The A/V Thread

Discussion in '-off topic-' started by Madmoiselle_Silk, Apr 17, 2018.

  1. sanctum -member- -member-

    sanctum
    Joined: 10 Mar 2005
    Posts: 1905
    Posted: Tue Sep 27, 2022 11:48 pm

    flowersofnight wrote:
    flowersofnight wrote:
    flowersofnight wrote:
    How many niches-within-niches of a special interest group is that?
     
  2. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Wed Sep 28, 2022 6:32 pm

    Only one really, just about every audio/video conversion tool out there is just a pretty wrapper around ffmpeg and/or lame .
     
  3. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Sun Nov 6, 2022 11:00 pm

    Today in Let's Be Aghast: the beautiful machine I purchased from Japan has a leprous flywheel and I'm not entirely sure how to tackle it.

    upload_2022-11-6_22-57-28.png

    What is that yellow crud? Why does it have lumps? ::kozi::
    I saw someone online making reference to very carefully sanding the flywheel while it's running, but that seems like it would very easily result in disaster. Maybe I'll use some hexane and see if anything happens to wash off with a good swabbing.

    EDIT: found a picture of one that looks much newer and unstained, but still warty:
    https://blog-imgs-128-origin.fc2.com/n/e/w/newprojectsc77/20190817114838d19.jpg
     
    Last edited: Nov 7, 2022
  4. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Tue Dec 20, 2022 11:54 pm

    Anyone heard of Plangent Processes?
    https://www.plangentprocesses.com

    It seems they have a proprietary way of eliminating distortion in tapes, as described here:
    https://www.plangentprocesses.com/_files/ugd/2aa449_58e0c825b6f74ebdbb65ecd1390c2067.pdf

    Essentially - all tapes have an ultrahigh-frequency "bias signal" running at a constant frequency. If you see that frequency varying, then you know the tape playback is a bit distorted, and you know how to correct it - just change the speed to make the frequency constant again.

    There's not much information about this publicly because it's mostly sold to big record companies working on old master tapes for famous acts like the Grateful Dead. Sounds like they charge by the minute for their services and supposedly it requires custom hardware - though I see a few mentions online of people just sending in digital files and leaving all the "magic" to their software.

    Last year there was a whole presentation from AES on the topic, including a demo of the hardware. I haven't watched the whole thing yet.
    https://www.aes-media.org/sections/pnw/pnwrecaps/2021/oct2021/index.htm

    Not that this is directly on-point to anything we do here, since they seem to deal entirely in reel-to-reel tapes, 1/4" or wider. Interesting concept though.

    I tried just capturing a tape at 192kHz to see if any bias signals show up, but got nothing. I think even if that would work you'd need a sound card with a higher frequency than that. You could only detect up to 96kHz signals that way and supposedly professional duplicators would go at least five times higher than the maximum frequency in the recording - so at last 100kHz for sure.
     
  5. cardy -member- -member-

    cardy
    Joined: 03 Nov 2004
    Posts: 337
    Posted: Sun Jan 1, 2023 6:18 pm

    Happy New Year y'all few scapers reading this!

    Looks pretty interesting. Too bad it's unavailable to most people.

    I've actually been reading up on the whole LD-decode / VHS-decode movement where instead of capturing the analog output of video via S-Video or Composite cables into a capture card, you instead tap into the RF signal of a device and capture that. You then decode the RF signal to produce a visua image.
    Supposedly it can get far superior results.

    https://github.com/oyvindln/vhs-decode

    I don't think it's something I'll ever attempt to do because of the sheer increase of processing involved per tape but it's interested to see if this will be a viable solution in the future.
     
    flowersofnight likes this.
  6. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Sun Jan 1, 2023 6:59 pm

    cardy wrote:
    Makes sense, that's like what people do on the Super Genesis Playsystem 64 to make HDMI adapters - just tap into the output pins of the video chip, essentially, and have a bit of FPGA circuitry to create an HDMI signal out of it.

    Much like with games, it does seem that this requires a sacrificial unit though, and I'm far too incompetent at soldering to try the mod myself. I guess it's a question of just how good a VCR you're willing to sacrifice for the experiment, given that good ones for even "mundane" purposes are expensive, scarce, and getting more so.

    Might be worth playing around with on a dollar-store garbage VCR just as a proof-of-concept, at any rate.

    And by the way, it amuses me how many decades of work the Domesday Project left in its wake, to work around that one catastrophic decision to use laserdiscs.
     
  7. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Mon Jan 2, 2023 1:59 am

    Tonight: Experiments With Azimuth

    upload_2023-1-2_0-49-28.png

    Here's a spectrograph of me playing back my (almost certainly bootleg) SANS LOGIQUE tape while playing with the azimuth dial like a laboratory ape. In an ordinary playback with the azimuth dead-center, that 15.7kHz line would be at a pretty much constant level. But on this side of this tape, you can actually increase the high-frequency response over dead-center by turning the dial about two "dots" to the right on my CR-70. The parts where it cuts out completely are where I turned the dial pretty much all the way to one side or the other.
    Curiously, unlike side A, side B is best with the azimuth centered.
    Since most of the audio signal is not very high-frequency at all, the azimuth probably doesn't matter too much, but I think that's the way to get the best out of this limited source material.

    My other deck (a DR-2) had a "ghost line" at precisely 21kHz, which this new deck doesn't. Not that you can hear that unless you're a bat, but it's interesting anyway.

    By the way, the "warty" flywheel doesn't seem to have mattered in the end. What mattered was getting a good belt (Vintage-Electronics was good, WebSpareParts was terrible) and getting the torque exactly on-point. In fact, I ended up calibrating the torque by playing a W&F tape and twiddling the torque adjuster till I got a minimum on the W&F meter. Once I did that, the torque meter was dead-on with no "jumpiness".

    Will be posting some audio files over in the downloads section in a bit, anyway. (EDIT: this is done)
     
    Last edited: Jan 2, 2023
    heutre likes this.
  8. cardy -member- -member-

    cardy
    Joined: 03 Nov 2004
    Posts: 337
    Posted: Mon Jan 2, 2023 9:11 am

    flowersofnight wrote:
    The good part of tapping the RF is that it apparently doesn't even require a good deck. Cheap ones can get the job done and even transfer SVHS tapes as all of that conversion of the signal happens in software later.

    It doesn't even need a TBC to stabilize the signals and that can be also done later on in software. I always thought the physical constraints of the tape medium introduced jitters and a bad signal that needed to be corrected in realtime. But I guess the TBC in itself is a hardware device which takes the signal and rebuilds it so that being a software step could make sense? I really want to send someone who has this setup a VHS I've already transferred to see an A-B comparison.

    Most of the comparisons I've seen do show that it could be good and you could eek out more quality but then the hardware / software knowledge is at such a high level that you're just parsing data which limits large scale capturing. I'm sure you can run batch scripts to do all the steps but those intermediate files are huge.

    flowersofnight wrote:
    It's quite fun to play with that azimuth knob, isn't it?

    I've come to expect that anything with less than a 10Khz signal will not really substantially change it's sound no matter how much you play with it. Degradation of the signal, whether from time, low quality tape or just simply being the way the source was, will never get those highs back.

    I have experienced what you have though, where one side A and B need differing Azimuth settings. I never quite understood how that could happen from a demo tape which was presumably recorded by the same person / deck.

    When the azimuth does work for me is with a well recorded tape, like an officially dubbed or high quality tape. Once you adjust that dial, you can see the frequency range start stretching out in the upper frequencies, and, most importantly, the treble (like voices and drum high hats) just sound a crisper and more bright.

    I'll be interested in hearing your transfer of Sans Logique. I always had the impression that it was never transferred properly and that hopefully we'll hear it crystal clear one day when done right.
     
  9. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Mon Jan 2, 2023 1:36 pm

    cardy wrote:
    They say that, but is it true? The quality of the video circuitry obviously doesn't matter since it's being bypassed, but doesn't the quality and condition (and number?) of the heads themselves matter still?

    cardy wrote:
    I wonder if it's because the bootlegger used an auto-reverse deck and just played back one of the sides in reverse. It's notoriously impossible to get both forward and reverse azimuth calibrated at the same time, even if you try.

    cardy wrote:
    Well, ever since talking to that guy who ran a small-time duplication plant I'm convinced that even the smallest runs of professionally-duplicated tapes will have been done on decks that were regularly calibrated, so they should be dead-center, assuming your own deck is calibrated. There will always be that potential bit of misalignment due to asymmetries in the tape shell etc. I suppose, but I'm not sure if that will be audible or correctable. I'll play around with it anyway, even though most tapes won't have such an easy "cheat" where there's a huge band of high-frequency noise to lock in on XD

    cardy wrote:
    It's here (including the azimuth done both ways on side A): https://www.scapeforums.com/index.php?threads/malice-mizer-demo-re-rips.6161/#post-340136

    I think that's about as good as can be done with the tape I have, but it is still low-quality material. It would be interesting to see what a genuine original sounds like, but I suspect the fantasy version we're all hoping for was only ever on the masters - if that.
     
  10. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Thu Jan 5, 2023 2:01 am

    Trash To Treasure

    I was experimenting with some of my most obviously fraudulent tapes and came across a pleasant surprise!
    This is Kneuklid Romance's first demo tape. The real thing was limited to 10 (!) copies and evidently my copy is not the real deal. If you look at a spectrogram you see a rather poor response:

    upload_2023-1-5_1-35-38.png

    However... if you turn the azimuth all the way to the left, those "ghostly" high frequencies come back into focus and we get a much more listenable track:

    upload_2023-1-5_1-36-41.png

    By comparison, the history DVD has the same track from this same performance, and there we see an extremely aggressive restoration:

    upload_2023-1-5_1-38-54.png

    The DVD version seems like it was compressed and/or had the frequency emphasis changed to make the vocals pop more - not sure. I have no idea what they were going for with the mastering here - every single track is like this, with a hard cutoff at 16.2kHz or so, but also high levels of sound right up to that limit, resulting in that "square" look. It doesn't seem to be an attempt to cut out 15.7kHz noise since some of the tracks have a very prominent line still present at that frequency.

    Incidentally, there's no telling what the correct azimuth really is for my bootleg, since I took it all the way to the leftmost limit on the CR-70 without seeing any falloff. Maybe it could even be improved further by someone willing to intentionally decalibrate their deck to move the azimuth even further left... but that someone is not going to be me XD
    I wonder if it was made on a Nakamichi ZX-7 or ZX-9 with manual record azimuth setting, and the guy just set it as wrong as it could possibly be.

    Anyhow, I'll post the MP3 for this (with a bit of noise reduction applied) in the downloads section.
     
    heutre likes this.
  11. cardy -member- -member-

    cardy
    Joined: 03 Nov 2004
    Posts: 337
    Posted: Wed Jan 25, 2023 4:01 am

    flowersofnight wrote:
    That's my thinking too. I just was recently back in New York, where I have all of my VCR gear...and I mainly transfer tapes on a SVHS JVC player which was prosumer grade back then. Out of curiosity I transferred a few tapes with a junky consumer VCR and while I inspected it on screen and it looked sharper what I found out eventually when I really watched the clips was that there was more screen tearing at the top, more chroma bleeding (the reds specifically) and also way too much oversharpening. In the end, my trusty JVC which I thought was producing softer results was much better. I'm still trying to figure out how RF capturing can still overcome problems introduced by the physical playback of any given deck.

    flowersofnight wrote:
    Agreed, I think it must've been so easy for bands to duplicate two sides of a tape with an auto-reverse deck.

    flowersofnight wrote:
    [/QUOTE]

    These rips sound great. I've made some comments over on that thread~
     
  12. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Fri Jan 27, 2023 10:27 am

    On the topic of "Does azimuth shifting raise the noise floor": not usually but I did come across one notable exception so far.

    See the Kneuklid Romance spectrograms above for an example of turning the azimuth all the way to one side without increasing the overall noise. But on the other hand... check out what happens on this copy of "Lie" by Cains:Feel ::gaku::

    Centered:
    lie_center.png

    Maximum response found at 2 dots to the left:
    lie_left.png

    The band of background noise barely visible in the centered spectrogram jumps up significantly in the 2-left one. Whatever frequency response improvement there is in the music (if any) gets swamped by the noise.
    I don't know what the reason for this is.
     
  13. cardy -member- -member-

    cardy
    Joined: 03 Nov 2004
    Posts: 337
    Posted: Fri Jan 27, 2023 6:22 pm

    I just mentioned in the other thread that I find the noise floor becomes higher when adjusting the azimuth on tapes with a smaller frequency spectrum 6-10khz which has lost it's high frequencies.
    For tapes with a good 10-15khz range, the noise floor changes are less noticeable or imperceptible.

    The results you have for the Kneuklid Romance song is something I don't feel like I'd be able to achieve with my deck. In general I never get such noticeable results like that when I adjust the azimuth.
    Not sure if it's the limitations of my deck or the tapes I'm playing.
    I'm always tempted to send one of my harder to transfer tapes to a more professional operation who know what they're doing to create a benchmark or reference to compare against.

    This Cains:Feel example looks extreme and I'd probably transfer it two different ways, once centered and once off, just to figure out later which one has the better sound.
     
  14. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Sat Jan 28, 2023 7:56 am

    cardy wrote:
    Well, I may only be posting the most bizarre and unusual results here XD
    That Kneuklid tape is a real outlier, you'd almost have to be trying to get that far off-center. It starts out the way most tapes end up if you go too far with the azimuth. I'm sure you've seen the way it "drops off" past a certain point on regular tapes.
    Sometimes on YJ you see decks for sale where in the photos the azimuth knob is turned all the way up. I wonder how many people are just leaving it "maxed out" at all times as if it was a volume knob.

    More commonly, the results of adjusting is just that you see a few more wispy "clouds" in the high-frequency regions, and a bit brighter 15K line if there is one.

    As for the Cains:Feel tape I think I'm just going with the centered version, I'm not convinced there's any new signal to be found among the noise after adjusting.

    cardy wrote:
    Hard in what sense?
     
  15. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Tue Jan 31, 2023 12:43 pm

    Here's a more typical "good improvement" case, from an oxbxjxe demo:

    Centered:
    obje_center.png
    Adjusted +2 right:
    obje_right.png

    The 15kHz line continues to be a mystery. It does show up even on some tapes that are indisputably authentic, like studio submission tapes.
     
  16. cardy -member- -member-

    cardy
    Joined: 03 Nov 2004
    Posts: 337
    Posted: Wed Feb 1, 2023 6:25 pm

    flowersofnight wrote:
    Seeing your results is great though and I'm enjoying reading about it.
    I wish there were more people out there who I can 'look over the shoulder' and see what they were doing to get their vinyl / cassette captures right. Even with lineages that often come from an analog transfer, there's all of those in between steps in how people use the tools which sometimes makes the difference.

    One great ripper on redacted mentions that they completely fast forward and rewind a tape before transferring in order to make the tape reel taut and more consistent when playing back. Tips like that are great to know about.

    What I meant about certain tapes being hard / difficult was that they sometimes already have a low frequency range to start, whether that was the case when they were new or deterioration happened over time, and twiddling the azimuth doesn't really change the sound appreciably.

    flowersofnight wrote:
    I think I've mentioned this before but I'm seeing that line in many official VHS tapes I've transferred and many other bootleg lives when I bother to look at the spectrals.
    There probably a lot of CRT's in the same room - probably computer monitors?
     
  17. flowersofnight -moderator- -moderator-

    flowersofnight
    Joined: 04 Aug 2004
    Posts: 13568
    Location: Vintage Live House, 1994
    Posted: Thu Feb 2, 2023 9:41 am

    cardy wrote:
    It makes sense I think. When I was calibrating W&F it was driving me round the bend how things would change if I had or hadn't just rewound, and I didn't know why. Maybe I'll try that.

    cardy wrote:
    I wonder about that - particularly when the spectrum lines are thick and "smeary". You'd think CRTs would be extremely close in frequency at all time because otherwise the picture would get distorted. (And I presume we can rule out the Doppler effect here ::meev::) A line that is still thin but very bright would be what I'd expect recording in a room full of TVs.
    You know what I bet would be a good place (perhaps the only remaining place) to test this? An arcade where they still have a lot of old-time boomer games with CRTs.
    Besides which, I'm not sure any of these no-name indies bands would be recording in a room full of computers in like 1994 XD If you look at some "Make Your Demo Tape" books from around then, they don't even mention computers at all. More likely it was like the stereotypical setup on TV where it's just some scraggly punks with guitars and one guy outside chain-smoking while fiddling with the sliders on the mixer XD

    I've actually seen a couple of cases where the line is not at 15.7kHz but actually something higher like 16.2. I wonder if that's "real" or just a regular 15.7 signal with change in tape speed. I was thinking maybe a CRT HDTV but that would be way into ultrasonics.