vhs-decode setup? A Basic Guide

[latreche34]

Quote:

Originally Posted by lordsmurf

There will be a potential % sharpness boost, but that's really it. You can cherry pick errors, benefits, etc, but it's all really fungible. You're still reliant on extant VCR heads, so the main

Not even sharpness, As I said above they use composite only VCRs to compare samples, Some are really cheap skate nerds that don't want to buy a decent VCR, Some are flat out deceiving like the member over at videohelp that got called out by members after finding out he is over processing the composite capture and using completely two different VCRs, Talking about comparing apples to oranges. And when he was challenged to do a real comparison by passing the same tape around to volunteer members to do each own sample he opted out.

[lordsmurf]

Quote:

Originally Posted by latreche34

Not even sharpness, As I said above they use composite only VCRs to compare samples, Some are really cheap skate nerds that don't want to buy a decent VCR, Some are flat out deceiving like the member over at videohelp that got called out by members after finding out he is over processing the composite capture and using completely two different VCRs, Talking about comparing apples to oranges.

When I state "% sharpness", I mean tiny nuances, the sort of thing videophiles (or audiophiles) drool over. Normal people don't care, including serious hobbyists and professionals. It's not game changing. You can easily falsely sharpen to match, but sharpening cannot be boosted before upscale. So the main beneficiary is upscalers, but only if source and upscaler is quality as well.

I saw that VH comparison. More like apples to a suspension bridge (TBBT reference). Complete BS, and then he doubled/tripled/+ down on it was called out.

I honestly don't think VHS tapes merit a "decode" at all. The focus should be on formats that lack quality extraction, such as Beta, U-matic, etc. I forget all of the speciality formats that were abandoned, where tapes exists, gear does not.

[i4004]

Quote:

Originally Posted by latreche34

Not even sharpness, As I said above they use composite only VCRs to compare samples, Some are really cheap skate nerds that don't want to buy a decent VCR, Some are flat out deceiving like the member over at videohelp that got called out by members after finding out he is over processing the composite capture and using completely two different VCRs, Talking about comparing apples to oranges. And when he was challenged to do a real comparison by passing the same tape around to volunteer members to do each own sample he opted out.

it's not even about composite-only vcrs, i think vhs-decode can be sharper only if it's doing some sharpening on it's own, simillar to how vcrs will do sharpening (or "peaking" as it's called in vcr electronics terminology) on their own. other ways of sharpening vhs are
a) better video heads
b) better heads preamps
c) better video processors inside vcr (the parts that convert RF signal from heads to composite or s-video output from vcr)

better vcrs usually have decent components on all 3 stages (are there any vhs vcrs that have crappy heads but good preamps and video processors? or any other ridiculous good/bad combination of the 3?), so any improvements with 'vhs-decode' probably boil down to better drop-out compensation. maybe. if vcr you're using is bad at drop-out compensation. many are.

i wouldn't expect 'vhs-decode' to magically improve my 35 year old vhs machine's sharpness, nor would i expect it to improve upon sony slv-e730, because that's already razor sharp.
i don't have tbc issues on my tapes so i'm not even discussing that. mostly PAL OTA recordings, looking better than any pre-recorded vhs i ever saw. "pre-recorded" meaning made at commercial duplicating facility on whatever gear they used. that always looks like 2nd generation, because it is 2nd generation (1st generation is master, its copy is 2nd generation).
that's probably why everybody thinks "vhs is crap", their vhs is copy (back in the beginnings it was bootlegging, never even a 2nd generation, 3rd and above were standard, before "proper" ie legal video stores opened), mine is OTA, they never saw decent vhs recording

just had a lil talk here
https://www.youtube.com/watch?v=mgIy...fFsA-yjF6brHSn
(that link should lead to comments thread under that video)
i'll copy/paste my reply here anyway, because yt replies system is sketchy....that link probably won't be working soon enough....

Quote:

@ivok9846
prije 23 minute
​ @TheRealHarrypm panasonic nv-hv60 has hifi fm test points, but it's already too new to be any good, if panasonics were ever really good. it doesn't have separate head preamp at all. as for head condition, i think it's doesn't make much difference, knowing what i did to my oldest vcr (35 years old, still works) and how it kept going over the years. i mean....better not to discuss it.....if you were to find unused and machine with 1000hrs i doubt differences would be great. or even perceptible. i dunno if it's even possible to find worn heads vhs machine, the tape is pretty soft. maybe on machines that were working full working hours, ie 8h every day, but not on home machines.

it would be interesting if we could swap a vhs tape, you capturing it via RF and me via my old system, and then compare the two, but...like i said, sharpness via slv-e730 and philips saa713x based capture card is not an issue at all. i kinda doubt i would get better results with sanyo vhr5100ee from 1989 and vhs-decode. why?
because sanyo must have worse heads of the two. even when it was new. and worse RF pre-amp. not deteriorated, just worse.....

regarding virtual tbc (i presume vhs-decode is doing): yes, digital system can align scanlines beginning, but it still can't align perfectly the complete line, because of magnetic tape non-linearities and stretching, ie every line has some horizontal wiggling going on, this will be visible when recording signal that has vertical lines in it, for example subtitles etc.
i'm just watching "vhs-decode vs 3 others: TBC stress test (Crafters)" on yt (and i see your comment), and it seems it's not even doing aligning of scanline start i just decribed above. maybe the recent version has tbc.
(must say, i like the amount of taskbar icons in your "VHS Decode, ld-analyse & tbc-video-export A Basic User Guide - Windows (2024)" guide.)

i mean i don't mind the command line (because i was using and still use avisynth..that would be father and mother of QTGMC...heh) but i don't have tapes with tbc issues....

gosh, those are high data rates, as i expected......280 or 560mbps....

it seems you're including some sort of drop-out compensation, we were doing that via avs filters.....in 'doom9 analog capture guide', i wrote the chapter '13 Removal of clicks and scratches' (linked in '7.2.3 Removal of clicks and scratches'. in 7.2 Processing the video using AviSynth chapter).
that's a good thing, i had some kodak tapes with plenty of drop-outs.

in your "Visual Comparisons" on wiki, all "conventional methods" pics look like some ridiculously crappy usb-mpeg2 capture device that's not even preserving interlacing, so image is vertically polished, blurred, ie it's not even full resolution, those camera date/time captions reveal that. most of my captures look like those bottom pictures. the "Lorrane Wedding Tape" looks like it was captured with contrast setting during capture cranked all the way to the max, so everything went black. better to do less contrast than more during capping, because more means unrecoverable erros like these.

regarding linear(mono) audio, it gets a bit complicated, right? i mean you need external clock hooked up to 2 cx cards, right?
why didn't you solve that in software too? usually capture utilities(vdub) can resample audio to match video clock, or vice-versa.

ie the actual capture utility that puts both rf (video) data and audio from pc sound card into one file you later process....

the comparison part would be fun, if i was in UK as most vhs-decode enthusiasts, but i'm not, and in the matter of their comparison (as i wrote above in my reply on yt thread), much like smurf i don't see it as a real comparison (why not 720x576 .png images from both?).
https://github.com/oyvindln/vhs-deco...al-Comparisons

maybe i'll send him a tape i'm just about to capture, so we can compare it later....

i mean stuff they're doing is not totally crazy, it makes some sense, but it's a bit convoluted....it's not easy.
if one had a lots of tapes, and tapes with lots of drop-outs, and if money could be made on it, it would be interesting project.

LS:

Quote:

When I state "% sharpness", I mean tiny nuances, the sort of thing videophiles (or audiophiles) drool over. Normal people don't care, including serious hobbyists and professionals. It's not game changing. You can easily falsely sharpen to match

i drool over sharpness, i really do. like completely and utterly non-normal people would. i was known to capture same footage, off of different vcrs, than compare same frames. exactly the same frames. and check exactly sharpness. with magnifying the details of the frame. and it was easy, and it was fun.
while it's not easy to "falsely sharpen to match" if you ask me. first of all you need to know how to sharpen (for example limitedsharpenfaster) and then you might find out there's no detail to sharpen in those worse vcr samples. while better machine needs no sharpening.

why do i do this? because i won't be keeping huffyuv OR mjpeg sources, it will be mpeg2 in the end. a rather decent bitrate, sure, but still lossy. so i need every percent of sharpness i can get.
funnily enough, i don't mind noise (these days we don't get noise in video anyway, all we get is blurry digital transcode...the damn thing deteriorates faster than vhs copies!!! 50% of yt video is garbage) so it kinda reminds me of noisy times, and also, noise actually adds some sharpness, that process is called dithering, applicable to both images and video. this is why when you have crappy, blurry and blocky digital video adding noise via ffdshow on playback actually helps. it literally breaks the blocks.

[lordsmurf]

Lately I seem to be spending more time discussing vhs-decode than it deserves.

Quote:

Originally Posted by i4004

it's not even about composite-only vcrs, i think vhs-decode can be sharper only if it's doing some sharpening on it's own, simillar to how vcrs will do sharpening (or "peaking" as it's called in vcr electronics terminology) on their own.

Correct.

Quote:

other ways of sharpening vhs are
a) better video heads
b) better heads preamps
c) better video processors inside vcr (the parts that convert RF signal from heads to composite or s-video output from vcr)
better vcrs usually have decent components on all 3 stages

Correct.

Quote:

(are there any vhs vcrs that have crappy heads but good preamps and video processors? or any other ridiculous good/bad combination of the 3?),

At this late date, yes. Many otherwise-good models/decks have shot heads. The unit was likely fine when new, but it hasn't been new in 20-30 years.

Quote:

so any improvements with 'vhs-decode' probably boil down to better drop-out compensation.

Nope, not even that. Dropout is about heads.

Quote:

maybe. if vcr you're using is bad at drop-out compensation. many are.

I would suggest most are, which is why lower-end consumer fodder decks are not suggested.

i wouldn't expect 'vhs-decode' to magically improve my 35 year old vhs machine's sharpness, nor would i expect it to improve upon sony slv-e730, because that's already razor sharp.

Quote:

i don't have tbc issues on my tapes so i'm not even discussing that. mostly PAL OTA recordings,

Uh, hmmm... I shall not discuss this here either. I disagree vehemently, but we can simply table that side discussion here. All I'll say is this: PAL timing errors present different from NTSC, so sometimes that color opinions on errors "not being present", as most discussion online about timing/TBCs are very NTSC-centric..

Quote:

looking better than any pre-recorded vhs i ever saw. "pre-recorded" meaning made at commercial duplicating facility on whatever gear they used.

Most commercial tapes, contrary to common assumption, are not recorded. Those are "duplicated", generally using contact duplication methods. Nothing is recorded, it's not real-time, VCRs were not used. That only existed for low-budget stuff, like the local wedding videographer/studio running off extra copies for your family members. Contact duping was an overall cheaper process to any sort of volume, even if you just consider the man hours.

Quote:

that's probably why everybody thinks "vhs is crap",

It's mostly due to
- cheap VCRs,
- cheap TVs (including crappy HDTVs),
- bad tapes (reused, rentals, no-name blanks, etc0,
- simply misremembering what videotapes looked like decades ago

Quote:

i'll copy/paste my reply here anyway, because yt replies system is sketchy....

It's actually not "the system" at time, but so,e Youtubers are wusses/snowflakes that do not want any dissent or discussion, aside from fellating praise, on "their" videos. That said, Colin/video99 isn't one of those to my knowledge.

Quote:

i drool over sharpness, i really do. like completely and utterly non-normal people would. i was known to capture same footage, off of different vcrs, than compare same frames. exactly the same frames. and
check exactly sharpness. with magnifying the details of the frame. and it was easy, and it was fun.

My discussion about vhs-decode potentially "having more sharpness" is taken fully out of context -- to the point where I just need to go back, find those old posts, and edited them for clarification.

I'm often pressed for time when posting online, so I tend to be briefer that I'd like. In this instance, brevity led to misinterpretation.

- Let's say you have 100 videos.
- 90 of them transferred perfectly with no extra effort. That's actually statistically common, when using a quality workflow. (ie, JVC/Panasonic S-VHS VCR with line TBC > DataVideo/Cypress type frame TBC > quality capture card/device.)
- Of those 10, you can cajole 7 of them to give results, usually by swapping the VCR, or manually changing some settings.
- Of the 3 left, you get what you get.

Now enter vhs-decode.

- Of those 3, let's say that the issue at hand was severe ghosting, and the video had a lot of text. Perhaps a VHS tape made by the school AV dept, of a HS play, with on-screen captions. The processing within VCRs can sometimes get confused, as most have a dual mandates of maintaining a sharp image, whily also reducing slop. When the signal is crap, it makes internal realtime processing decisions hard. I have seen some very cherry-picked abnormal instances where vhs-decode is able to make an image look "sharper" (not really) due to specific noise handling. So in this case, less ghosting around text. Not some sort of perfecting removal/de-ghosing, but just different (better in this case) suppression. The irony here is I've seen many more samples of vhs-decode, where is has actually added similar problems to tapes that were fine even in low-end VCRs.

But here's the real kick in the teeth:

- Those 90 tapes that were "transferred perfectly" would not be as good using vhs-decode, due to how primitive the decoding software is (and may always be). Some may be "the same quality" (arguable), but some will be obviously worse. So overall, lesser quality.
- Same for the 7 that were eventually trasferred fine.

It is a different system entirely, and it's not as smart as hardware from decades ago. Different, not better. And let's be honest here: would you expect some weekend hobbyists to outperform what was accomplished by major corporations with R&D engineers, even if from decades past? And VCRs got heavy R&D, unlimited budgets, for decades.

Quote:

while it's not easy to "falsely sharpen to match" if you ask me.

There are so many Avisynth sharpeners out there these days, that it's often not hard to match or near-match sharpness profiles from other VCRs/gear. Some of what we saw in the late 90s and early 2000s was the pinnacle of the format.

Quote:

why do i do this? because i won't be keeping huffyuv OR mjpeg sources, it will be mpeg2 in the end. a rather decent bitrate, sure, but still lossy. so i need every percent of sharpness i can get.
funnily enough, i don't mind noise

MPEG can be near-lossless in the way that ProRes422 or DNxHD is. I keep some Huffyuv for super important videos, but I also save a lot (of my own videos) down to 4:2:2 MPEG-2 at broadcast bitrates. Though not so much anymore, as we can now buy a 22TB HDD for like $350, and I'm not that cheap. Spending all day re-encoding to save a few dozen GB is just idiotic. File archiving, storage, depends on factors. Just be pragmatic about it.

Quote:

50% of yt video is garbage)
. this is why when you have crappy, blurry
(these days we don't get noise in video anyway, all we get is blurry digital transcode...the damn thing deteriorates faster than vhs copies!!!

It's a feature, not a bug. DVD-Video and DV artifacts were blocks. Rather than degrade to blocks, H.264 and beyond decided to instead turn video to mush. "Look kids, no blocks! Enjoy your oatmeal mush."

[i4004]

Quote:

At this late date, yes. Many otherwise-good models/decks have shot heads. The unit was likely fine when new, but it hasn't been new in 20-30 years.

yeah, you could say that, but i was thinking about (then) new models, ie low, middle and hi range/class vcrs were separated....
that distinction was erased later on, when they just became....cheap....number of heads didn't mean much at that time....
ie after 2000 you can easily have 6 head vcr that looks same or worse than 2 head machine a decade earlier.

Quote:

Nope, not even that. Dropout is about heads.

i have some kodak tapes that claim otherwise....worn heads have different symptoms....clogged heads too.
i mean even visual inspection of tape is reletively good parameter of number of drop-outs you'll get, ie if you can see small dots on tape, it'll have drop-outs on playback.

Quote:

Uh, hmmm... I shall not discuss this here either. I disagree vehemently, but we can simply table that side discussion here. All I'll say is this: PAL timing errors present different from NTSC, so sometimes that color opinions on errors "not being present", as most discussion online about timing/TBCs are very NTSC-centric..

"not being present" in a sense that i don't have much more than few percent of horizontal jitter and no "flag-waving" or vertical roll. ie no obvious flaws, just the usual vhs imprecision.
i don't think pal or ntsc is too important, as i have seen equally decent ntsc recordings, and they were equally OTA. ie my main point is ota offers higher quality.

here's ntsc that's obviously ota:
https://www.youtube.com/watch?v=7uZPtqx4D68&t=1598s
although the guy digitizing it didn't bring the color back to its place, but other than that, it's ok quality.
something "oddity archive" will never achieve.
in one part because he probably never rocorded ota.

i did some (ntsc) tape in pal60 mode, no issues with timebase instabilities either.

Quote:

Most commercial tapes, contrary to common assumption, are not recorded. Those are "duplicated", generally using contact duplication methods. Nothing is recorded, it's not real-time, VCRs were not used. That only existed for low-budget stuff, like the local wedding videographer/studio running off extra copies for your family members. Contact duping was an overall cheaper process to any sort of volume, even if you just consider the man hours.

that depends on what duplicating company actually has. i never heard about contact duplication untill recently, but i did hear and saw plenty of images of commercial duplication facilities with tons of vcrs.
so if you had more of these than contact duplicators, which is a possibility (esp. given that only otari and sony offered these devices?), then you have more VHS that was recorded on actual machines than via contact duplication.
it would be interesting test to make (contact vs. video copying recording, frame-level comparisons), but...like i said, never saw really good looking commercial tape. it just looks like it's not "master". ie it looks like vhs->vhs copy, not like some better format->vhs.
dunno about american market...i didn't collect ntsc tapes.
to that extent pal does have 100 lines of resolution more, and no color-hue issues, again for ota recordings.
(dunno if ntsc had color hue issues over satellite? probably not)

Quote:

It's mostly due to
- cheap VCRs,
- cheap TVs (including crappy HDTVs),
- bad tapes (reused, rentals, no-name blanks, etc0,
- simply misremembering what videotapes looked like decades ago

tvs were pretty well equalized when it comes to image quality, offcourse some were bigger (but not as big as in usa, for example not much rear-projection tvs in europe). vcrs were better in 90s than what came later. ie all of them were better, in average.
but what almost nobody did was record with vcr's tuner. that was too complicated, so vcrs were mosty used as vhs-players. i believe that to be the case in usa too.
that and timer recordings.....geez, that was only for rocket-scientists! hehe.....

Quote:

It's actually not "the system" at time, but so,e Youtubers are wusses/snowflakes that do not want any dissent or discussion, aside from fellating praise, on "their" videos. That said, Colin/video99 isn't one of those to my knowledge.

yeah, some "creators" are on ego-trips, which reminds of some admins on some forums (for example it was nice at ars-technica forums, until you wanna talk (non-leftist) politics...huh, but also on some purely technical forums...mindless censorship..oh yeah, also on doom9 back in the day...i oncecorrected doom9 so i was exiled....hehe), but i was more thinking about the "system" itself, where changing one word is all it takes for post to be published, even though word you used originally is quite fine too. not swearing or whatever....
so there are two levels of censorship on yt, automatic "spelling detection"..heh....and then there are creators banning anyone for whatever reason.
i just open the link to post (the time when it was posted is a link to that particular post) in browser incognito window to check if it's visible to everyone.
but the mere fact one needs to check that is a PITA.

Quote:

- Of those 3, let's say that the issue at hand was severe ghosting, and the video had a lot of text. Perhaps a VHS tape made by the school AV dept, of a HS play, with on-screen captions. The processing within VCRs can sometimes get confused, as most have a dual mandates of maintaining a sharp image, whily also reducing slop. When the signal is crap, it makes internal realtime processing decisions hard. I have seen some very cherry-picked abnormal instances where vhs-decode is able to make an image look "sharper" (not really) due to specific noise handling. So in this case, less ghosting around text. Not some sort of perfecting removal/de-ghosing, but just different (better in this case) suppression. The irony here is I've seen many more samples of vhs-decode, where is has actually added similar problems to tapes that were fine even in low-end VCRs.

vhs-decode not as a processor, but just digitizer, ie something like this
https://www.youtube.com/watch?v=BPwiVVWj2Ls

or
https://www.youtube.com/watch?v=CQ-u979_6Cw
where everything except vhs-decode destroys the sharpness by doing too much sharpening (grass has more detail on vhs-decode). and vhs machine everything was taken from is low end panasonic from mid 90-s (i had simillar model, already too much plastic at wrong places)
and oh yeah, that diamond device is crap that is almost blind to sync pulses, there are plenty of other yt clips that prove it.

vhs vcrs doing oversharpening is also a big theme in itself. in that aspect vhs-decode should help because normally sharpening will be done after heads pre-amp in vcr, so vhs-decode shouldn't do any of that, dunno about shenanigans some do to advertise vhs-decode....

as for titles, in this particular place in europe foreign stuff is subtitled, so i know a thing or two about what sorts of processing makes subs look bad. i also know ghosting from reflections in aerial. the latter affected me more so i've found some avisynth filters to ameliorate it....

Quote:

It is a different system entirely, and it's not as smart as hardware from decades ago. Different, not better. And let's be honest here: would you expect some weekend hobbyists to outperform what was accomplished by major corporations with R&D engineers, even if from decades past? And VCRs got heavy R&D, unlimited budgets, for decades.

well.....frankly....i would!
because, fundamentally it's software vs. hardware battle, where time is not on hardware side, not at all.
vcrs couldn't have motion-compensated denoising, because that thing didn't exist then, it's a purely digital process, can't be done on lines of image (like vcrs can do), but on 8x8 or 16x16 blocks. vcrs never had enough memory to do that.
and, frankly, most of pre-motion-compensated dnr on avisynth was crap. AND if you have lots of noise, even mc dnr will blur.
but i'm pretty sure that hardware dnr doesn't work at all if you ask me. too much artefacts and/or blur. blur is my biggest enemy. when in doubt, i just crank the mpeg2 bitrate and preserve the noise too.

that reminds me of pioneer kuro plasmas, praised all over, and then you take a look at sd image they produce.
and how are they supposed to look when it's the first generation of plasmas, again the time not on the side of hardware.
offcourse it doesn't have to do with plasma or lcd panel type, but the mere fact video-processors were bad at that time.
they couldn't deinterlace and upscale well. they could, later....and offcourse they (modern tvs) STILL can't do noise reduction, but the difference is some at least can defer dnr fully, while many only decrease it, never comepletely can turn it off.

many things made possible by avisynth or vapoursynth couldn't be possible in analog domain. at all.
but the aim of vhs-decode shouldn't be to process, minus aligning lines (for tbc effect) a bit and have better drop-out compensation (which shouldn't be hard to do in digital domain: much harder (ie more expensive) in analog), but just to "rip" the video signal as-is.

but, as you say, the main problem is "what if vhs-decode is actually not better than ok vcr and ok digitizer?"
(afaik doesn't have to be high-end vcr)
i would like to see a test where particular tape is done both ways and then compared.

ie the issue is "what exactly are we trying to fix with vhs-decode?". or "what do i get for all that complication?"
and there is something rather unnerving about (mono) audio path. something it's hard to "decode" from their wiki page.
https://gitlab.com/wolfre/vhs-decode-auto-audio-align (that's not vhs-decode wiki, this is another project!)
"oh, gosh, do i need 3 capture devices to capture video, hi-fi and linear audio?"

being a sharpness-freak, i can't ask for more than i already have and i don't expect vhs-decode to beat it.

maybe the -decode project is better suited for things like this:
https://www.youtube.com/watch?v=TAdFHwGlAQQ
because that content is worth saving, if there's not better hd version...heh..i think i have one or two episodes on vhs....offcourse it was recorded "off the air." ...

Quote:

There are so many Avisynth sharpeners out there these days, that it's often not hard to match or near-match sharpness profiles from other VCRs/gear. Some of what we saw in the late 90s and early 2000s was the pinnacle of the format.

along the lines of what i said about hardware dnr above, better question would be "would high end vcrs be able to match limitedsharpenfaster?".
because, as in your example with school graphics above, too much (of bad) sharpening can create halos.

Quote:

Spending all day re-encoding to save a few dozen GB is just idiotic.

it's also about ease of playback, everything plays mpeg, what about huff?
i mean sure, you can play it on pc with qtgmc on-the-fly with newer pcs, but not on tvs....
some of which, as you'll remember, do have decent processing and you can turn off dnr completely.
AND cce is fast. VERY fast. it's almost as if someone made mpeg2 encoder in assembler.

Quote:

It's a feature, not a bug. DVD-Video and DV artifacts were blocks. Rather than degrade to blocks, H.264 and beyond decided to instead turn video to mush. "Look kids, no blocks! Enjoy your oatmeal mush."

you know, going down that path one might end up in all sorts of weird thoughts, one of which might be:

how many people today are actually watching worse video image quality today than in analog days?
(a blasphemy!)

1-their (cheap chinese) tvs have crappy video processors, with totally wrong settings, for example crappy dnr applied to overcompressed video, so as soon as something starts to move, you get abstract art on screen, nothing is visible
2-most content will be overcompressed to boot, they lack bandwidth.....so much so that transition to 4k might be longest in hostory of formats, if it happens at all.....there are almost no 4k channels on european satellites....

this sacrilige was complete truth some time back when transition to hd was still not done (europe probably has a 10 year lag after usa in that aspect, still many, MANY sd channels are transmitted on satellite, for some reason), today a bit less so because its mostly hd, at least on terrestrial. on satellite sd channels are still eating into the bandwidth of hd.
beyond that there's iptv, again overcompressed.....hehe....
and h265 didn't help much.

[lordsmurf]

Quote:

Originally Posted by i4004

yeah, you could say that, but i was thinking about (then) new models, ie low, middle and hi range/class vcrs were separated....

Almost all decks from the 2000s/10s are low-end Funai, used by multiple brands. It was literally the only manufacturer left.

Quote:

i have some kodak tapes

More awful stuff. A few leaked into my collection as well. Kodak made nothing somewhere in the late 80s to early 90s, all were rebadged cheap tapes. Those had dropouts before even being used. See also Fuji Pro (and the "Pro" part is laughable).

Quote:

i don't think pal or ntsc is too important,

It really does amuse me how the old myth of "PAL is better than NTSC" finally went away in the 2000s, once people saw how equally bad/wrong both were at values, timing, etc.

Quote:

but i did hear and saw plenty of images of commercial duplication facilities with tons of vcrs.

Hear/see where? Commercial duplication, where decks were daisy chained together, were mostly for small-run productions. It simply was not as economical as a contact dupe with scale. But I have to add that most European countries were "small scale", so I can see a higher concentration of manual duping in PAL lands.

Quote:

tvs were pretty well equalized when it comes to image quality,

Nah. CRT era, HDTV/LCD era, no difference. It's just as variable now as it was back then.

Quote:

vcrs were better in 90s than what came later.

Or before (70s/80s = blah).
The best decks came in the 90s, but the early/mid/late really depends on manufacturer. An exception to the rule is JVC S-VHS. which had some of the best decks in the early 00s as well (but also some really lousy ones; model/line matters).

Quote:

vhs-decode not as a processor, but just digitizer, ie something like this
https://www.youtube.com/watch?v=BPwiVVWj2Ls
or
https://www.youtube.com/watch?v=CQ-u979_6Cw
where everything except vhs-decode destroys the sharpness by doing too much sharpening (grass has more detail on vhs-decode). and vhs machine everything was taken from is low end panasonic from mid 90-s (i had simillar model, already too much plastic at wrong places)
and oh yeah, that diamond device is crap that is almost blind to sync pulses, there are plenty of other yt clips that prove it.

I hate videos like that. The first is not impressive whatsoever, and the second is extremely misleading (as there's zero context). That goes beyond vhs-decode, too. Only newbies are duped by this.

Quote:

shenanigans some do to advertise vhs-decode....

And if it was truly a superior method, shenanigans would not be needed. Big claims require big evidence, but all we get is misleading "proof".

Quote:

ie the issue is "what exactly are we trying to fix with vhs-decode?". or "what do i get for all that complication?"

I sometimes think it's the answer to a question nobody asked. It's now the mid 2020s, and most video conversion was finished in the 2000s, with redo and later-comers in the 2010s. The "way of the future" is two decades too late, and the project is on that same downwards trajectory.

I think the real value lies in formats that do not have quality transfer equipment, namely Beta(max), but also others.

To me, the -decode for VHS was just reinventing the wheel -- or more like an old tire, on a 1985 Ford Taurus, one of the most boring cars ever. If this project is really about money, then uber-budget viable methods, cheaper than vhs-decode, already exist. If it's about quality, then beating out premium workflows hasn't been done either. So what was the goal? That's a good question.

Quote:

it's also about ease of playback, everything plays mpeg, what about huff?
i mean sure, you can play it on pc with qtgmc on-the-fly with newer pcs, but not on tvs....
some of which, as you'll remember, do have decent processing and you can turn off dnr completely.
AND cce is fast. VERY fast. it's almost as if someone made mpeg2 encoder in assembler.

You're pulling out names that most people don't even know anymore. Do I know you on VH? You remind me of SatStorm, but I've not heard from him in 10+ years.

Quote:

you know, going down that path one might end up in all sorts of weird thoughts, one of which might be:
how many people today are actually watching worse video image quality today than in analog days?
(a blasphemy!)

We live in an era of sheeple eating up disinformation. Sadly, yes, many do watch inferior quality now, none the wiser. And viewed smaller than a 13" CRT TV.

Quote:

and h265 didn't help much.

H.265 was actually worse in many ways. It's just another thing that is promised to be better, but "better" is not really accurate. It's just different, both good and bad. Like this project, like so many other things in the realms of video.

[i4004]

let me just answer this to myself:

Quote:

"oh, gosh, do i need 3 capture devices to capture video, hi-fi and linear audio?"

nope, but you actually need a lil electronics project consisting of 4 pcbs seen in image taken from
https://github.com/oyvindln/vhs-decode/wiki/CX-Cards

this just blew my mind, and not entirely in a good way!

offcourse, on top of that you need to modify one cheap cx card that will be capturing video.
and rf amp board: https://github.com/oyvindln/vhs-deco...-amplification
maybe.

back in the "ancien régime" you would just use pc's existing audio inputs (just as i still do), capture software would do the rest, you end up with one file that has both audio and video inside, imagine that!!!

[latreche34]

The image processing of a high end S-VHS VCR or a Hi8 Camcorder with built in TBC/DNR is as close if not equal in certain cases to that of the VHS-decode, The problem is that these newbies that go buy a thrift store VCR for $15 and face an awful capture experience, think VHS-decode is magic, While it does process the RF signal decently like a high end VCR chip, it doesn't produce a final YUV video signal that can be used by normal video tools, You end up with huge FLAC files that need to be processed and all chroma and luma parameters has to be tweaked or else you end up with video like in the link above or like this one with massive ringing and ghosting.

[i4004]

Quote:

Almost all decks from the 2000s/10s are low-end Funai, used by multiple brands. It was literally the only manufacturer left.

the chassis (mechanism) suddenly became thin metal that twists and causes mistracking....quite a development!
electronics didn't fare much better either. too much integration was cost cutting measure....vcrs should have separate head-preamp that's shielded, not everything in one chip.

Quote:

Hear/see where? Commercial duplication, where decks were daisy chained together, were mostly for small-run productions. It simply was not as economical as a contact dupe with scale. But I have to add that most European countries were "small scale", so I can see a higher concentration of manual duping in PAL lands.

you just answered yourself. saw ads in magazines on video technology, they were mentioning particular duplicatng machine's models. so one just needs to know which was made in greater amounts, "professional" copying decks or contact duplicators.
yes, contact is faster, but 100 decks working non-stop also has ok output.
dunno how big were the runs anyway.....maybe few thousands for every title?
in europe more wouldn't make much sense....
were there ever any numbers of sold vhs tapes?

contact duplications still sounds weird to me, i mean how do you heat tape to over 100°c without destroying plastic?

Quote:

Nah. CRT era, HDTV/LCD era, no difference. It's just as variable now as it was back then.

hardly. crts did no processing, so what's left is color balance, which was usually ok.
there was no deinterlacing and upscaling as factors of image destruction.
it was always "1:1" pixel mapping. with lines instead of pixels.

but you're right about sharpness nitpicking in a sense it would be mighty hard to gauge it on crt tv and vcr plugged to it.
it's one thing to compare static screengrabs, and another to look at crt with constantly moving images.

Quote:

The best decks came in the 90s, but the early/mid/late really depends on manufacturer.

once you see flat cables runing out of "upside-down" drum assembly, it seems it was over....
dunno about jvc, it's wasn't that big in europe.
there is a clear disctinction in markets, it's obvious some japanese makers concentrated much more on bigger (us) market.
there was also less of hitachi and mitsubishi.
thsis is also visible today when scanning the used machines market.
there were some german companies that were just putting their lable on japanese machines, esp. i later days.

Quote:

I sometimes think it's the answer to a question nobody asked. It's now the mid 2020s, and most video conversion was finished in the 2000s, with redo and later-comers in the 2010s. The "way of the future" is two decades too late, and the project is on that same downwards trajectory.

but was it done with qtgmc?
but yeah, not much market there, esp. in smaller countries.
maybe some work on documentaries, but that's still niche.
for example i recently watched documentary on williams sisters and their father, tennis stars.
and not a second of sd content was deinterlaced. and there's lots of it, their home footage.
it's like having a woman director who has less chances of being interested in video tech.
not a second was deinterlaced. not one.

vhs-decode could of been done better, they kinda forgot the audio capturing part, probably because they started with laserdisc (domesday project archiving) which doesn't have separate audio heads like vhs.

but in many yt comparisons they actually have better video quality than other capture methods.

their wiki is not structured well, too complex block diagrams (mixing hardware and software processing, pretty bad) and too much text in wrong places. it was extremely hard for me to decipher what exactly are they doing with the audio.
it's buried deep....people will buy these cx cards and then figure out their audio is out of sync....
or maybe they won't, because you first need to mod the card.

Quote:

You're pulling out names that most people don't even know anymore. Do I know you on VH? You remind me of SatStorm, but I've not heard from him in 10+ years.

nah, i just registered on vh to reply to harry of vhs-decode, with the audio objections.
was briefly on doom9, some more at arstechnica, some local satellite/pc/electronics forums.

we did this:
http://www.doom9.org/index.html?/capture/start.html

Quote:

You are now reading the (v4.0) version of the guide. This time the changelog is substantional. Many things are added or rewritten by the Version4 team (Wilbert, trevlac, arachnotron and i4004 (Ivo)).

always the same nick.

resurrected my interest in vhs mainly because this newer machine got broken, and it's not easy to find vcrs with ep speed in europe. stumbled upon video99 fixing that very problem (magnet unglued from inside the drum...ridiculous) and saw vhs-decode mentioned. also looking to find another sony (but same model as i already have) to serve as backup, or "for parts", because i'm still not done capping vhs. the content i picked back then suits me, feminist/minority talk of today's tv doesn't, lots of lying happening there.
i have good stuff that can be watched many times. AND tons of stuff that's digital. era of vhs recording stopped once i got pc in 2000. i was sure to include capture card upon building my first pc. infact i dunno would i even have pc if it wasn't for video.

Quote:

We live in an era of sheeple eating up disinformation. Sadly, yes, many do watch inferior quality now, none the wiser. And viewed smaller than a 13" CRT TV.

i'm quoting this for truth.

Quote:

H.265 was actually worse in many ways. It's just another thing that is promised to be better, but "better" is not really accurate. It's just different. both good and bad. Like this project, like so many other things in the realms of video.

they're doing 1080p50 h265 on terrestrial vs. 1080i50 h264 on satellite here. and h264 looks better.
265 was touted as 4k codec, but it didn't really happen, claims of "50% more efficient" are usually just marketing.
just like divx/xvid was 50% better, but they had to downsize dvds to 640x360 to do it.
also an ancient history.

manono (veteran of doom9) had a post where he says only mpeg2 can deal with noise.
offcourse, i knew that much before he said it because i was comparing codecs on analog captures. they always have noise.
and new codecs don't like the noise. yeah, one can tweak h264 a lot, but still....
but h264 is a good hd codec.

there are also limits to what is possible, which collide with human capacity to fantasize (and that's a nice subject for philosophy/psychology).
you will never be able to have good quality digital video at 200kbit/s. sd or hd.

Quote:

it doesn't produce a final YUV video signal that can be used by normal video tools, You end up with huge FLAC files that need to be processed and all chroma and luma parameters has to be tweaked or else you end up with video like in the link above or like this one with massive ringing and ghosting.

the weird thing is it seems to take a lot of cpu, people are talking about 10+ hrs for a movie.
that's just converting from raw rf to "normal" video files.
dunno about processing, i don't think that's mentioned here:
https://www.youtube.com/watch?v=Xb128g617sg
shouldn't be needed, ie command line utils should just convert without processing. the fact they have toys like vectorscope doesn't matter much.

it's just converting rf file to video file.

https://www.youtube.com/watch?v=mWGEBpDBGg8 ?

the video is 2x sharper on vhs-decode.
and not "unsharp-mask sharper", but truly more detail.
and, sure enough, much more noise, but like i said, these go hand in hand many times on vhs.

[latreche34]

Audio and video cassettes were massively produced using contact duplication, The reason why most of us didn't know about it is because it was an industrial task, unless you are in the business no one would tell you about it, The cost per unit is very cheap and the turn around times are very short but requires large batches minimum orders. Duplication by recording machines and multi copy duplicators was done by small production houses for corporate, religious, education and other small scale media entities.

Quote:

Originally Posted by i4004

https://www.youtube.com/watch?v=mWGEBpDBGg8 ?
the video is 2x sharper on vhs-decode.
and not "unsharp-mask sharper", but truly more detail.
and, sure enough, much more noise, but like i said, these go hand in hand many times on vhs.

It looks good on video until you read the description:

Quote:

Our subject matter is an old NTSC VHS tape from 1989(!), which had a copy(!!) of output from a VHS-C camcorder (presumably using composite) onto a regular VHS tape at EP speed.
Left half was captured using a composite signal, right side is the RF capture.

Do you think this is a fair comparison?

[lordsmurf]

Quote:

Originally Posted by latreche34

Do you think this is a fair comparison?

Like having a taste test, but one is food, the other is a cat turd -- not both food.

[i4004]

Quote:

Originally Posted by latreche34

Audio and video cassettes were massively produced using contact duplication, The reason why most of us didn't know about it is because it was an industrial task, unless you are in the business no one would tell you about it, The cost per unit is very cheap and the turn around times are very short but requires large batches minimum orders. Duplication by recording machines and multi copy duplicators was done by small production houses for corporate, religious, education and other small scale media entities.



It looks good on video until you read the description:


Do you think this is a fair comparison?

i dunno who made more vhs copies, contact or duplicators, dunno for europe, dunno for usa, one would need someone from both camps to share some memories, or numbers, but i do know i never saw commercial tape that looked anywhere close to decent.
copies of my ota recordings look better than that. made and copied on usual 2 head vcrs.

as for comparison, you mean because he's capturing composite?
yes, in that sense it's fair, he's using same tape and same vcr for both processes.

but it's not fair in a sense he doesn't mention composite capture device OR codec used.
if it's diamond vc500 + mpeg2 on-the-fly....oughm.....
also not fair in a sense he didn't set contrast properly on composite cap, so highlights are blown out, it's also a nice way to destroy details....

and the vcr he's using ( https://www.youtube.com/watch?v=BC316-Xqh90) is already a crap to avoid.
to that extent it's interesting to think about where vhs-decode helps more, on decent older decks, or "recent" crap where they're bundling rf amp in y/c processing chip.
probably the latter.
oh yeah, notice vcr has "reality REgenerator".
advertsing gimmicks is interesting theme, it's not just usual electronics, it's special, with special names.

"fair comparison" is if these boys from uk send me tape they did with vhs-decode so i do it too.
then compare.
on frame level, offcourse.

[lordsmurf]

Quote:

Originally Posted by i4004

i dunno who made more vhs copies, contact or duplicators, dunno for europe, dunno for usa, one would need someone from both camps to share some memories, or numbers, but i do know i never saw commercial tape that looked anywhere close to decent.

The contact method can result in degraded copies if the processed is not properly cleaned/maintained, and not enough masters are used. Also in the quality of the blanks used for the duplication. Major studios were sticklers (even though some of them were still bitten by shoddy operators), but small operations were tightwads who got what they paid for. Or rather, the consumers got screwed, and didn't get what they though they were paying for.

Quote:

or "recent" crap where they're bundling rf amp in y/c processing chip.
probably the latter.

To an extent, it really is about putting perfume in the outhouse, all while ignoring the fact that you're in an outhouse (and not a bathroom with "modern" plumbing).

[i4004]

yeah, i've read that sony system required master swaps every so often.
dunno about otari.
a decent parameter would be number of imported machines, if it's small, and they didn't work 24/7 then they probably couldn't account for majority of copying, even though it was fast.
better parameter yet would be meters of tape delivered to copiers (distributors)....that would be "money shot" of this dilemma.

Quote:

To an extent, it really is about putting perfume in the outhouse, all while ignoring the fact that you're in an outhouse (and not a bathroom with "modern" plumbing).

probably boils down to simple things, a la "we probably won't be designing new/cheaper/worse heads, let's just sell what we have and be done with it, they'll save on electronics and mechanism...and they won't have discrete rf amp anymore". where "we" is heads mfr, and "they" is vcr mfr.
i think futaba made most of the heads.

[lordsmurf]

Quote:

Originally Posted by i4004

probably boils down to simple things, a la "we probably won't be designing new/cheaper/worse heads, let's just sell what we have and be done with it, they'll save on electronics and mechanism...and they won't have discrete rf amp anymore". where "we" is heads mfr, and "they" is vcr mfr.
i think futaba made most of the heads.

No, it has nothing to do with manufacturers, or even the developers of this project. Much in the way how the ES10/15 is used as a half-baked "TBC replacement" (it's not), certain cheapskate elements try to use this project as a way to still use their awful low-end VCRs. vhs-decode probably can resurrect some % of the destroyed signal (as destroyed by the crappy VCR), but the real quality will come from ideal VCRs. The same ideal VCRs used for standard workflow. But actually, a subset therein, as it rejects JVC decks. When you're using AG-1970/80, or comparable PAL decks, you make almost no gains, and in fact negative gains.

That's where we're coming from. If you want quality, you get quality workflows.

If you want to poke around garbage e-waste VCRs, then vhs-decode can maybe help you salvage your VCR to suck less. Not be good, just suck less.

The output is wildly variable, and even just today the lead developer stated that vhs-decode is "just a community project with limited resources so the software and hardware still needs a lot of work yet". And he gets much respect from me, because he's grounded in reality, unlike some of the project fanboys that want to turn it into a cult. It's a tool, not a religion or political movement.

[Gary34]

Quote:

as for comparison, you mean because he's capturing composite?
yes, in that sense it's fair, he's using same tape and same vcr for both processes.

REALLY?!?! How was that a fare comparison? They are using unknown gear in unknown condition over composite to represent traditional capture.

-- merged --

I have just seen unrecommended unrefurbished gear tested against decode.

-- merged --

What capture software did they use for the conventional capture?

[aramkolt]

Sort of related, but this guy uses VHS_Decode to rescue an obscure movie (Ricky one):https://www.youtube.com/watch?v=pEzmbw_Y-Tw

He actually references recommendations from this site including AIW, line TBC, Frame TBC

I do plan to do my massive comparison of devices and VHS_Decode will be a part of that. I'll be transparent with my capture methods and post raw capture samples of each. I will have a basic panasonic omnivision over composite as a sort of "standard" of what most people would use with a variety of capture cards to show how much the VCR vs capture card matters. Should be worse, but how much worse than recommended equipment?

[lordsmurf]

Quote:

Originally Posted by aramkolt

Sort of related, but this guy uses VHS_Decode to rescue an obscure movie (Ricky one):https://www.youtube.com/watch?v=pEzmbw_Y-Tw

There is a lot of halo/ringing in those vhs-decode captures.
The artificial sharpen knob was cranked to 11.

I see that repeatedly, especially with NTSC captures. That is not acceptable for archival, or even sub-archival use.

[Aya_Rei]

Quote:

Originally Posted by lordsmurf

There is a lot of halo/ringing in those vhs-decode captures.
The artificial sharpen knob was cranked to 11.

I see that repeatedly, especially with NTSC captures. That is not acceptable for archival, or even sub-archival use.

Certainly does look that way, a clear example being the scene with the mobsters. The haloing being visible on the edges of their black hats. While I always seem to get footage with halos even with a proper workflow, the halos mainly show up on hi-8 camcorder footage, easily visible when it's the edges of objects like shrubby against a bright clear sky or text on signs.

This message might be a bit off topic as I don't have a lot to say over vhs-decode, the Ricky 1 result is at least better than some old 240p YouTube upload from the early 2010s, but can probably look even some % better with a traditional workflow instead

[Feedbucket]

Oversharpening seems to be a trend in the Decode samples I've seen. Wonder if it's systematic or if people are choosing to render that way.