VideoHelp Forum
+ Reply to Thread
Page 6 of 7
FirstFirst ... 4 5 6 7 LastLast
Results 151 to 180 of 197
Thread
  1. Originally Posted by videeo View Post
    So... When I setup the ES20 for pass-through are there any settings I should look out for? I haven't used a DVD recorder before...

    I'll be connecting via composite output on the VCR to the ES20 (rear input). Then s-video output on the ES20 to the capture card.

    Anything I should look out for on the capture card settings (Diamond ATI 750 HD USB)?
    Assuming you're using Virtualdub, the most important setting to look out for is probably this:
    After going to File->Capture, go to the Audio menu (I think) and disable audio playback during capture. Back when I used my TV Wonder 650, I got massive frame drops until I did that, even on a fast computer. Disabling audio playback is annoying, but it won't affect audio capture. (I'd be more precise about the name of the setting to disable, but I'm away from my capture computer at the moment.) My current All-in-Wonder setup doesn't have any problem with audio playback, but that's because audio is captured through my sound card.

    If you ever get any dropped or duplicate frames (Virtualdub keeps track of how many are lost or inserted on the PC side of the equation), try the following settings: http://forum.videohelp.com/threads/343892-Audio-Video-sync-offset-when-transferring-ta...=1#post2143847

    Also, go to Video->Capture Pin (I think) to make sure you capture at the resolution and colorspace you want (probably 720x480, YUY2). You can set the software proc amp from here (more on that in a second). I forget exactly where to go to set the input, but make sure you select the correct one (S-Video vs. composite). Go to Video->Compression (I think) to pick the codec Virtualdub will encode your captured stream with. I use ffdshow's Huffyuv in YUY2 mode (you may have to download and install ffdshow-tryouts).

    As far as the software proc amp is concerned, it's pretty handy. (It's also high-quality: It either uses great dithering or takes effect before quantization to 8 bit channel depth, because I never see any image or histogram degradation from it.) Hopefully your TV Wonder 750 will have a working Video->Levels dialog box so you can make adjustments while previewing your video input...if not, you'll have to set the proc amp non-interactively from Video->Capture Pin (I think). It's more of a hassle, but you have the same controls at your disposal.

    To set your software proc amp, you may find it helpful to hit Video->Histogram and watch the histogram at the bottom. The red part on the left side shows the black that's pushed below the visible level (luma 16 = RGB {0, 0, 0} for typical playback), but it's not lost forever even after you capture it unless it's totally cut off on the left. The red part on the right side shows the white that's pushed above the visible level (luma 235 = RGB {255, 255, 255} for typical playback). Technically speaking, your VCR will likely clip whites to some degree, and the TV Wonder's AGC will do the same...but as long as you don't let anything get cut off, you won't clip any more.

    Brightness is an additive control, and it affects both ends of the spectrum equally. Set brightness first (then tweak afterwards once you have good ballpark brightness/contrast levels). You don't want a whole lot of red showing, but black levels can be tricky to set, because the head switching noise at the bottom of the picture is very black, often much darker than the darkest black in your actual video. If there's a hump on the left followed by a dip, followed by a rise that leads into the rest of the histogram, I'd probably start the blue part of the histogram at the beginning of the rise and keep the leftmost hump and dip in the red. If you have the brightness so high that the head-switching noise is "in the blue," you'll end up with a washed out picture without any true black in the actual video content.

    If you have trouble finding the right black point and you're afraid of black crush, try going to Video->Cropping and temporarily cropping out the head-switching noise and borders. This will give you a clearer idea of what the histogram looks like for the real picture. Just make sure to uncrop afterwards unless you want to capture with strange frame dimensions! I've forgotten before, and it cost me a few hours of redos.

    Contrast is a multiplicative control. Adjust contrast so that no white value ever goes into the red range...or at least not by much. You'll want to avoid blowing out your whites any more. (That said, there might be superwhites in your head noise too; if so, you'll hopefully be able to differentiate them from the actual signal. This may be worse if you increase the proc amp's sharpness.) After you change the contrast, you may have to tweak brightness again to bring the black point back where it was...then tweak contrast again, etc. The adjustment should get smaller and smaller each time though until you arrive at the settings you're looking for.

    When you set your proc amp settings, make sure to use a tape that has lots of dark and bright scenes, so you get a good idea for the full range of video signal coming from your equipment. Keep in mind that the optimal settings will likely differ if you switch VCR's or introduce any new equipment to your chain (different TBC, etc.). Moreover, consumer camcorders tended to be terrible about following standards, so if you capture home movies from different camcorders, so you may have to switch up your settings when you put in a movie made by a camcorder different from the last one. All that said, the AGC in your TV Wonder should keep these adjustments smaller than they might be for me (AGC meddling occurs before the signal ever reaches the software proc amp).

    Once you're ready to capture, go to File, then um...Set Capture Name maybe? I forget the exact name, but anyway, this lets you say where you want to save your capture and under what file name. Then hit Capture->Start Capture (or something similar), and your computer will go to work capturing the input, encoding it, and saving it to your drive. You can manually stop it at any time with the same Capture menu, and you can set a timer in the Capture->Stop Conditions (or is it Tools->Stop Conditions?) dialog box.

    Quick tip: Once you're done with your settings, exit out of capture mode and exit out of Virtualdub, then restart Virtualdub. That way, you can be sure your settings will stick. It's annoying messing with tons of options only for Virtualdub to crash without ever saving your settings.
    Last edited by Mini-Me; 4th Apr 2012 at 12:24.
    Quote Quote  
  2. Originally Posted by videeo View Post
    So... When I setup the ES20 for pass-through are there any settings I should look out for? I haven't used a DVD recorder before...
    Disable the noise reduction filter in the DVD recorder. It causes ghosting. At least the one in my ES15 does.
    Quote Quote  
  3. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    jagabo: My ES20 DNR hasn't caused ghosting, so far. Its DNR is similar to that of the ILS encoder/decoder chip in the ES10 (same chip. ILS, right? LSI? I might have the letters out of order, but same chip in the ES10 and ES20. The ES15 had a Panasonic-made chip). IMHO the ES20's DNR visibly softens the image, kinda looks like a temporal filter. For fairly clean tapes, I turn it off. For grungy ones, test first to see what you get. In the samples videeo has submitted so far, I don't see the need for the ES20's DNR.

    Originally Posted by videeo View Post
    So... When I setup the ES20 for pass-through are there any settings I should look out for? I haven't used a DVD recorder before...

    I'll be connecting via composite output on the VCR to the ES20 (rear input). Then s-video output on the ES20 to the capture card.

    Anything I should look out for on the capture card settings (Diamond ATI 750 HD USB)?
    I don't know the ATI in detail, but the ES20's default input/output settings are the ones to use for typical NTSC sources. The others are for various foreign webcam and tape IRE levels, which you won't have to worry about for now.
    Last edited by sanlyn; 21st Mar 2014 at 19:30.
    Quote Quote  
  4. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    Thanks, that's a great tutorial on setting the levels Mini-Me. Yes, I'll be using VirtualDub for capturing.

    On the ES20, do I just power it up and just pass-through the signal, or do I need to be in a certain mode to ensure that the TBC is on?
    Quote Quote  
  5. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Power on, set ES20 input to Line 1.
    Last edited by sanlyn; 21st Mar 2014 at 19:31.
    Quote Quote  
  6. Member Deter's Avatar
    Join Date: Dec 2007
    Location: United States
    Search Comp PM
    Not sure about the ES10 or ES20, however on the Panasonic DVD recorder that I have the S-Video outputs have two controls for the Black Levels. It depends on the video which settings looks best. It is Light or Dark. If I want the middle level you have to record to that DVD recorder. (Not sure about the ES20 unit) Never record in 2 hour mode. Either XP mode or FR mode, more than likely this unit will only be used as a passthrough.

    You can also play with aspect ratios 4:3 and 16:9

    The DNR feature on these recorders are not the best, just turn it off.

    1st thing you want to make sure is that you get a good signal via the s-video cables. You just need to check the picture when it is being played on the TV.

    Again PLEASE CHECK YOUR PICTURE WITH YOUR S-VIDEO connections make sure you are getting a good clean signal. Bad Cables may = Inference.

    Sound wires, again make sure they are good, they don't have to be gold plated, your free cables you get from comcast may not be the best.

    You are running wires through 3 or 4 different units....
    Sometimes you miss a connection so make sure both channels of the audio are working.


    VCR to ES20 to PC

    If you use an external TBC in the mix.

    NOTE:
    if you only send a video signal to the external TBC.

    (Very good chance it will un-sync the audio by a few frames, you need to check this and make sure it is in sync)

    The good news is the problem is normally a set amount of frames....Lets say it is 7 frames out of sync...Than you just move the audio by that amount.

    For me it is easy to sync the audio, however for the avg user it may be a very hard to re-set, figure out the sync point or impossible to detect.

    If the video is totally messed up with the audio sync

    (for example)

    At the start it is off by zero frames and at the end of the video it is off by 1 minute. You have a problem and need to re-do the video and figure out were it got messed up.


    I have never used the ES10 or ES20, but I got my Panasonic DVD recorder back in 2007. On a few tapes I was having tracking problems at the bottom of the screen. Than by mistake I was using the unit to record something else, than played back a tape with bottom picture tracking problems. On the playback the screen was cropped different and the picture was perfect, no tracking issues. Plus the unit has a Digital Tuner and I always like the sharp picture of that PANASONIC DVD RECORDER.

    I still use that recorder a lot for recording DIGITAL TV.

    FINAL NOTE:
    The picture tearing at the top of the screen, I have only seen JVC decks have this problem, my PANASONIC VCR's, it has never happened....

    QUESTION:
    How is the picture on the SR-W5U/SR-W7U units?
    How do VHS tapes look?
    How does the higher end stuff play back?
    Do the videos have normal VHS errors? Or do they play back like DVD's?
    Last edited by Deter; 5th Apr 2012 at 21:51.
    Quote Quote  
  7. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    ES20 arrived. Here is a raw capture (DNR off) from the park bench scene posted earlier. The signal is a lot more stable, and the left border is pretty much a straight line, the right border is better, but not completely straight. I tested out the DNR, and it wasn't that great as noted earlier for this source. I am much happier with the capture, there is some annoying combing in some areas when the clips I tested are IVTC'ed. I'm thinking there still must be some alignment issues with lines within the frames causing it. Could a better VCR or source help? What do you guys think?


    test-19-es20-snippet.avi
    Quote Quote  
  8. VInverse() is good at getting rid of residual combing artifacts after IVTC.
    Quote Quote  
  9. Originally Posted by videeo View Post
    I'm thinking there still must be some alignment issues with lines within the frames causing it. Could a better VCR or source help? What do you guys think?
    I see that crap a lot as well. For some reason the fields aren't always aligned properly within the frames. It's nothing to do with TBCs, or with VCRs, I don't think, as I've seen it on three of them and two of the VCRs are outstanding. If it's bad I use QTGMC, select the better of the two fields/frames in a pair, and then decimate if needed. If that residual combing isn't bad then I use the same as jagabo recommended, vinverse.
    Quote Quote  
  10. It's caused by horizontal sync time base errors. Remember fields are recorded sequentially on VHS tape. So the two fields that comprise the frame have different horizontal jitter. Hence the residual comb artifacts, even if the IVTC is perfect.
    Quote Quote  
  11. Originally Posted by jagabo View Post
    It's caused by horizontal sync time base errors.
    Then a TBC should correct it, shouldn't it? I'm running two, both a line TBC in the VCR as well as an external frame TBC. If I'm reading what you're saying about the cause correctly, the line TBC in this Panasonic VCR should fix it. But it doesn't. It fixes a lot, so it's not as if it's broken or of poor quality or anything like that. I don't think.

    And if you're also saying a VHS tape stores fields rather than frames, then thanks because I didn't know that.
    Quote Quote  
  12. Originally Posted by manono View Post
    Originally Posted by jagabo View Post
    It's caused by horizontal sync time base errors.
    Then a TBC should correct it, shouldn't it?
    TBCs aren't perfect. There's still a little jitter left, even in the best of circumstances. With second or third generation tapes they may do little. If the jitter is embedded into the signal they can do nothing.

    Originally Posted by manono View Post
    And if you're also saying a VHS tape stores fields rather than frames, then thanks because I didn't know that.
    All interlaced analog video (ie, all standard definition video) is transmitted as a continuous, linear, one dimensional signal; one scanline at a time, one field at a time. It travels over composite or s-video cable as scanlines and fields.

    http://zone.ni.com/devzone/cda/tut/p/id/3020
    Name:  b2ae74b4555.gif
Views: 245
Size:  14.5 KB

    VHS is a linear analog recording of that analog signal onto tape. On playback the player just reads that signal off the tape and sends it down the cable. The TV scans that signal onto the two dimensional screen as scanlines and fields, making a picture. A capture device captures two successive fields and weaves them into frames.

    When a VHS tape is recorded and played back the speed of the rotating drum that reads the signal varies a little bit. This causes the duration (position and length as seen on the screen) of each scanline to vary:

    Name:  vary.jpg
Views: 284
Size:  12.8 KB

    A line TBC works by digitizing the sync pulses as well as the picture information between them. It then measures the distance between successive sync pulses and stretches or shrinks the entire line to match the standard.
    Last edited by jagabo; 7th Apr 2012 at 06:55.
    Quote Quote  
  13. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Either the original tape is damaged for a stretch & player not tracking it well, or the source might have been produced that way. I think it's the first rather than the second, because the O.P. has submitted earlier versions played on another VCR with fewer color problems, better tracking, and less head switching noise at the bottom.
    Last edited by sanlyn; 21st Mar 2014 at 19:31.
    Quote Quote  
  14. Member Deter's Avatar
    Join Date: Dec 2007
    Location: United States
    Search Comp PM
    The picture looked ok to me, it had a slight flicker (which could be caused by a poor or lose cable)

    Don't have the codex for this type of AVI file, played it in Virtual Dub.

    Please understand when you play VHS recordings on a TV, you don't see the excess scan lines.

    It looked pretty normal.

    You normally have scans lines at the bottom of the picture, and the edges are usually never perfect, even on digital stuff. Most of the time it makes no difference.

    However if you want to watch it on your PC and never be bothered by the scan lines, than just add boarders like so. Personally I don't do this cause the playback is for the TV and you always lose 5 to 10% of the picture, depending.




    Attachment 11857
    Quote Quote  
  15. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Originally Posted by Deter View Post
    The picture looked ok to me, it had a slight flicker (which could be caused by a poor or lose cable)
    From the earlier posts, the slight flicker is partly on the tape and partly the fault of the VCR. The owner has another VCR with better tracking, but doesn't like it.

    Originally Posted by Deter View Post
    Don't have the codex for this type of AVI file, played it in Virtual Dub.
    You have the codec somewhere, or VirtualDub couldn't play it. It's compressed with Lagarith.

    Originally Posted by Deter View Post
    Please understand when you play VHS recordings on a TV, you don't see the excess scan lines.
    Yes you do. Not everyone enables overscan on their TV. Many turn it off to watch 100% of the image. There are easy ways to handle it, though (see previous posts).

    Originally Posted by Deter View Post
    It looked pretty normal.
    Not to me. Image was more stable, but it has many of the same color problems as earlier versions -- and they were not easy to fix in processing. If you don't see the problem at your end, you need a better monitor.

    Originally Posted by Deter View Post
    You normally have scans lines at the bottom of the picture, and the edges are usually never perfect, even on digital stuff. Most of the time it makes no difference.

    However if you want to watch it on your PC and never be bothered by the scan lines, than just add boarders like so. Personally I don't do this cause the playback is for the TV and you always lose 5 to 10% of the picture, depending.
    See previous scripts that covered head switching noise, border noise, and centered the image in the frame.




    Attachment 11857
    Last edited by sanlyn; 21st Mar 2014 at 19:31.
    Quote Quote  
  16. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    Originally Posted by Deter View Post
    Please understand when you play VHS recordings on a TV, you don't see the excess scan lines.

    You normally have scans lines at the bottom of the picture, and the edges are usually never perfect, even on digital stuff. Most of the time it makes no difference.

    However if you want to watch it on your PC and never be bothered by the scan lines, than just add boarders like so. Personally I don't do this cause the playback is for the TV and you always lose 5 to 10% of the picture, depending.
    The sides will still be visible since black bars are on the sides of the frames. The uneven border will show, at least on my TV when I playback MPEG2. I mentioned the left and right sides since I take it as an indicator of how well the line TBC is able to get the lines to the theoretical width along the whole frame.

    So I was able to borrow a nice JVC unit with TBC. It's a nice unit with low mileage. It seems like it is more capable than the ES20. I have not tried the ES20 in series with the JVC unit yet (I only have 1 s-video cable), it might be too much processing...

    I did some color filtering (along with noise filtering) with MSU Color Enhancement with it set to darken highs and brighten lows a little. I also turned on color enhancement to the first level and activated auto levels. If I look at what it does on a frame-level basis, I think it does an OK job especially considering it is more automated. However, when I play back the clip, I notice that there is a temporal aspect to it, where it takes a frame to adjust to a scene change. It also looks like its doing some other fluctuations with levels as well. I was wondering is there another tool (VirtualDub or AviSynth) that can provide an adaptive technique of color enhancement and levels, while also being clever in detecting scene changes, but remaining stable? I have attached a clip below. It was captured in AUTO mode, would it be better to capture in EDIT (it seems a bit more noisy)?
    Attached Files
    Quote Quote  
  17. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Originally Posted by videeo View Post
    The sides will still be visible since black bars are on the sides of the frames. The uneven border will show, at least on my TV when I playback MPEG2. I mentioned the left and right sides since I take it as an indicator of how well the line TBC is able to get the lines to the theoretical width along the whole frame.
    What is the "theoretical width"?.

    Originally Posted by videeo View Post
    So I was able to borrow a nice JVC unit with TBC. It's a nice unit with low mileage. It seems like it is more capable than the ES20. I have not tried the ES20 in series with the JVC unit yet (I only have 1 s-video cable), it might be too much processing...
    With two line tbc's in tandem, the second tbc will see few, if any, line errors. But the ES20 will add its own processing to the image and might help correct some audio sync timing if the VCR gives you a problem there.

    Originally Posted by videeo View Post
    I did some color filtering (along with noise filtering) with MSU Color Enhancement with it set to darken highs and brighten lows a little. I also turned on color enhancement to the first level and activated auto levels. If I look at what it does on a frame-level basis, I think it does an OK job especially considering it is more automated.
    It didn't look OK to me. The candy-cane color effect is a hallmark of JVC's DigiPure (I don't know why they use the term "pure", but many people like the effect). You can darken the darks and brighten the lows with any of several better filters without screwing up the color. The filters and autogain have shifted gamma too far to the right most of the time. Red and blue look more like plastic neon than they did with the other players. As I said, many people like these effects. IMHO it gets rather dreadful at times, and would be impossible to correct.

    Originally Posted by videeo View Post
    However, when I play back the clip, I notice that there is a temporal aspect to it, where it takes a frame to adjust to a scene change. It also looks like its doing some other fluctuations with levels as well.
    It doesn't "look like" fluctations, it is fluctuating. It's often called "pumping", and it happens every time darker or brighter objects enter or leave the frame. The color balance changes with every shot and every change in the camera's view. None of the colors are accurate. The most annoying part was the scene with the kid in denim; he changed color 5 times in a few frames, varying between pink, green, red, orange and blue. The background colors changed as well (orange shrubbery ? Interesting, but not something I'd want to deal with).

    Originally Posted by videeo View Post
    I was wondering is there another tool (VirtualDub or AviSynth) that can provide an adaptive technique of color enhancement and levels, while also being clever in detecting scene changes, but remaining stable?
    There are other similar filters. They all work the same way. Are you aiming for color "enhancement", or color correction? None of them correct color properly.

    Originally Posted by videeo View Post
    I have attached a clip below. It was captured in AUTO mode, would it be better to capture in EDIT (it seems a bit more noisy)?
    AUTO mode softens the image. Upsampled on any HDTV, the image will look somewhat out of focus. Most JVC owners would dub or capture in EDIT mode, which supposedly transmits the image as-is. But that's not quite true; with the tbc/DNR circuits enabled, what you see is digital artifacts from JVC's primitive processing. If you turn off the processing, you disable the tbc. This reminds me why I stopped using JVC's long ago. In EDIT mode motion is blurred, often with artifacts like ghosting and trailing, and then sharpened; in AUTO, motion is simply blurred, which is obvious in this video. IMHO the softening and loss of fine detail and textures is characteristic of JVC's processing. It's the opposite of the over-sharpening often done by other VCR's. After fine details and subtleties of texture are stripped away, you don't get them back.

    Many people prefer the characteristic "look" you get with this combination of player and filters.
    Last edited by sanlyn; 21st Mar 2014 at 19:32.
    Quote Quote  
  18. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    Originally Posted by sanlyn View Post
    What is the "theoretical width"?
    Who knows, but if all the lines are all the same, then I should end up with straight sides on the frames.

    Originally Posted by sanlyn View Post
    With two line tbc's in tandem, the second tbc will see few, if any, line errors. But the ES20 will add its own processing to the image and might help correct some audio sync timing if the VCR gives you a problem there.
    I did another test with both in tandem and the JVC in EDIT mode. It is attached.

    Originally Posted by sanlyn View Post
    It didn't look OK to me.
    You're right, I tested it on my TV and it didn't look good. At this point I don't really trust my eyes since I keep trying to correct colors and then when I check on my TV, it looks bad. I guess I need help here. I think overall if I can clean up a little noise and fix the levels and color, it should look good.

    Originally Posted by sanlyn View Post
    Many people prefer the characteristic "look" you get with this combination of player and filters. The advantage to processing your own video is that you get to do what you want with it.
    If I can do it better than automated tools...
    Attached Files
    Last edited by videeo; 10th Apr 2012 at 21:06.
    Quote Quote  
  19. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    I found my laptop can output s-video, so I am now able to get real-time feedback on the changes I am making to color.

    It seems when I recompress using Lagarith, the color somewhat gets faded, is there a way to fix it?
    Quote Quote  
  20. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Neuther Lagarith nor huffyuv have any effect on color, in themselves. Make sure in Avisynth that you set Lagarith to the desired colorspace. RGB 24 is its default, which is also VirtualDub's default. But Lagarith can be set for RGB32, YV12, or YUY2, and often it doesn't re-set to RGB unless you tell it to. In VDub's top menu click on "Video......Compression......." Select Lagarith, then select Lagarith's "Configure" options.
    Last edited by sanlyn; 21st Mar 2014 at 19:32.
    Quote Quote  
  21. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Originally Posted by videeo View Post
    It seems when I recompress using Lagarith, the color somewhat gets faded, is there a way to fix it?
    This shouldn't happen. As I mentioned, it might be due to colorspace changes in the background. To open a YV12 clip into VirtualDub, you shouldn't let VDUb use its default conversion to RGB. You should open that clip with Avisynth using a script like this:

    AviSource(C:\path to video\filename.avi")
    Assumme TFF()
    ConvertToRGB24(matrix="Rec601",interlaced=true)

    This insures that some VDub assumptions (such as Bottom Field First for YV12) don't occur. VDub will usually take the Rec601 matrix by default, but it won't hurt to specify. Rec601 is the color matrix for standard definition video. Many DVD players playback with Rec709 (High Definiton), but that's not a concern for the time being. There's not much difference between Rec601 and Rec709. A few TV's allow you to set an "Auto" colorspace for playback (the TV analyzes the video and uses a precise matrix), but most don't.

    The example script is for interlaced video. You'd use "interlaced=false" for progressed source. You might be opening the clip from a d2v project, which reads mpeg or other formats directly, before they convert to AVI. So instead of AviSource you'd use something like MPEG2Source. Modify the "source" statement accordingly. In this case we're working directly with AVI.

    You can view that file in VirtualDub and then save it in its original colorspace. So if the original AVI is YV12 or YUY2 and you want to keep it that way, you would do this in VirtualDub's top menu: Click "Video" then click "Color depth...". Select the preferred colorspace. RGB24 is the default. If you don't want that, select the clip's original colorspace (YV12 or YUY2, whichever applies). Then under "Video" select "Compression..." and use Lagarith or huffyuv. Huffyuv uses RGB or YUY2. If you want YV12, choose Lagarith and set Lagarith for YV12. Then Under "Video" select "fast recompress". In this case VirtualDub has converted to RGB for display ONLY. That RGB conversion will be ignored when the file is saved.

    If you apply any VirtualDub RGB filters, the video will be converted to RGB and then back to whatever colorspace you've specified. You want to avoid multiple colorspace changes whenever possible. Do all the work you require in YUV first, then go to RGB and stay there. If you're using something like HCenc to encode after you've done your RGB work, load that file into HCenc with an Avisynth script that converts from RGB to YV12 properly. Even if you're using a high-end encoder like Premiere Pro, MPEG will be encoded to YV12 anyway regardless of the colorspace used for other work. Yes, that's a conversion but it's not a multiple back-and-forth deal, which can wreck a video.

    Originally Posted by videeo View Post
    Originally Posted by sanlyn View Post
    What is the "theoretical width"?
    Who knows, but if all the lines are all the same, then I should end up with straight sides on the frames.
    I get your meaning. The borders should be straight. They might shuttle side to side a bit (that's what's called a "projector" effect), but it shouldn't wiggle or bend.

    Originally Posted by videeo View Post
    I did another test with both in tandem and the JVC in EDIT mode. It is attached.
    The ES20 is doing some DNR and comb filtering, but its tbc likely isn't doing anything. It does tend to push red and change levels a bit. Panasonic is famous (or infamous) for pumping the black level somewhat. More on that later.

    Originally Posted by videeo View Post
    You're right, I tested it on my TV and it didn't look good. At this point I don't really trust my eyes since I keep trying to correct colors and then when I check on my TV, it looks bad. I guess I need help here. I think overall if I can clean up a little noise and fix the levels and color, it should look good.

    Originally Posted by sanlyn View Post
    Many people prefer the characteristic "look" you get with this combination of player and filters. The advantage to processing your own video is that you get to do what you want with it.
    If I can do it better than automated tools...
    Originally Posted by videeo View Post
    I found my laptop can output s-video, so I am now able to get real-time feedback on the changes I am making to color.
    You can do it without automated tools. And you're right, you can't trust your eyes -- not at first, anyway. I find it impossible to work with video on a laptop, except for copying. In any case, color work and even black and white are difficult at first, for several reasons. Mainly, most people haven't developed a basic knowledge of color. The eyeballs are smart enough to know when something "ain't right," but an untrained eye can't say why. There's trial and error involved, and some research to do (most of it free, thank goodness). But it doesn't take long. You look at images that seem haywire, and one day something clicks: you sit up and say, "So that's how it works!". It's much easier and faster from there.

    For years I've looked in forums for detailed color tutorials. Nothing. Now, most people say "I only have a couple of videos. It's not worth it." If the source or the memories they invoke don't mean that much, one can always pay someone to archive them. The shop might have better equipment, but at cheap/quick rates the results won't be much better than your own if done correctly, and I've seen where the paid results can be awful. Might as well do it yourself, tape-to-DVD and be done. Don't destroy the tapes; you'll likely be sorry you did. But if they have value for you or yours, either pay big bucks for a good restoral or invest some time and effort with it. Likely it will never be perfect, even at high pay rates.

    Another reason people can't trust their eyes: they don't understand monitors. They don't know what those gizmos are doing to levels and color, nor why their video looks one way on a PC but completely different on TV -- or different from monitor to monitor. There are ways of getting this information. You don't have to be a physics major. A few test patterns can reveal much. Some procedures require special tools, but many users get decent results without them. Many think it's less trouble to fight it out with the equipment, shooting blind. Experience shows that it's easier to get some information first. When you have some free time (does it even exist?) try some of these links and old threads. Here's a recent post that explains the problem with monitors:
    http://forum.videohelp.com/threads/345226-Nervous-Video?p=2154037&viewfull=1#post2154037

    That post refers to a popular website. You might want to give it a try and browse the pages indicated by the icons across the top of the site. Can be a real eye opener. Has free stuff you can use:
    http://www.lagom.nl/lcd-test/

    Eventually this free stuff gave me fits. That's because I'm not just picky, I'm crazy. Most people wouldn't go thru the trouble I did (but many do). For not much cost you can get good stuff that works wonders. Photogs and videophiles would go for a piece of hardware called a colorimeter -- a puck-like device used to measure light and color from a monitor, and the software that goes with it. Companies that make this stuff do a thriving business, mainly because most of it actually works and many find it essential. This post describes the results of using that stuff. Has pictures, too:
    http://forum.videohelp.com/threads/335402-VHS-capture-critique?p=2083260&viewfull=1#post2083260

    So much for the hardware. Even if you don't actually do this stuff, you'll have a better understanding of color. In itself, that goes a long way. I'll try to work with your latest video and get up some samples for color work, later.
    Last edited by sanlyn; 21st Mar 2014 at 19:33.
    Quote Quote  
  22. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    test-32-jvc-edit-es20-snippet.avi has serious problems (field parity, among other things) and does not IVTC properly. Some bad color, especially at scene changes. The video looks soft and over-processed (thanks to JVC, mostly). There is no reason to use two tbc's and it's causing problems.

    Use YUY2 for analog capture, not YV12. I know you're going to YV12 for processing anyway (some plugins work in YUY2), but YUY2 captures more video info. Open the clip in VirtualDub and move one frame at a time, you'll see what I mean. Then use SeparateFields() and you'll see some bad stuff.
    Last edited by sanlyn; 21st Mar 2014 at 19:33.
    Quote Quote  
  23. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    Thanks sanlyn...

    Originally Posted by sanlyn View Post
    Use YUY2 for analog capture, not YV12. I know you're going to YV12 for processing anyway (some plugins work in YUY2), but YUY2 captures more video info. Open the clip in VirtualDub and move one frame at a time, you'll see what I mean. Then use SeparateFields() and you'll see some bad stuff.
    I think I did use YUY2 with the HuffyUV codec. This clip I probably saved with Lagarith, which is the default codec I am using after cutting clips to post or after doing some processing. The color issue I mentioned earlier with Lagarith is bothering me, if I save in HuffyUV, it does not have the faded output as compared to the output preview in VirtuaDub before re-compressing. In Lagarith I am saving as YV12.

    Originally Posted by sanlyn View Post
    test-32-jvc-edit-es20-snippet.avi has serious problems (field parity, among other things) and does not IVTC properly. Some bad color, especially at scene changes. The video looks soft and over-processed (thanks to JVC, mostly). There is no reason to use two tbc's and it's causing problems.
    Here is the script I just used to check out what you are talking about:

    Code:
    #- ----------------------------------------------
    #   Source is progressive hard-telecined 3:2 YV12
    #- ----------------------------------------------
    
    AviSource("G:\VIDEO\test-32-jvc-edit-es20-snippet.avi")
    
    #- -------------------------------------------------------
    #   Inverse telecine/decimate. Result is progressive video
    #-  in its original state as 23.97 fps movie film
    #- -------------------------------------------------------
    AssumeTFF().TFM()
    TDecimate()
    
    #- -------------------------------------------------------
    #   RGB conversion for VirtualDub. Omit if not wanted
    #-  and then save as YV12.
    #- -------------------------------------------------------
    
    ConvertToRGB24(matrix="Rec601",interlaced=true)
    I see what you mean, could it have been caused because I happened to cut this clip in a bad spot? I didn't think that was possible if I was working with a lossless file, but maybe the IVTC routine got tripped-up this time.

    I'll post another clip with all the JVC stuff disabled (if possible).
    Last edited by videeo; 11th Apr 2012 at 13:50.
    Quote Quote  
  24. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Your video will look worse if you disable the EDIT tbc and processing. That's the whole point of using the JVC to begin with. You should eliminate the ES20, or use one or the other - not both. You can't get away from softening effects with JVC; that's the way JVC plays tape, and that's why their owners insist on their use.
    Last edited by sanlyn; 21st Mar 2014 at 19:33.
    Quote Quote  
  25. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    Here's another "raw" snippet from the JVC in EDIT mode. I am now noticing the levels "pumping" on scene changes, but it might be from the capture card since I saw some commentary about it here in the forums. It could be from the JVC though since I haven't noticed it being this intense before.
    Attached Files
    Last edited by videeo; 11th Apr 2012 at 21:46.
    Quote Quote  
  26. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    Here is an mpeg of the tandem setup that is IVTC'ed and has some noise and color filtering. One thing I noticed is the file played over s-video through my laptop looks different than when I play the same file through my PS3. I think I'm slowly going insane...

    At this point I'm thinking I should figure out what setup I should use to get the best capture. I think I would rather trade less noise / stability with as little detail lost as possible over color issues, since those seem to be easier to fix in the long run. So I'm thinking it is either between tandem with JVC in EDIT or the JVC in EDIT on its own.
    Attached Files
    Quote Quote  
  27. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    For fun I tried the DVD burning capability of the ES20 to see how it looked, and it wasn't too bad. Since the tape is 50min long, I could burn it in XP mode. I viewed the DVD on my Macbook and noticed the color looked warmer than on my Thinkpad, so that machine is definitely not the one to use for final mastering. I burned it with th JVC in EDIT mode, but if I do it again, I would try AUTO mode.
    Quote Quote  
  28. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Originally Posted by videeo View Post
    At this point I'm thinking I should figure out what setup I should use to get the best capture. I think I would rather trade less noise / stability with as little detail lost as possible over color issues, since those seem to be easier to fix in the long run. So I'm thinking it is either between tandem with JVC in EDIT or the JVC in EDIT on its own.
    The best ways would be:
    1) The JVC in EDIT with no other line-tbc in circuit, to lossless avi (will be a smoothed image, with "JVC color" and motion blur)
    2) One of your other VCR's and the ES20 as pass-thru (good detail retention, less effect on color than JVC, less motion blur)

    I'd go with number 2. Your choice, but don't use both tbc's together. Using both will overprocess the image and caused some weird field polarity, color, and timing issues. 1) and 2) are time-proven methods used by most people for years. This is the only time I ever heard of anyone using both together, and it obviously created problems.

    AUTO mode is what most people would avoid for capture. You might just as well play it through a cheap Sharp VCR, which is what many JVC owners do when their JVC rejects bad tape. The image would look just about the same, but no tbc in the Sharp player.
    Last edited by sanlyn; 21st Mar 2014 at 19:33.
    Quote Quote  
  29. Member
    Join Date: Mar 2012
    Location: US
    Search Comp PM
    sanlyn, did you see the mpeg that I posted? Are there still field parity issues? I'm thinking maybe where I cut the clip tripped-up the IVTC because I didn't see it happen when I processed the full capture (as posted).
    Quote Quote  
  30. Banned
    Join Date: Oct 2004
    Location: New York, US
    Search Comp PM
    Test 32 looks better, videeo. Sorry, I'm working on a client's PC repair now, tuned in on my laptop. Looks like a black-level problem in a couple of shots, but that's more from the tape than anything else (saw that in earlier captures. Sloppy production work, IMHO). Hard to tell on my laptop screen, will tackle it when I get home late tonight. But looks much better!
    Last edited by sanlyn; 21st Mar 2014 at 19:33.
    Quote Quote  



Similar Threads