the ADVC-300 are notorious for over-processing video, leaving you with smeary, ghosted, posterized quality. It's similar to what happens with certain Panasonic DVD recorders. The JVC S-VHS VCRs have none of these issues. At very worst, sometimes the JVC Norm/Auto picture settings will default to a soft image, but that varies from tape to tape. There's no ghosting or posterization going on.

I highly suggest the ATI 600 and ATI 650 cards. The comparison of video cards against test patterns doesn't really tell a complete story, when it comes to capturing video. Most videos are screwed up anyway, so you'll be adjusting the color prior to ingest (capture) via hardware, or during post-capture production in a video editor or restoration program. If the video card is off by a slight amount -- and virtually everything is, which is why videos vary so much -- then simply compensate during pre- or post-capture workflows. And if you're really paranoid, there are ways to adjust the ATI gain, via registry tweaks.

Furthermore, luma gain, chroma gain and IRE are often played off one another. I recall two specific eras of hardware -- the 2nd generation Panasonic DVD recorders (2003-2004) and 3rd generation LiteOn clone DVD recorders (2006) -- that reduced luminance gain in order to compensate for excessive IRE values. While reducing luma does reduce overall brightness, it can shift the image tint if done in too elementary of a method. Because luma (Y) carries green color data in YUV, these so-called "fixed" DVD recorders had green tint shifts in the shadows and blacks. I've seen other equipment that shifts red or blue. My own HDTV has a nasty blue shift (completely caused by a lamp upgrade), which is thankfully fixable due to advanced RGB controls (reduce blue gain and intensity).

Nothing is perfect in the realm of video conversion, and the search for perfect hardware is only done in vain. Trust me, I've tried!

That's not entirely accurate. The "built-in chip" is really not any different from the chips on a sound card. In fact, chips on dedicated sound cards tend to be far superior than the cheapy chips used on consumer-grade capture boards or motherboards. It really comes back to the quality of chipset, as well as the quality of the computer's internalized communications. The real causes of skew are when the audio card has too many milliseconds of skew, or if the capture software handles a/v input poorly. In addition to dropped frames, of course.

Some of the WinTV cards have really bad gain. Just horrible. The PVR-150 is the worst offender, both with audio and video. It's not at all a suggested card. The PVR-250 and PVR-350 have quite decent image quality, excluding Half D1 being blurry (not just soft, but literally blurry).

I've never liked that comparison because it's missing context.

The BT8x8 (AVerTV card) is indeed quite lousy. There are a number of workaround drivers that you can play with, and get quite decent quality, but it's most definitely inferior to Philips, ATI, Hauppauge and some others. It's just not in the same league. It's comparing a cheap $25 card to cards that cost well over $100 apiece. Apples to oranges comparison.

The issue I have with the Philips (AVerMedia EZMaker) is that it has halos and echos, which points to artificial sharpening. In fact, I'd bet it happens early in the processing, before the image is down-sampled internally. I don't recall that chipset's native ingest off-hand. A number of the "AVI only" type cards are made with 768x576 palettes, as opposed to the near-native NTSC/PAL resolutions of ATI. The ATI cards do not process (butcher) the video further, in terms of "adding detail" or any such nonsense, which is why they've often been referred to as "soft" when compared against cheaper cards.

The biggest omission in these comparisons is that squabbling over "detail" is a wasted conversation when it concerns the capture of videotapes. None of the videos have the kind of detail information that you'll find in a test pattern. Even S-VHS, Super Beta, BetacamSP and Hi8 will have trouble pushing out pixel-accurate details to fill out a 720x480 or 720x576 Full D1 image palette. The important goal is to capture all of the image detail available, while equally NOT including random high frequency noise (which some people mistake for "detail"). The Philips and ATI cards can both do this quite adequately.

While comparing against test patterns makes for interesting discussion amongst software developers, video engineers, hardware manufacturers, professional video restorers, etc -- it's mostly meaningless information to casual or even advanced users. It's not applicable knowledge of any kind, as it won't really affect what they can or cannot do with videotape conversion.

DV is about 13GB/hour
Lossless AVI (Lagarith or Huffyuv) is about 30-40GB/hour.
Uncompressed YUY2/YV12 is about 75GB/hour.

Lossless is about 2:1 (or 75:35 rounded to 75:37)
DV is about 5:1 (or 75:13 rounded to 75:15)

MPEG-2 can be whatever you make it. The benefit of MPEG-2 over DV is less colorspace compression. You can match DV 5:1 compression by using 25Mbps MPEG-2 I-frame only, or compress temporally and get it down into the 15Mbps range with similar quality.

The problem with this approach is that EDIT doesn't always provide the cleanest image quality, which is the main reason to own a JVC S-VHS VCR. In fact, you generally should keep it in NORM (never AUTO, as that means Calibration is on, which is bad). Never SHARP, rarely SOFT. Chroma noise will especially disappear when the TBC is engaged, and NR is set to NORM. And the removal of chroma noise is very desirable to do in hardware, as software is largely unable to process this noise out effectively. The best chroma noise reduction filter via software is available with Avisynth, but even it is weak compared to good hardware processing. And all VHS tapes suffer from some degree of chroma noise.

Whether it's the ADVC line of DV converters, or even just a DV camera, always use WinDV in Windows XP. I've not yet tested it in Vista or WIn7 (that I can remember, at least). Scenalyzer is another option.

Downloads available here:
- http://www.digitalfaq.com/forum/vide...35-sclive.html
- http://www.digitalfaq.com/forum/vide...-software.html

Premiere is a video editor with capturing put in later as an add-on function. As such, the software is tuned to editing, not capturing. You generally want a tool built for the purpose of capturing. (Or in the case of a tool like VirtualDub, something that was built with the dual goal of capturing and processing.) You never want to use an editor or authoring software -- especially not some "all in one" solution -- as capturing is a poorly implemented add-on function in almost all cases.

It's a competent DV input device, absolutely. Overpriced, definitely. Overhyped, assuredly. These were always consumer devices, as compared to the Canopus DV Storm and similar discontinued products of the same era. Remember that Canopus DV boxes were created in the era of Pentium III computers (late 1990s and early 2000s), and are not even remotely current technology. The only reason for their long lifespan is that (1) it's using DV standards, and (2) it's fully externalized as a hardware appliance.

That's definitely an older method at this point in time. A number of Avisynth methods are now deprecated (obsolete), including TDeint and some of the older framerate change methods. It's also not just the motion you have to worry about, but the quality of the resolution resize and the deinterlacing method.

Generally speaking, it's often better to just make a PAL DVD from a PAL VHS tape, and allow modern hardware to adjust the size and framerate. Software methods of conversion are permanent, while hardware improves with each successive generation. The WDTV boxes and Samsung Blu-ray players handle progressive PAL source beautifully. So a simple deinterlace is just as effective now as a framerate adjustment combined with an audio pitch edit.


... that's the long reply I've written in months. :)

Hi guys, I tried to do a chart in my message, but was not convincing.

Here is resume of advantage and disavantages of 2 paths (DV-AVI vs AVI lossless).
I did I chart in power point that I want to present to "my client" so they will decide wich way to go.

Deck; JVC-HR-S7600EK (right now I have no full frame TBC)

Aim: 80% VHS-PAL capture and 20% VHS-NTSC

Path 1
DV-AVI Canopus ACDV-300
Advantages:
Take less space
Don't need to go back to windows XP.
"Easy" audio/video synchro
User friendly.

Disavantages
Recuded or no ? post-processing possibility. (ex: desinterlacing for steaming; noise reduction filter)
Have to compress again to burn DVD.

Path 2
Huffy-AVI ATI All wonder x600 pro PCIE
Advantages:
Lossless compression via huffy
Post-processing possibilty (less destructive if there is plan for desinterlacing
More stable format for archives
1 step compression to burn DVD
More marging toward technologic evolution.*

*Right now path 1 seem to suffice for their need in term of IQ result. DVD signal is carried by standard RCA connector to a Hitachi projector.

Disavantages
Take more space.
Audio/Video synchro may be picky
Many tuning to do
Go back to windows XP

If you find innacuracies or falses assumptions or things to add in my analysis feel free to correct me. Thanks

Overconcern with audio sync:
Loss of audio/video sync is not really a common issue, and hasn't been for many years now. Most sync errors were caused by cheap hardware from a couple of generations ago: VIA motherboards, $10 no-name sound cards, BT8x8 video chipsets. That sort of rotten hardware just doesn't exist anymore.

"1 step compression" doesn't make much sense:
Both DV AV and Huffyuv AVI must be encoded to MPEG-2 to make a DVD (or AVCHD/H.264 or MPEG-2 for Blu-ray).

Both are user-friendly, but...
The ADVC is dummy-friendly to install and use.
The ATI AIW card can be difficult to install, but it's very user-friendly to use for capturing. (Use VirtualDub for AVI.) This site has plenty of documentation regarding proper installation, setup and use of ATI AIW cards.

Both methods are sound. I think your pro/con list is well-done. :)

Let the workflow decide:
Huffyuv is superior, but DV may be "good enough" depending on the overall workflow requirements. DV would be a lousy choice for restoration work, for example, as the colorspace compression screws up color and detail problems even more. For straight archiving of already-excellent sources, it will suffice. Certain formats look great as DV: Video8, Hi8, etc. VHS and S-VHS, however, tend to look better as non-DV.

I do have VIA motherboard indeed. But my video chipset is Philips and my sound cards are M-Audio Audiophilie 2496 and Terratec Aureon Universe 7.1. The 1st one is pro, the 2nd one is semipro. I still get gradual sync problem. The other computer at my parents house has built in soundcard on Intel based chipset motherboard and another Philips based video chipset. At my place I have different sync problems using both sound cards and both Samsung and JVC VCRs. At my parents I have sync problem using Hi8 camera.

The only time I don't have sync problem is when I set JVC VCR to Auto mode instead of Edit, and plug audio to Tettatec sound card. Then it is indeed perfect. But every other combination gives gradual sync problem: either audio is ahead of behind.

I am convinced that it has to do with a particular VCR and soundcard's clock. The gradual pattern of sync is specific for combination of a VCR and a SoundCard.

You mean I should capture excellent sources with DV or otherwise?

--Leonid

That you can do so --- not necessarily the same as "should". ;)

I don't know. I used to think certain things were attributable to specific VCRs or VCR model lines, but long-term experiences and research have shown those early conclusion to be based on false logic on my part. I'm afraid that may be the case here. I just don't see how the VCR's audio output could in any way affect the clock on the audio card. It's more likely just a general breakdown of hardware cooperation, caused by the cheap parts and components used by certain eras of hardware (VIA, again, being a major contributor to crappy performance experiences).

The most important aspect is that you've found a workable solution. :thumb:

Interesting. I always thought that for the bad quality source the inferior codec would not make any difference. I.e. for the blurry source it wouldn't matter whether you convert with Huffyuv or DV. But for the high quality source it is absolutely necessary to preserve all details and thus use superior codec which is lossless. Can you elaborate why you think otherwise?

No, no. The VCR doesn't affect the clock of an audio card.


Agreed in some way. The combination of hardware pieces makes the unique difference between audio and video.

First off, when I am talking about the delay it's up to (less or equal to) 80 milliseconds per hour. I.e. after one hour of capturing, my audio would be 80 milliseconds behind my video. That's what I experience using Terratec and M-Audio. It's not a big deal, but if you know it, you may notice it. If the delay over 100 milliseconds, it becomes noticeable.

As to why.... Do you agree that 2 different high end sound cards have 2 different internal clock? That means that if you digitize one hour of analog audio, the length of the wav files will end up to be different up to several milliseconds.

Now 2 VCRs don't run at precisely same speed as 29.97 fps. There are slight variations. One might run at 29.9669 and the other might run at 29.9701. This fact alone contributes to the difference in the length of an audio.

Can you recommend the USB ones with 3d Comb for NTSC composite?

I agree. But if the sound quality is not as important, then internal chip will do a job. For example when you don't work with music but digitize sports or Hi8 footage where the sound is bad to begin with or not as important. You just don't need to go extra step to sync audio and video.

Actually the gain (white peak) was ok. It's just the image was blurry.

I'll post the AVISynth script in a separate thread.

I always make PAL DVDs from PAL VHS tapes. However, some of people I do DVDs for are unable to watch PAL interlaced footage in good quality. Thus in addition I create NTSC DVD. But if you can recommend the software player that does the job of conversion and perhaps doubles frame rate, I would definitely appreciate it. Thank you.

--Leonid

I almost never feed capture cards composite signals anymore. Even if the VCR is composite out, another device in the loop will switch it over to s-video. That's usually a second JVC S-VHS VCR, as needed. I have a dozen of them, after all.

Definitely post that Avisynth script in another thread. Maybe we can help you tweak it. :)

For the PAL, what I'm getting at is that you just need to convert PAL to progressive. PAL interlaced and progressive are both 25fps, so it's an easy task, using a good Avisynth deinterlacer like QTGMC. Even YadifMod+NNEDI2 can yield decent results for non-animation, if you're in a rush or on a low-power CPU computer. I don't see a need to double frame rates. I don't even own a DVD player anymore, which doesn't come with USB capable of supporting MPEG-4 (Divx/Xvid, sometimes H.264 as MP4/MKV), supporting PAL and NTSC. All of them play region-free (homemade) PAL DVDs.

Hi guys,

If I follow you DV capture come with 4:1:1 color space, meaning that if I capture good PAL-VHS tape (4:2:0) I am loosing some color "potential" quality.

Happy new year!

- NTSC DV= 4:1:1 = colorspace compression lower than all other formats (4:2:2 for most analog, 4:2:0 for most discs/DVB)
- PAL DV = 4:2:0 = colorspace similar to DVD/Blu-ray discs 4:2:0 (but not 1:1 identical, since co-siting method is different)

For PAL, DV is not as big of an issue.
It's still best to leave it as 4:2:2, however, until that very last processing encode (for distribution online, on disc, broadcasting, etc).

- 4:4:4 exceeds most sources (all but film, 2K/4K digital)
- 4:2:2 is ideal
- 4:2:0 is acceptable
- 4:1:1 is noticeably lossy
- 3:1:1 is excellent, used in some Sony proprietary formats (similar to 4:2:2 ratio)

An analog tape (PAL, NTSC or SECAM) is closest to 4:2:2 when measured with that digital measurement.

One day down, 364 more good ones to go! :D

Can we says that PAL-DV vs PAL-huffy, is the same as comparing picture vs photo strip?

I tried to find "Image" to vulgarize for my "old" client. :)

I don't think I quite understand the question. I think you're getting at the old photographer cliche: picture vs photograph.
Is that correct?

If so, then I don't think it's really quite that bad. But it's probably close enough to serve for the purpose of comparison for a layman (the client).

Sorry as you probably know since long, english is not my first language.

I mean Film negatives vs photos or in a mordern way....RAW vs JPG :)

I think I know what you mean since English is not my 1st language either :) . I think Huffyuv vs DV is like PSD (photoshop files) vs jpeg. Or BMP vs Jpeg. But BMP is totally uncompressed whereas Huffyuv compressed, but compressed losslessly.

Funny that it's called lossless, lossless vs what? I mean if it's lossless what is the point of using YUV 4:2:2 uncompressed? :confused:

You don't understand. From losslessly compressed AVI you can restore the full information. I.e. you can recreate the uncompressed AVI without loss of any information. From DV you can create uncompressed AVI but with loss of the information that was thrown out when the file was compressed to DV.

Thanks, so if you want to edit a huffyvideo in premiere, first you have to (or better to) uncompress it?

Quite interesting...(I know better digital photography and not much video)...I guess that huffy is just a way to save space then...

Regards

I don't work with Premier. But I think working with Huffyuv avi directly there should be fine.