Current status of ld-decode / vhs-decode (true "backup" of RF signals)

[dellsam34]

General question to European VCR owners who are experimenting with vhs-decode. Have you guys tried capturing a native NTSC tape? Technically it should work since most European and Asian VCRs are two speed VCRs (525/625), If the speed is correct the VCR should pickup a pure NTSC 3.58 at the RF level.

[Brad]

General question to European VCR owners who are experimenting with vhs-decode. Have you guys tried capturing a native NTSC tape? Technically it should work since most European and Asian VCRs are two speed VCRs (525/625), If the speed is correct the VCR should pickup a pure NTSC 3.58 at the RF level.

Yes, it does work as expected. oln has noted though that these VCRs may not optimally read NTSC Hi-Fi, especially in EP mode.

Here is one of Harry's examples. The "reference" capture is PAL60 from the VCR, poorly interpreted by a DMR-ES10. Those Panasonic DVD recorders don't properly support PAL60 input.
https://archive.org/details/dream-times-minayo-chan-1987-vhs-ntsc-j-fm-rf-archive

---


Hopefully I'll have some worthwhile AG-1980 samples ready to post this week. Maybe tonight if I'm really lucky. The rewolf amp does resolve the amplitude drop + crosshatch issues.

[dellsam34]

The HiFi issue can be mitigated by physically tweaking the stator angle on VCRs with the stator on top, I've had success doing so on some problematic tapes.

[pasteur10]

about capturing cx card.
I have found 3 versions of them.
which ones do correctly?


[Attachment 86245 - Click to enlarge]
https://www.aliexpress.com/item/1005008463187953.html


[Attachment 86246 - Click to enlarge]
https://www.aliexpress.com/item/1005008509975212.html


[Attachment 86247 - Click to enlarge]
https://www.aliexpress.com/item/1005008188029346.html


also i have hackrf. it can capture rf signals.
has anyone tried it before or is it possible at least?

[i4004]

this is not a valid test: dvb-s (mpeg2) encoder might of had totally different setup when you were (re)recording it from satellite. ie lower bitrate.
dvb-s .ts rip will always look better than vhs recording of it, or s-vhs recording, or sp beta recording, doesn't matter which analog format u use. it's just that better formats will yield less deterioration.

What is not valid is your understanding.

Let me explain once more the test in that link.

On one side there is the rip of a DVB-S broadcast at 480x576 resolution, MPEG-2 encoded at good bitrate 5/6mbps average bitrate. And from source master in good shape. In principle, from a technical point a view, slightly better than the max quality of the S-VHS standard, which allows rafely 400 point of horizontal resolution inside a line and 576 interlaced lines. Definetely a nice video when whatched.
(BTW, the rip is in .pva format, not in .ts; at that time there was no hardware able to rip the full stream)

On the other side there is exactly the same broadcasted program, transmitted few years earlier, when I did not have the hardware to rip the digital stream. So the flow to produce the digital file was DVB-S set-top box Y/C output to high-end S-VHS VCR recording on high quality Sony VXSE tapes, then an analog capture of the tape via its Y/C output to a capture card.

When comparing the two files there is almost no difference in term of quality, meaning that the whole process of capturing, but also of recording, is not introducing significant deviations from the original. Impressive!

As a side note, we all know that on Astra and Hotbird in Europe there were / there are channels with crap quality, at low resolution, too low bitrate and not optimized encoders and low preformances studio equipements (cameras and broadcast hardware). It was not the case for Canal Jimmy (my source for the test I linked) inside D+ bouquet in late 90s/early 2000s.

dvb-s .ts rip will always look better than vhs recording of it, or s-vhs recording, or sp beta recording, doesn't matter which analog format u use. it's just that better formats will yield less deterioration.

Yes in principle, not in practice in the test I showed, given the specifications I gave.

In order to see a difference the reference must be a D1 720x576 at 8/9mbps bitrate from a perfect master. Then, of course, the (limited in comparison) S-VHS specifications will not be able to match the original quality of the source.

i wholeheartedly disagree with LS there: you don't need TBC. at all.

Concerning the need of an external TBC, I am one of the few disagreeing with lordsmurf that is always needed, because I do not see a need for tapes in pristine shape like mines. But is not so common, so what you wrote in term of TBC "you don't need TBC. at all" is a non sense.

edit: and while comparing, as you can understand, I can also move frame by frame, field by field the reference and the analog capture at the same time and verify that not only the quality of the second is in pair, but also that there are not dropped or inserted frames or whatever, both with AviSynth scripts and by visual inspection. As you probably know these potential issues are of great concerns when capturing analog.

dvb-s mpeg2 bitrate inside transponder was dynamically allocated, ie some channels were prioritised. for example "less important" channels got more bitrate only if "more important" channels had static video that required less bitrate.
you're literally not comparing the same thing, it's apples to oranges.
on the other hand, i'm not contesting s-vhs has decent quality. but on the third(!) hand s-vhs has same crappy chroma bandwidth as usual vhs, AND dvb-s has crappy 4:2:0 color encoding! this is where recording dvb on vhs becomes rather questionable. it seems to me vhs actually needs some noise, to "excite" it, and dvb has none. dvb recordings on vhs for me were unbareably soft. good thing that didn't last long! ie we started ripping it over capturing it from vhs.
offcourse, these (s)vhs color issues become more apparent once you start copying, ie master is ok, copy is instantly recognizable as copy.
just like recompressing digital video. but digital video has big advantage, you just copy the digital file, and then it's 1:1 to "original". infact it is original. offcourse that advantage gets lost when vhs-decode boys upload raw files...hehe.....250gb.....

tbc, still you:

And the A/D D/A is not the only aspect; any additional element introduces (neglettable or important) degradation. Better do not risk if not necessary

so do we need tbc or not?

Off topic:

BTW, if we could have access to the digitized data at the output of the TBC circuits inside the VCRs it would be a wonder: no capture card needed, just a packing data/format conversion/what else.

this is peculiar thought! why do you think adcs in those 90s machines were better than digitizers on (cheap) capture cards some 10 years later? for example bt8x8 and saa713x.

esteemed lordsmurf (btw. no tongue in cheek there):

The A<>D process is not harmful any more than colorspace conversions are harmful.

have you tried 5x colorspace conversion vs. 5x a-d+d-a process? heh...

Consider this:
- If TBCs only cost $10, almost nobody would state TBC isn't needed.
- If TBCs were embedded in all capture cards (and all cards cost, say, $300 minimum), almost nobody would complain.

point1: no, i just don't buy stuff i need not.
point2: tbcs essentially ARE INCLUDED, by design: cheap cards have adc front end, just like TBC.
it's just a matter of which adc you used. to illustrate:
https://youtu.be/S1ndneq38xY?t=136
(excellent, moving camera is enough to confuse elgato! incredible stuff!)

You could even think of TBCs as insurance. With frame TBC, you know you're getting the best "quality" (actually capture integrity, signal quality, not visual). Without TBC, it's really just guessing, or overly relying on drop/insert counters (which can be deeply flawed with some cards/software), unless a deep-dive verification is performed post-capture. The net effect is that TBC removes project time needed.

how? what makes you unsure when you lack tbc? we need to go back, to the beginnings:

That makes zero sense. "In the beginning" (of modern ingest/capture, meaning the late 90s, early 00s), the items were sold new in stores. Nothing was scarce. I remember, because I was there.

but were they any good? what was your first capture card? and what was wrong with it?
mine? vpx3225 chip on riva tnt2 graphics card.
good image quality, dropped frames for no reason, couldn't do 25fps, not even near.
(yeap. i remember part number, vpx was made by "micronas")

i mean, i think i see the issue:
https://www.digitalfaq.com/forum/marketplace/8253-sale-ati-600-a.html
you love ati!

Wrong assumption. Released retail material has never been my focus.

so what was? what do you have such a big troubles with that you need tbc every time?
1st generation copy? that doesn't have tbc issues even on consumer grade vhs hardware.

This is not a true statement. And it seems you make the blanket statement based on your own narrow (and unverified) experiences. If TBC was not needed, why do you think I, and others, buy them? Do you really think we're just stupid?

this is not an argument if you can't explain why do you need them. it's like saying "i drink beer, others drink it too, it's excellent, it's healthy!"

and my experience is not narrow or unverified: let's just quickly verify it:
here is cheap card capturing white noise, ie no sync signals at all:
https://www.youtube.com/shorts/oDWUVkI6sek
here is...well, see for yourself:
https://www.youtube.com/shorts/8R0IWcxA_mw

here's cap of ffwd:
https://youtu.be/z-EkBTh0pNg?t=117
(you'll notice change in voice pitch when audio resampler tries to keep pace with video, you'll also notice sync is not lost as soon as normal playback continues)
here's hd version, if you wanna inspect how good is yt at sd vs hd compression, and how good or bad does it resample pal source to ntsc's 480 lines:
https://www.youtube.com/watch?v=RSJIKgPzNps
notice how crappy and mangled that video signal is: now, come again, why do we need tbc?
i mean you have portions of that clip that were capping just one field of signal, because one head was dirty! lol! for example flicker at 2:08

were there dropped frames there?
sure!
how many?
a lot! on crappy video portions.
but...so what?
are we capturing white noise or cueing video or one-field vhs?

from your reply to lollo:

It's not so much that it's "always" needed, but rather "almost always". There are exceptions. However, everybody always wants to think themselves the exception. So the general rule is that TBC is needed, with exceptions. But the exceptions do not overrule the rule.

no, you can't prove this.

TBCs allow us to babysit captures less. We can let the capture hardware do its thing (VCR>TBC>card captures out video), with monitoring, but with no real invention needed.

You could even think of TBCs as insurance. With frame TBC, you know you're getting the best "quality" (actually capture integrity, signal quality, not visual). Without TBC, it's really just guessing, or overly relying on drop/insert counters (which can be deeply flawed with some cards/software), unless a deep-dive verification is performed post-capture. The net effect is that TBC removes project time needed.

insurance is quite an english/american concept, and it's a good one! esp. when it comes to health insurance!
another good saying of yours is "if it ain't broken, don't fix it!"
i would say in this case try with decent capture card....sync issues?
sure, get tbc.
but not before you try without it.

This line of argument is extremely recent, last few years, started by newbies

i'm rather more interested who started the "everybody needs tbc!" mantra, and i think ti was probably you.
https://forum.videohelp.com/threads/111648-ATI-AIW-drops-frames-but-only-on-one-computer and i hope yor standard process was not capping straight to mpeg2!

to that extent:

You just need a standard workflow: VCR > TBC > capture card.

If you have XP PCs, then ATI AIW are outstanding, some of the best cards ever made. When you say "9800 PRO", be sure it's "All In Wonder". but not just am ATI graphics card. It has to be AIW to have the video. Remember, AIW is just video, audio is slaved to a quality audio card, such as Turtle Beach Santa Cruz.

i think i alread wrote this somewhere, maybe on digitalfaq: i once tried ati-aiw on pal, and it was biggest pos ever.
incredibly blurry images. composite input. they were not even trying. has to be something along these lines:

GV has PAL fans, but not really NTSC -- which isn't unusual, lots of cards do favor one format, then "also do" the other (but compromised in some way, either video quality or usage/UX quality).

(my bold)

as for ld-decode and tbc, to neatly bring us back to subject: would you concede that aligning lines in software is cheaper and probably can be done better than doing same thing in hardware from 90s?
i mean i didn't do deep or any dive on this (i didn't dload those rf files that are 250gb in size to inspect are lines shimmering left/right, hell no! hehe), but what would your guess be? who would be able to do it better, today's cpus and xx gbs of ram and (relatively) simple algos, or 90s tbcs?

So at what point is it really about "better" quality? -- which is actively being debunked by others, with true 1:1 S-VHS vs. vhs-decode samples,

it was dubunked? somebody has done both vhs-decode and analog cap with the same machine and tape?
and vhs-decode looks worse?

dellsam34
Re: Current status of ld-decode / vhs-decode (true "backup" of RF signals)

General question to European VCR owners who are experimenting with vhs-decode. Have you guys tried capturing a native NTSC tape? Technically it should work since most European and Asian VCRs are two speed VCRs (525/625), If the speed is correct the VCR should pickup a pure NTSC 3.58 at the RF level.

i'm not experimenting with vhs-decode, but one can capture pal60 (pure ntsc tape) without problems in europe, again with a cheap capture card.
feature on vcr is called "ntsc playback on pal tv".

Not sure how hard to reverse engineer those chips. Maybe it's impossible, maybe it's because no one has tried it yet, Who knows. But if extracted, it seems that it's much easier to deal with digital pockets of scan lines than an entire process of RF capture and conversion.

again, as i said to lollo, why would you do that?
were those adcs inside vcrs even sampling video at 27mhz?

MrCreosote

I'm such a noob here. I have a lot of gear, SVHS VCR's, DVD HDD recorders, XP PCs, but not much in capture cards.

I'm working with SVHS and wondering, compared to VHS, is it worth the extra effort for RF-capture?

I looked for the ATI AIW 9800XT and Radeon 7500 cards and found not much on eBay. Is the 9800 PRO as good? I just don't know. A list of OK cards would be nice. Then read (pardon garbage memory) Brookstone? BF878 Conexant card someone wrote RF-capture firmware for? or? (don't quote me)

I remember somewhere, in some catalog, maybe Heathkit, ??? Lafayette Radio ??? where they had hi-fi systems (turntable, amp/receiver/tuner, speakers for various budgets. Would LOVE that for the Food Chain of capturing SVHS.

I definitely NOT a $$$ pro TBC maniac and honestly, RF-capture would require more time than I have left on this earth although the improvement over VHS is rather amazing BUT is it as dramatic on SVHS? I know REDs suck. (I recall somewhere about 60 lines for reds?, for VHS?)

Best "bang for buck" VHS capture? (I have read the Retro Game captures are NOT good for VCR. RetroTINK even expensive, not ideal.)

you mention brooktree bt848 and 878.
no nobody will rewrite drivers for those, but i just checked sampling frequency of vhs-decode projects:
https://github.com/oyvindln/vhs-decode/wiki/Signal-Sampling
they're actualy using LOWER sampling freq. than bt8x8!
https://en.wikipedia.org/wiki/Brooktree

vhs-decode uses something else, some cheap surveillance card with newer conexant chip (conexant bought brooktree) (which seems to be not so cheap now?) and a re now making alltogether different hardware.

i used cheap bt878 card a lot, in pal land, no issues with quality or reliability.

yes, color on vhs is crap, yes, especially visible on reds. it's even less than 60 lines. vhs can't stack 60 vertical lines of different colors on screen. and neither can s-vhs. and it quickly falls apart if you copy it on another (s)vhs machine.

frankly i doubt anybody can demonstrate big difference between good hardware setups (be it vhs or s-vhs....you'll still be capturing vhs recordings ie not s-vhs recordings....s-vhs has the edge by separating luma and chroma, but only if the recording had is separate too, ie if vhs tape was made via composite copying what exactly are you gaining on s-vhs machine? to that extent i'm just watching some of the stuff from ITV archive on yt: i'm finding composite video with composite video issues
https://www.youtube.com/@ITNArchive early 90s european production) and vhs-decode.
but many can demonstrate difference between crappy vcrs and crappy capture cards vs. vhs-decode!

https://www.youtube.com/watch?v=CQ-u979_6Cw vhs-decode is clear winner there, sharpness and stability.
this was captured from Panasonic NV-SD220. i had that vcr. i have machine producing sharper images now.
even when capturing tapes recorded on sd220, it's sharper.

pasteur10, that's rf, but radio rf!
can't be used for video.
software defined radio is ment to decode radio rf modulations.

abour cards:
https://github.com/oyvindln/vhs-decode/wiki/CX-Cards
notice the "capture hardware" links on the right....below "software installation" links.

[dellsam34]

i'm not experimenting with vhs-decode, but one can capture pal60 (pure ntsc tape) without problems in europe, again with a cheap capture card.
feature on vcr is called "ntsc playback on pal tv".

My question was specifically about the RF pickup not about using the VCR normally.

[MrCreosote]

YES to crappy VHS/SVHS chroma - such a disappointment - reds somewhere I read like 60 lines of resolution - don't know if VHS or S.

Having done countless convergences of Advent Video Beam TV's, it's Blue we can't see.
Red and Green are sharp.
Couldn't they have compromised Blue instead when designing the VHS system? why red?

[lollo]

dvb-s mpeg2 bitrate inside transponder was dynamically allocated, ie some channels were prioritised. for example "less important" channels got more bitrate only if "more important" channels had static video that required less bitrate.

While true there there was / there is a dynamic allocation of the bandwidth in term of bitrate for single channels, I can assure you that for what concerns my tests, there is 0 (zero) impact.
I ripped from satellite hundreds, I repeat hundreds of programs, and never measured major modification to average bitrate for channels in D+/Sky Italy bouquets.

you're literally not comparing the same thing, it's apples to oranges.

The apples and the oranges are in your understanding, the test are pertinents and appropriate. Their outcomes are valid, and outstandings.

on the other hand, i'm not contesting s-vhs has decent quality. but on the third(!) hand s-vhs has same crappy chroma bandwidth as usual vhs, AND dvb-s has crappy 4:2:0 color encoding! this is where recording dvb on vhs becomes rather questionable. it seems to me vhs actually needs some noise, to "excite" it, and dvb has none. dvb recordings on vhs for me were unbareably soft. good thing that didn't last long! ie we started ripping it over capturing it from vhs.

Again, you did not understand the contest. The source is always the 4:2:0 MPEG2 broadcasted by satellite through a DVB-S transport stream. Ripped from digital, or recorded to S-VHS. So any "defect" because color space is in the source (the reference), and then also in the analog captured and "compared" video.

offcourse, these (s)vhs color issues become more apparent once you start copying, ie master is ok, copy is instantly recognizable as copy.
just like recompressing digital video. but digital video has big advantage, you just copy the digital file, and then it's 1:1 to "original". infact it is original. offcourse that advantage gets lost when vhs-decode boys upload raw files...hehe.....250gb.....

Side comments, not relevants.

so do we need tbc or not?

It depends on the source.

this is peculiar thought! why do you think adcs in those 90s machines were better than digitizers on (cheap) capture cards some 10 years later? for example bt8x8 and saa713x.

Beause I am an analog designer and know the chips inside the old cards, the new cards and eventually the VCR. Not assessing anything in general, but different hardware provides different performances. And in the VCR you are closer to the signal, rather than intercepting it outside after some inevitable loss. But it was just an idea, nothing engineered.

[dellsam34]

We don't know at what chroma sub those late 90's ADCs sample at but at the time chip technology was advanced enough to store a lossless digital scan line, Perhaps a specially designed chroma pattern recorded into a VCR and played back can decypher the chroma subsampling of a line TBC if one knows how to design the graphic patterns.