VideoHelp Forum
+ Reply to Thread
Results 1 to 27 of 27
Thread
  1. Member buckethead's Avatar
    Join Date: Jul 2008
    Location: United States
    Search Comp PM
    First off, a hale and hearty thank you to all the kind and knowledgeable folks here who take the time to share their views. Recently, I discovered the wonderful world of neatvideo and all it can do. This has inspired me to recapture my vhs collection (shoot me now). While capturing this time, I want to make sure that the black level devil is minimized. Here is my planned chain :

    1) Purchase a Panasonic AG-1980, which is on the way. Should I use tbc on all tapes, or only the ones that look shaky without it? Should dnr be used here, or should I save this for neatvideo ?

    2) Run the output to a Sign video proc amp. I am hoping to set brightness, contrast, chroma and hue here as accurately as possible. From what I've learned on this forum, I believe (duck and cover) that the output from the proc amp should be 7.5 ire for vhs video.

    3) Possibly a Sign video dr-1000 detailer/enhancer. I'm thinking about software sharpening instead, but if this is a must-have, I'll do it.

    4) Capture in XP mode on a Panasonic e80. I know some here have a real hatred for this model, but so far it has served me well. Alternately, I could use my ez27, if that is better.

    5) Import a short sample clip via dvd-ram to tmpgenc 4.0. Here I plan to use the contour filter, which works magic on vhs, and possibly some fine tuning of brightness, contrast, chroma and hue, so long as the black level is not altered.

    6) Import this file to Virtualdub and neatvideo. I'm guessing that my pc monitor should be a large crt for this step, as well as the previous step.

    7) Burn this file to dvd-rw using tmpgenc 3.0. Now, play this dvd on a full sized tv monitor to make a final check. If all is well, possibly burn the complete clip to dvd-r gold disk. (Necessary/best brand?)


    I also have several PAL tapes which I will play through my Panasonic AG-W3. I have no seperate TBC, so I'm hoping my Prospec dv-773 will help.

    Okay, all tips, hints and hellish experiences are welcomed ...
    Quote Quote  
  2. Geez, I capture straight from one of my beloved Mitsubishis to my E80H.

    Follow the link below for all you need to know. Personally, I think you're going through way too much trouble.

    http://www.digitalfaq.com/dvdguides/capture/intro.htm
    Quote Quote  
  3. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    Do you have any hair? You won't when you've finished!

    Seriously, if you want to do this much effort, check what format you digital files are in. You have an MPEG-2 encoding step in your DVD recorder. If TMPGEnc 4.0 is also outputting MPEG-2, that's two generations. You'll have to encode again after neat video to burn onto DVD - that's three generations.

    Two generations of MPEG-2 is one too many (though results can be fine, with care) - three is going too far IMO - what does the Contour filter in TMPGEnc do? Can you use HuffYUV for the intermediate stages here?

    Some people demand lossless all the way from capture to final encode, others are happy with a generation or two of MPEG-2 along the way - I take a path somewhere between these two extremes and use DV until the final encode.

    Cheers,
    David.
    Quote Quote  
  4. Member themaster1's Avatar
    Join Date: Nov 2006
    Location: France
    Search Comp PM
    Capture lossless>vdub+filters+frameserver>tmpgenc

    This way buddy

    vdub filters: msu deblocking (or smart deblocking it depends), neatVideo, gradation curves, collor mill, msu color enhancement, HSV, resize

    You got pretty much all the tools needed here
    Quote Quote  
  5. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    Why would the OP need to deblock a lossless capture of a VHS video?!

    Cheers,
    David.
    Quote Quote  
  6. Member themaster1's Avatar
    Join Date: Nov 2006
    Location: France
    Search Comp PM
    lossless codecs are not w/o loss: 2-3% compression i guess this answers your question.
    Quote Quote  
  7. Member buckethead's Avatar
    Join Date: Jul 2008
    Location: United States
    Search Comp PM
    According to tmpgenc : Contour
    Reduces the picture color gradations within a specified range within a sharply changing
    color area, increasing the amount of fine detail.

    To my eyes, the contour filter has a focusing effect, that's the best way that I can explain it. It drives up the noise content, but voila, neatvideo kills that in its sleep.

    I'll have to look into capturing using virtualdub. Up to now I've avoided pc capture due to horror stories of dropped frames, sync problems and other gremlins.

    I knew you would get me to thinking !!
    Quote Quote  
  8. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    Originally Posted by themaster1
    lossless codecs are not w/o loss: 2-3% compression i guess this answers your question.
    Not really. You are mistaken. Lossless is lossless. That's why they are called lossless.

    I hope they achieve better than 3% compression - but this compression is without loss, just like Zip, for example.

    Cheers,
    David.
    Quote Quote  
  9. Member Ethlred's Avatar
    Join Date: Feb 2008
    Location: United States
    Search Comp PM
    Not really. You are mistaken. Lossless is lossless. That's why they are called lossless.
    They are called lossless because they are mathematically lossless. Ignoring roundoff errors. But there are round off errors. At least according to what I read about lossless steps in mpeg and I think all codecs are likely to have similar errors.
    Quote Quote  
  10. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    Originally Posted by Ethlred
    They are called lossless because they are mathematically lossless. Ignoring roundoff errors.
    You don't have to ignore "round off" errors because there are none. The difference between the predictors and the actual data is stored exactly; reconstruction is exact. This part is 8-bit integer arithmetic - hardly a challenge!

    http://neuron2.net/www.math.berkeley.edu/benrg/huffyuv.html

    Cheers,
    David.
    Quote Quote  
  11. Member AlanHK's Avatar
    Join Date: Apr 2006
    Location: Hong Kong
    Search Comp PM
    Originally Posted by buckethead
    If all is well, possibly burn the complete clip to dvd-r gold disk. (Necessary/best brand?)
    "Gold" DVDs are just hype.
    Best brand is Taiyo Yuden. Verbatim is well thought of (and may sometimes be rebadged Taiyo Yuden).
    Quote Quote  
  12. Member themaster1's Avatar
    Join Date: Nov 2006
    Location: France
    Search Comp PM
    Originally Posted by 2Bdecided
    Not really. You are mistaken. Lossless is lossless. That's why they are called lossless.
    I hope they achieve better than 3% compression - but this compression is without loss, just like Zip, for example.
    I guess this picture will enlighten you then



    And i guess this will enlighten you even more.

    By the way, i use to work on mpegs more so this is absolutely required while on real time caps avi lossless, it's more a precaution.
    Quote Quote  
  13. Member Ethlred's Avatar
    Join Date: Feb 2008
    Location: United States
    Search Comp PM
    From the site themaster1 linked to:
    Some codecs are capable to accept input data in several color spaces (RGB, YUY2, YV12, etc.), however, compression is not always "lossless". Some codecs "silently" perform the conversion of color spaces, thus, raising a degree of compression. Quality losses coming out of such conversions are not visible, but surely exist. In the given testing work codecs is compared separately for various color spaces, thus in each color space full absence of losses is guaranteed!
    So as far as I can see HuffYUV v2.1.1 is lossless, unless you perform color space conversion, unlike the Mpeg I was referring to in my previous post. Perhaps I am missing something here. What you circled was a compression ratio. If it is straight integer, as it appears to be, then it is possible to be lossless. Removal of redundancy is not a sign of loss of data.

    Than you 2Bdecided for your information and the link.
    Quote Quote  
  14. Member buckethead's Avatar
    Join Date: Jul 2008
    Location: United States
    Search Comp PM
    Ooh, another newbie question, this time about virtualdubmod. After using neatvideo and saving as an avi file, which compression codec should I be using ? Up to now, I've been dealing with one or two minute clips which yield 4 or 5 gig files. Later, I will be using one and two hour clips, which really scares the begeezus out of me. Is huffyuv the one for minimal loss and manageable file sizes ?
    Quote Quote  
  15. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    I think Ethlred has covered the lossless is lossless issue.

    buckethead - HuffYUV is lossless, but these are still big files (about half the size of uncompressed). If you don't have the space, and your video is compatible (frame size, colour space, frame rate, interlaced), then DV-AVI can be a good choice, but is quite lossy - fixed 25Mbps.

    There are some great "perceptually lossless" (i.e. lossy, but not very lossy!) codecs, but they're not free. The Canopus HQ codec is really good. Whereas my machine crawls when using HuffYUV AVI as a source, it flies with a Canopus HQ AVI, and the data rate is lower.

    MJPEG at a high datarate is another option.

    I think most people using "free" software use lossless for intermediate files - HuffYUV or Lagarith. Many people using certain commercial software packages use the bundled perceptually lossless codec.

    Hope this helps.

    Cheers,
    David.
    Quote Quote  
  16. Member themaster1's Avatar
    Join Date: Nov 2006
    Location: France
    Search Comp PM
    Originally Posted by Ethlred
    So as far as I can see HuffYUV v2.1.1 is lossless, unless you perform color space conversion, unlike the Mpeg I was referring to in my previous post. Perhaps I am missing something here. Removal of redundancy is not a sign of loss of data.
    Well i'm always amazed with the arrogance such as expressed with the following words declares full quality losses absence

    I will consider their assertion when i'll see 10 different tests from 10 differents companies if you don't mind.
    Quote Quote  
  17. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    If lossless data packing wasn't lossless, you'd have people shouting that ZIP was broken. It isn't.

    If lossless audio packing wasn't lossless, you'd have people shouting that FLAC was broken. It isn't.

    If lossless video packing wasn't lossless, you'd have people shouting that HuffYUV was broken. It isn't.

    Which part of "mathematically lossless" don't you understand?! You get the same numbers out that you put in. They are integers - they don't get rounded!


    Arrogance? I'm not the one questioning millions of users, the software developers themselves, and the dictionary!

    Cheers,
    David.
    Quote Quote  
  18. Member Ethlred's Avatar
    Join Date: Feb 2008
    Location: United States
    Search Comp PM
    I will consider their assertion when i'll see 10 different tests from 10 differents companies if you don't mind.
    You could run your own test.

    There is most definitly the possibility of lossless compression. What is so hard to comprehend about that? Rar Zip 7z Ace and others have been doing it for a long time. Some of the techniques are as simple as run length encoding and that used occasionally on the Apple ][. Twenty five years ago at least. Of course there were a lot less colors on the Apple so run length worked pretty well. Decompression and writing to the screen was slow enough to be visible so it was really clear that it was used.
    Quote Quote  
  19. Video Restorer lordsmurf's Avatar
    Join Date: Jun 2003
    Location: Want my advice? PM me.
    Search Comp PM
    Originally Posted by buckethead
    First off, a hale and hearty thank you to all the kind and knowledgeable folks here who take the time to share their views. Recently, I discovered the wonderful world of neatvideo and all it can do. This has inspired me to recapture my vhs collection (shoot me now). While capturing this time, I want to make sure that the black level devil is minimized. Here is my planned chain :

    1) Purchase a Panasonic AG-1980, which is on the way. Should I use tbc on all tapes, or only the ones that look shaky without it? Should dnr be used here, or should I save this for neatvideo ?

    2) Run the output to a Sign video proc amp. I am hoping to set brightness, contrast, chroma and hue here as accurately as possible. From what I've learned on this forum, I believe (duck and cover) that the output from the proc amp should be 7.5 ire for vhs video.

    3) Possibly a Sign video dr-1000 detailer/enhancer. I'm thinking about software sharpening instead, but if this is a must-have, I'll do it.

    4) Capture in XP mode on a Panasonic e80. I know some here have a real hatred for this model, but so far it has served me well. Alternately, I could use my ez27, if that is better.

    5) Import a short sample clip via dvd-ram to tmpgenc 4.0. Here I plan to use the contour filter, which works magic on vhs, and possibly some fine tuning of brightness, contrast, chroma and hue, so long as the black level is not altered.

    6) Import this file to Virtualdub and neatvideo. I'm guessing that my pc monitor should be a large crt for this step, as well as the previous step.

    7) Burn this file to dvd-rw using tmpgenc 3.0. Now, play this dvd on a full sized tv monitor to make a final check. If all is well, possibly burn the complete clip to dvd-r gold disk. (Necessary/best brand?)


    I also have several PAL tapes which I will play through my Panasonic AG-W3. I have no seperate TBC, so I'm hoping my Prospec dv-773 will help.

    okay, all tips, hints and hellish experiences are welcomed ...
    It's a shame other posters went on a silly tangent about codecs.

    To answer your question, or some of it at least,

    - Yes, XP mode best choice on that recorder
    - Pansonic mucks up luma too, so the IRE settings in-machine sort of suck as a "correction" method, using a proc amp can be better to adjust away from whichever "dark" setting you pick on it.
    - NeatVideo is a time-consuming pain in the ass, I've never used it for that reason. I stick to VirtualDub or TMPGEnc filters.
    - Output from proc amp should be offset to whatever the recording machine needs as compensation, if that's the reason for adjustment. Don't worry about numbers as much as your own eyes on a well-calibrated display. Sometimes the numbers lie anyway, if the original was shot crappy yet "meters" otherwise. That happens.
    - Panasonic AC3 is fubar, so you'll often need to convert it to something else. I have to use Womble MPEG Video Wizard, convert to MP2 at a higher bitrate (less chance of loss), then dump into audio editing software to remove hiss and other noise (if any). You've overlooked this.

    More or less what you're doing sounds like a decent method. I'd watch for too much filtering or re-encoding in software. You can do it, just remember to set non-final video files to high bitrates if compressed (I-frame only MPEG-2 at 9000-15000k) or just use uncompressed AVI if you've got hard drive space (75GB/hour). There's a reason I have almost 5TB of space.
    Quote Quote  
  20. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    That's all good advice - except I wouldn't call I-frame only MPEG-2 "high bitrate" at 9Mbps - DV AVI takes 25Mbps for roughly equivalent data and compression, and even that occasionally causes visible differences. I wouldn't encode to MPEG-2, at any bitrate, for an interim stage. 25-50Mbps would be fine, but MPEG-2 encoding also takes time; DV or lossless is far quicker. Beware colour space issues with DV.

    Cheers,
    David.
    Quote Quote  
  21. Member buckethead's Avatar
    Join Date: Jul 2008
    Location: United States
    Search Comp PM
    I hate to admit it, but audio quality is a low priority on these transfers, because most were captured from glorious mono tv broadcasts.

    I just got all the gear and am starting to play. The AG-1980 puts out a beautiful picture. I need to decide whether to use the detail mode or the nor mode and how much to let the Sign Video DR-1000 detail and/or noise suppress. I'm using a starting tack of enhancing until noise just appears and then backing off a tad from there and then seeing if Neatvideo can work its magic without softening the underlying details too much. The msu smart deblocker is a great tool for multiple generation (tapes from old trades) as well.

    I have both a BVP-4 and a GTH correctors, but must admit that the Signvideo units put out a much cleaner, noise free image.

    I'll post updates as the project continues ...
    Quote Quote  
  22. Video Restorer lordsmurf's Avatar
    Join Date: Jun 2003
    Location: Want my advice? PM me.
    Search Comp PM
    MPEG-2 I-frame only can easily be superior to DV compression. The bitrate is not the sole issue at hand here. As you said, beware of colorspace compression problems with DV25.
    Quote Quote  
  23. Member buckethead's Avatar
    Join Date: Jul 2008
    Location: United States
    Search Comp PM
    Two points of interest :

    1) I noticed that the dvd-ram files when imported to tmpgenc 4.0 were 704x480. Should these be resized in tmpgenc or virtualdub to 720x480 before burning ?

    2) I love dvd-ram for two reasons. First, I can add to a disk anytime I want and play it back on any ram player without finalizing. Secondly, if I trade with others, I can extract from a disk in tmpgenc. If my new archiving medium is dvd-r or dvd-rw, can I still do something similar ?
    Quote Quote  
  24. Video Restorer lordsmurf's Avatar
    Join Date: Jun 2003
    Location: Want my advice? PM me.
    Search Comp PM
    1 -- Not necessary, 704 is valid DVD spec res.
    2 -- Extraction from DVD-R/+R is as easy as DVD Decrypter in IFO mode, pulling it to singl stream.
    Quote Quote  
  25. Member buckethead's Avatar
    Join Date: Jul 2008
    Location: United States
    Search Comp PM
    Ok, now for the black level devil.

    I set my proc amp for 7.5 IRE black and 100 for max brightness. My panny e80 is set to "darker" for s-video input. When played back on the e80 with the output set "lighter" as recommended, things look ok, ie similar to what my monitor shows while recording. When played on a JVC player, the blacks appear too dark.

    Should I be shifting the levels in virtualdub with an input of 0 to 255, to an output of 16 to 235 ? I believe I'm reaching critical mass at this point.

    Also, lordsmurf, can dvd+r be played on a standalone dvd player without being finalized ?
    Quote Quote  
  26. Member 2Bdecided's Avatar
    Join Date: Nov 2007
    Location: United Kingdom
    Search Comp PM
    Originally Posted by lordsmurf
    MPEG-2 I-frame only can easily be superior to DV compression.
    I agree, just not at 9Mbps (AFAIK). Or do you have a very clever encoder in mind?

    Cheers,
    David.
    Quote Quote  
  27. Video Restorer lordsmurf's Avatar
    Join Date: Jun 2003
    Location: Want my advice? PM me.
    Search Comp PM
    MPEG-2 I-frame VBR with max 15000k + min 9000k, I didn't say CBR 9000k
    A good encoder like Procoder or MainConcept has no problems here.

    DVD+R must be finalized.
    Quote Quote  



Similar Threads