I take it the D-VHS doesn't use LSI, because of both chroma and mosquito noise. Back in the day, I could see it being as good as an ATI AIW in DVD-spec bitrates. But not now, not even 10 years ago.
There are worse machines, but I'll probably side with sanlyn on this. |
I've seen worse, to be sure (thank heaven it wasn't a DVD/VHS combo). But cleaning up lossy encodes is not the way I want to live my life. Anyone who thinks this capture doesn't need a good scrubbing isn't paying attention. But the thought of what it will look like later, well.... it's a good argument for lossless capture, even with a fairly clean tape source.
I realize the sample isn't a finished project. But in cleaning it up, by the time I got to the 4-frame horizontal dropout at frame 980 I'd be ready to throw in the towel trying to work with the lossy compression noise and low acutance. On the other hand, maybe there are further plans for the sample that haven't been mentioned yet. It would be interesting to know what they are and to see some results. |
The capture is lossy and it is not meant by any mean to be for editing, clean up or long term storage. It is just to show the D-VHS capability using its native format for capture which is MPEG-2 720x480 4:2:0 compared to other lossy formats such DVD and DV. If I would capture that for archival I would go lossless then MP4 for wider compatibility. Don't judge the book by its cover.
|
As far as I know, there were only 3 PAL decks, 1 from JVC, that I recentely finally bought and two from Thomson.
And one of them have HDMI, or digital output, not even component. Archiving, restoration, denoising, relax man, nobody cares, if people would want that, they wouldn't go to me, they would go to professionals who would charge them 5 tiems as much. I chose D-VHS for quality of transfer, so it would look as good as it can and I won't bother with any further adjustments. |
Quote:
As for "nobody cares", that's not true. There are pros, advanced users, and even rank beginners here who can make a typical capture look pretty good, and some have learned how to repair some really ugly videos. Many threads show how it's done. Some do it because they want to restore cherished memories. Others do it because they take pride in their work. But non-participants are also welcome. Quote:
|
Obviously, because people who contact me do not need that, they need simple VHS to DVD, or AVI transfer and the output is good enough and so far I fortunately didn't encounter badly damaged tape with signal degradation, so I live in a happy bubble for now.
The other thing about the pain of restoring already compressed video, it does happen when capturing VHS via firewire, but PAL D-VHS doesn't have any kind of digital output, only S-video, so it works as regular S-VHS deck. I do not deal in still photography at all, so that would be right. Shooting live video is another thing, as I do a little more effort there. Bottom line is, I do not do restoration of VHS film masters for archive purposes, I deal with average Joe who want their tapes on DVD or in a file he could watch and be amazed that it is even possible and how good for him it looks like. Usually those are tapes that were shot and never played. And I obviously do not bother with tapes with worse signal, I just tell them it's the fault of the tape, I'm the budget option and it doesn't include fancy enchancement of every scan line, for that there are professionals with professional prices. I do not have my memories on analog tapes, so I do not really need to do that, the only analog tapes I have are from 2017 and 2018 and were filmed on S-VHS for a purpose, so that specific look was achieved, when I have finish working with them I will definitely show them somewhere. But since they are more or less fresh they will not have many signal problems. To each their own then I guess. |
Thanks for all the detail about your lack of interest and indiscriminate standards in video and graphics. True, the average joe misses out on a lot in this world, doesn't see or hear very well, and forums like this aren't for him. I don't need further convincing that most of what this forum has to offer is not directed at you, since there are a great many sources of information about hardware players and many of them go into far more detail about playback quality than you'll find here. I just don't understand why you're so defensive and insistent about it.
Unfortunately not many here are willing to lower their standards down to yours, so i guess you'll just have to put up with it. I think we're all smart enough to coexist. |
I didn't say that I lack interest, because this is not accurate.
I do what the Joe needs. Some of it is directed and I learn what I need. I have explained my side of the story, so as there wouldn't be some sort of misunderstanding that I might be an ungrateful bastard. By all means please don't lower the standards, I might find myself one day in the state where I'm ready to implement them or even contribute, rather than ask which gear to buy. And I cannot disagree with the last statement. |
There is no requirement to further enhance the videos, I'm all for basic restoration such as TBC, DNR (if needed), color correction, Anything more than that it comes with a price and the video start to become washed out. I'm not a professional however, I did some family videos few years ago using the Edirol VMC-1 and when I realized my mistake I did it again using a USB capture device to lossless, Still waiting the MP4 conversion though when I learn how to use those script programs.
So the bottom line advanced restoration is not for everyone, especially when scripting is involved. So if your clients cannot afford a restoration job and are happy with what you offer to them I don't see anything wrong with that. By the way there are none D-VHS PAL decks with HDMI output, But if you use the firewire output you will get MPEG-2 4:2:0 720x576 for PAL. Both DVHSTool and CapDVHS can be used for capturing. VLC is another option. |
1 Attachment(s)
Quote:
Bowen-Lichnovsky Wedding 1992 This is my sister's wedding from 1992. It was taped by a friend of theirs, using two cameras. The original has long since vanished and what you see here is at least a 3rd generation copy of the two tapes back to back in EP mode. To keep it somewhat relevant to the topic, the capture was performed by a D-VHS VCR...but in analog (S-Video) mode. Workflow was Mitsubishi HS-HD2000U>>TBC-4000>>ATI AIW X1800 (PCIe) on a Windows XP homebuilt machine; capturing lossless (HuffyUV) in VirtualDub. The colors here are typical of both cameras' tapes taken within the sanctuary (in the reception, with more light, colors were much better). Sis was pleased that I was able to save this tape at all when I first did it 2 years ago (for her 25th) with a Hauppauge HVR-2250 for capture and no TBC. Now that I have better equipment, I'd like to see if I can re-do it with better quality. Any suggestions for color correction and cleanup? |
Quote:
- some chroma offset - reduce pink - try to reduce harsh highlights, running into illegal values There is some faint interference pattern noise going on, not sure where it originates. Assuming not the capture setup, then there are Avisynth filters to try and attack it. That's one of harder errors to tackle. All things considered, not bad. I often see worse. This really isn't on-topic to this thread, should maybe be moved to its own. :wink2: |
Quote:
|
Possible, The manufacturer can decide what to do with a digital port, For instance some US models can input MPEG-2 and DV streams and output them as long as they are not copy protected, But none of them can output digital D-Theater cassettes and analog VHS cassettes with macrovision via firewire.
|
4 Attachment(s)
Quote:
And there are no "illegal values" in the sample. They disappeared a long time ago, likely when this out-of-spec video was recorded to RGB, which destroyed highlight detail and wiped out some dark chroma values. All that remain are hot-spots and featureless off-color blobs. Look at the YUV histogram and saturation vectorscope below: original - border removed, with YUV levels and vectorscope: http://www.digitalfaq.com/forum/atta...1&d=1546406792 (Above) Frame 148 of the original sample, lower border removed to avoid affecting the histograms. Users here spend lots of money and lavish heavy praise on high-end software, so they should be familiar with YUV levels histograms (left) and saturation vectorscopes (right). In the YUV histogram the upper white band shows no luma levels below y=16 and none above y=235, so no "illegal levels" remain in the video. There is no dark or bright data to be recovered. The dynamic range of the images cannot be extended or normalized -- the original data range no longer exists. The histogram's middle "U" band (yellow-blue) shows depressed, limited blue and a smattering of constricted yellow. The "V" band (green-red) shows over-extended red and a small amount of ragged green. The right-hand vectorscope shows pretty much the same thing as the histogram and has no indication of oversaturation or illegal levels. What it does show is that some yellow--orange, a lot of red, and some magenta (which contains a modicum of dark blue) is all anyone can find to work with. The right-hand quadrant of the vectorscope shows zero pixel values from middle green all the way around the edge of the color circle to dark blue. In Avisynth I tried decreasing red/magenta saturation a little (although neither is over saturated) and increased saturation of whatever green and blue might remain. I also used Avisynth's ColorYUV to increase the gain of the U channel and decrease gain of the V channel. The image below shows the results: original - modified http://www.digitalfaq.com/forum/atta...1&d=1546406833 Increasing blue\green saturation across the board shows that some noisy, corrupt blue data begins to appear. The scattered blue remnants are badly displaced both vertically and horizontally. There was no increase in green or yellow data. Want to reduce red? I did that with Color Finesse in AfterEffects, and below is what's left of the previous image after reducing red, wherein the image starts to approach gray: http://www.digitalfaq.com/forum/atta...1&d=1546406871 There's always the old monochrome treatment for video whose chroma has been irreparably ruined. But the genius who botched this video killed luma as well. With shadow detail crushed and highlights badly clipped forever, the resulting 4x3 monochrome with repaired borders looks almost like bad newsprint. Below, at least the b&w version lets you see more clearly the noisy mottling and garbage in the background. monochrome 4x3, borders repaired http://www.digitalfaq.com/forum/atta...1&d=1546406963 Throw the video into any NLE you want, big or small, "pro" or not, even if it's named after a famous Renaissance artist. I think you'll find that, lossless or not, the video was borked a long time ago. Except for removing a little noise, all one can do otherwise is make it look worse. And yes, I agree, it's a shame. |
Yep, this is a great example of doing more color correction just makes it worse. Been there, done/tried that. :laugh:
I'd mostly worry about chroma noise/offset, maybe attenuate the white to not be illegal values. I'm not sure if B&W is better or worse, just different. :hmm: It was probably screwed up when it was shot, not just from being nth gen (copies of copies). I doubt the master was good. |
Quote:
I am curious, though, about what you mean when you talk about this tape being "recorded to RGB"? I captured direct to HuffyUV from my copy of the tape and I thought I selected "direct stream copy" for the clip; did I accidentally slip RGB in there somewhere? But if you're referring to when the tape was made 27 years ago, my best understanding (I never met the man) was that it was not done by a professional videographer but by a "friend of a friend" with two VHS cameras. He set one up on a tripod in the balcony (not the angle pictured here) and hand-held the other. As far as I can tell he did no editing or processing at all, just popped the two tapes in a VCR and copied them back to back onto another tape in EP mode. I should probably keep my big mouth shut here in front of this audience, but I'm actually kind of proud of what I managed to do with this old tape two years ago with my very limited equipment. It's (much!) too big to upload here, but if you want to watch two hours and twenty minutes of someone else's wedding...have at. |
I didn't say you captured to RGB. I said that the effect of sharp cutoff of luma and chroma data at exactly y=16 and y=235 looked "as if" some device using RGB conversion or some other limiting device had caused severe clipping. Given the history you've recounted, there's no telling what caused it.
[EDIT] However, in looking at the MPEG music video posted earlier, I noted that the playback machine clips y=16, including black borders which are usually darker than that but are cutoff at y=16 in the music video. Those black side borders are SMPTE limits for 4:3 aspect ratio video in that frame size and should be zero black (i.e, super-black), not y=16. I'm suspecting that the player is doing it, unless the original tape has been thru some other processing. Except for a few special shots inserted during the performance, the tape itself has a limited dynamic range visually and relatively low contrast. Shadow detail looks rather murky, even for VHS. |
Quote:
I agree, cleaning up some color noise and tape grunge is the only workable area. |
Lordsmurf remarks referring to the wedding video sample:
Quote:
|
Quote:
|
Site design, images and content © 2002-2024 The Digital FAQ, www.digitalFAQ.com
Forum Software by vBulletin · Copyright © 2024 Jelsoft Enterprises Ltd.