vhs-decode setup? A Basic Guide
4 Attachment(s)
Hello all,
I got VHS Decode working last night, I'm going to do a step-by-step of what I've done (it's not the easiest if you're not familiar with using Github, Linux etc) but I'm going to do a step-by-step of the things that I've done and hopefully it'll be replicable for a good number of people. Just as a grounding, I'm no expert in this, I "know" my video machines, am a technical guy (but not in this field; by qualification), Amateur Radio licence holder (so I have a good foundational understanding of RF), moderately-competent programmer and blessed with a bit of free time and a business with a room I can set this stuff up in. I'm not an expert on anything to do with this in any regard, just a keen amateur. VHS-Decode is something I've been fascinated with, I won't retell the whole story, and I am only a step ahead of you with this, I'm very much a novice with this, but as I go along I thought a guide on how to set it up might be useful for those who've not got their hands dirty with 'hardcore' Python, Github or hardware tinkering. What is VHS Decode? Firstly, props to the guys who've been behind this (and by extension the LD-Decode team) - you're absolute geniuses, I genuinely am in awe at what's been accomplished here, each and every single person who's contributed to this. It's the guys & gals like this who mean helmets like me who sit on the internet can play with this stuff too. Remember to show your appreciation too if you take this further and do what you can to feedback to the community. :worthy: VHS Decode is an application that means that you can stream the video tape straight from the tape-head (not out the back of the machine) bypassing all of the electrical gubbins in the video player. I guess it would mean practically any old 'junk' video machine will do (although not recommended) and sending the RF (radio frequency) information encoded on the tape straight in to a PC. You then let the PC do the heavy-lifting, no worries about TBCs, amplifiers and the rest, just let the PC reassemble it all and turn it in to a video file. This turns the "pseudo-lossless" AIW captures etc from conventional means into lossless in the normal understanding of the term. That's the very basic concept, in a nutshell, condensed, squeezed and vacuum-packed. By heavens, it's deeper than that, but I'm trying to make it digestible for tinkerers, not RF or programming experts. I'm writing this so that somebody with fundamental knowledge of normal day-to-day computing can get this up and running on their machine. This is being installed on a Linux system in my instance, I'm not providing support for anything else, but I'm sure it's entirely doable. I am not skilled enough to be able to offer individual support, and I am generally only one step ahead of you with this. I have a fair experience of Linux command lines but I "know what I know" and I'm still far from an expert, but this should be enough for somebody to at least get a signal off their machine and onto their screen. This is the tricky bit I guess, once it's set up and working the fun starts. The Basic Stuff You'll Need
Anyway enough of the life history of the computer, I'm using but just to provide a bit of colour and show that for tinkering at least, you can use any old PC that you have laying around. One word though, you'll need something with a PCI slot (so, that's basically anything from the 2000s) but you'll need a full-height PC, not a modern desktop office machine (I've shown "Wes" [Image 2] one of our capture machines as these do not accept the 'old style' full-height cards. An SSD wouldn't go amiss, they're cheap and cheerful now, I put a Kingston 120GB in HB for experimentation with that cost about £15 so not expensive. The Capture Card Oh, you'll now need a card, thankfully these are dirt cheap now, I picked up one for £2 on eBay, and a couple more for no more than £10 each. These are old TV Tuner/capture cards, not explicit capture cards like the type more commonly referred to on here: practically valueless. They're plentiful, and I'd pick them up whilst they're still cheap. The actual card I'm using is pictured [Image 3] - this is a £5-odd Hauppauge WinTV PAL-I 34705 REV J198 PCI which I got to work with almost no fuss, but don't get hung-up on finding the exact card, you've got potentially hundreds to chose from. There's a rule though, you need to find one with a Conexant chipset, a 2388x series (it can end 1/2/3 with no difference for what we're doing, my one is a '1') [Image 4] to be exact, thankfully this is almost ubiquitous with these cheap cards. It's the large black chip in the bottom right, at the moment anything else just won't work. Don't get too alarmed if you see this referred to as a second-rate capture card, because remember we're not capturing video, we're just using it as a way of turning the wave-form from the tape in to 1's & 0's for the computer to process later. Right, get that nailed in to the PC, and let's get started with setting it all up. Don't worry about an OS at the moment, I used Linux Mint and I'll walk through that in the next post, don't worry, it's dead easy to do. If you can't wait you'll find a ton of Linux Mint installation guides online, and if you're more comfortable with Ubuntu use that, or if you're an admitted member of the Linux master race... Well, you do you. You will need a GUI version of whatever you use though, as... well, looking at video in a terminal sucks. We'll get this doing something more useful in the next section. More soon, RR |
Thanks for the guide, should make my life a lot easier! Thought you had to get a doomesday duplicator card to do RF captures but now that I know you can just use any CX2388x card I'll definitely give VHS-decode a go! vhs (and ld) decode certainly are magical pieces of software.
|
Thanks a lot!
I am following VHS Decode as the only real candidate to improve my existing capture flow since long time, but sometime is a bit difficult to understand all the evolution, specially on https://discord.com/channels/6655572...34485975351307 A summary was needed! |
Quote:
I'm slowly unpicking it all and trying to condense it into a 'need to know' guide, hopefully, the more people on board the more quickly it'll take off, I'm going to do the next post on Linux Installation and compiling and installing the applications, then there'll be a bit of detour into critical video theory before we start the capture and analysis stuff. I've had another play this afternoon with interesting results. |
My knowledge on VHS decode is limited to what I read, but if you need any help to compile the guide just let me know :wink2:
|
I have no expertise in VHS decode. I'm open to the possibility VCR decode will improve conventional playback but feel its advantages haven't been described even theoretically. In what specific areas of the processing chain is conventional processing known to degrade the recorded signal? Unless that is known it seems like we can only know if improvements are possible when VHS decode is fully implemented. Why not instead identify areas in the analog chain known to degrade the signal (targeting the weak link(s) in the chain) and digitise only them? Does it have to be a totally digital process? Why not a hybrid system? Again I'm no expert.
|
VHS-Decode video quality is mediocre it is miles behind, I can capture EP tapes better than what vhs-decode does for SP tapes. SingMai FPGA is ahead of the game but so far we've seen only a composite device and with SDI output only and lack of USB 3 or USB 4 they literally re-invented the wheel, SDI boxes existed a decade ago with the same 3D comb, 3dNR, TBC technologies.
SD video is Rec.601, There is no reason for change, The only change I've seen in Analog Devices chips for instance is the addition of HDMI, HD, 4K ... SD video has been standardized in the 80's and will never change. |
Quote:
Quote:
Quote:
Quote:
Quote:
Quote:
In the old days, 90/00s, cameras had to have pass filters to correct intolerable noises and artifacts. The OLPF (optical low pass filter) would smear both detail and color, though you'd only see it when you "pixel peep" (view at 100%), or print large. Eventually, cameras that lacked OLPFs were made. (You could remove OLPF on older cameras, but artifcats were intolerable.) Even then, those early sans-OLPF cameras were still slightly artifacty, viewed/printed large. It's much better now. VCRs are the same, essentially VHS OLPFs. vhs-decode is an attempt to capture without those filters, and get to the more unmolested rawer signal. It's not 1:1 identical, there's no mere OLPF to remove. But it's the same concept. everybody has huge HDTVs, and can easily see imperfections and flaws of VHS sources (even with the best conversion methods, some issues are inherent to the format simply due to VCRs), so any % of improvement is desired. The net result, in theory, is naturally sharper video, less chroma artifacts, no dot crawl, no unneeded artificial filtering, etc. Only needed filtering, like DOC and TBC (sans NR). But getting there is a challenge, and this project is nowhere near that goal. Yes, some samples look nice, but it's curated, specific to the sample. Not general use. Quote:
Later? Don't know. I think video newbies are way too optimistic. It reminds me of the overly insane giddiness for EVs, many of which have massive errors, and don't at all work as initially promised. For example, Tesla is more of a cult with a "dear leader" than car maker, and I don't want to see vhs-decode turn that way. I don't want vhs-decode to eat tapes and have the hardware catch on fire, all while promising magical abilities to turn VHS into HD. We need to keep the project grounded in reality, if it's ever to succeed. Quote:
|
The beauty is that raw data captured and saved right now can be reprocessed later as the decode software gets better.
Question: I was to believe that you need a Domesday duplicator unit to do these captures, is that not the case? I have one, participated in a group buy, but never heard of anyone using a normal capture card. |
Quote:
This is another statement that I've heard since the 90s: "the data will get better later with better software". Never happens. The nature of the capturing, even RF, requires interpretation. The best interpreter will have to makes some decisions in hardware, before software. If you fill up the HDD now with captures, hoping software can cure all that ails it, don't be shocked later if re-capture is required. Been there, seen that. This is just the nature of computing. I hate to see the "everything will someday be possible to fix in software!" attitude for this. hodgey/oln is obviously very cognizant of limitations with this, and he's a developer. His openness recently, in another thread, for FPGA to resolve some issues, is exactly what we need to see. Not magical thinking that software can do everything, as it simply is not so. A software-only solution is probably DOA, or non-DOA and disappointing. (This is not too dissimilar from bitcoin mining, where software and even normal hardware is not workable. Special hardware required. Software cannot do/fix everything.) Another potential problem is low-end BT/Coxexant chips being used for this. That probably won't work long-term, as those chips have their own limits and problems. |
Quote:
|
Quote:
Quote:
Quote:
I'm not a GUI Linux person myself, I used Ubuntu Server for many of my projects before switching to serverless, I've got no dog in this fight, I just used the OS suggested in the repository notes. Mint is pretty lightweight too, going back to my comments about 'finding an old PC'. I'm not sure what you mean by "shell-with-GUI" as terminology is important. Mint is lightweight, well supported, demonstrably works and is a nicely laid out interface for those who aren't familiar with Linux. I could have picked any given core distribution, but that's not really helpful for anybody reading this. As I said though if you are a Linux expert, do whatever works for you, I'm just using the OS suggested in the GitHub repository, it's also very easy to work with, lightweight(ish) and demonstrably works. I make no apology. Are you in the development group by the way? |
I noticed that VHS decode has some (small) similitude with the diatribe edit=off versus edit=on.
Fans of edit=on says that is better to capture without all VCR NR old hardware filters because a better result can be obtained with software post processing. I personally use for 90% of my tapes edit=off beause the capture video is "better", and because in post- processing I was not able to make the video captured with edit=on as good as the first: some small fine detail is always removed with filtering; but it could be a limit in my usage of post-processing. (For 10% of my tapes edit=on reduces a lot the well known problem of frames blended at scene change, so I use it) However, I think VHS decode project is worth, because in theory you extract the signal at its real "source", while at the Y/C output lot of processing happened internally, and part of it cannot be removed even with edit=on, and because post-processing filters will improve (and many can use them better than me :)) I agree that today we are far away to what can be achieved with a good classic flow, but things may change. I personally hope that will change because I do not see any better achievement in term of the "visual quality of the video" when I compare my videos with captures obtained with a much more sofisticated flow such as SDI capture, high-end capture cards, Panasonic HDMI output capture, etc. Of course in term of "stability" of the captures the more sofisticated flows have advantages. I rely then on VHS decode, thinking that while probably not a breakthrough, together with dedicated post-processing can improve the "visual quality of the video" |
Quote:
To the greater point about why? - Honestly, instinctually I think in the future this could be a infinitely better way of performing captures (with our without adjunct hardware, that's to be seen) - the cold reality is we're faced with the following at present: 1) Decent capture cards - getting very elderly now, the prices seem to be very 'controlled' on them, plenty of misery trying to get them to play nicely with modern hardware and that's not a situation that will improve. 2) Decent DV - points above about capture cards, NTSC especially sufferers here, PAL is good on DV but NTSC is horrid. 3) Use SDI, not for the faint of heart, requires a massive investment to get working hardware for domestic videotape capture. We're slowly switching over to SDI, but it's not an especially cheap game to get in to. 4) HDMI, the results get more-and-more interesting, this I think is where it'll eventually head, it's cheap and established technology with what ostensibly appear to be ever improving results, it'll always have detractors but look around for some rather intesting experiments that have been done with it, it's heading for this answer in my humble opinion. 5) Direct RF capture, always going to be a 'nerdy' passtime whatever happens, introduces it's own problems but weighed out by a full clone of the information of the tape. Who knows where it's going, I don't! 5) USB dongles and similar junk - they're alright enough for somebody who just wants to quickly view something on the computer, but nonsense for anything beyond that. Then we add in additional hardware, TBCs, good quality consumer-grade units cost a lot of money now and to set up a serious rack is supremely expensive. Then the powerful, but often impenetrable software we use for processing - AviSynth etc are very good but they keep a lot of people out of the hobby these days, there are of course great alternatives but it always comes back to AVISynth. This might be self-deafting with RF Capture though, I'll fully concede that. AVISynth is great, truly great, before anything is misintrepreted. |
I know vhs-decode is nowhere near as good as a conventional capture setup, for now at least, but how does ld-decode compare with the traditional way of capturing, for laserdiscs?
|
Quote:
Quote:
There's nothing inherently bad about USB cards. USB is just a signal carrier, much like AGP, PCI, PCIe, Thunderbolt, etc. You have both good and bad USB capture cards, as it has been with all others before it. You forgot one of the most important problems: hardware fail. It's expensive and/or time-consuming, if even reparable at all. This is the one that keeps me up at night. It's the reason I have a stack of TBCs, VCRs, and capture cards, as backups. I'm not swayed by the costs argument. People will spend $1k/year on fancy coffee, something they piss away (literally). But $1k for a video setup, to archive cherished videos forever? "OMG! The sky is falling! Wee-wah-wee-wah! EXPENSIVE! ROBBERY! AAAHHH!" It's quite frankly completely head-up-ass in terms of value. And when you're done with the project, you can sell the gear to recoup costs. You don't throw it in the trash, it's not lost money. Quote:
Quote:
|
Quote:
Don't misunderstand this - I'm not saying it's the best option, but one I imagine many will take. Quote:
Quote:
Quote:
Quote:
Hope this clarifies some points. |
Quote:
You may think I'm too unforgiving with some of these scenarios, but I think you're far too forgiving. And overly forgiving is how you get crap to view. I, for one, get tired of seeing bad quality on broadcast TV. Because more and more, broadcasters screw up interlace, fps conversion, etc. It often begins away from work, learning and doing wrong things to video. Raping video quality. The corporate world watches Youtube too, and searched Google, for better or worse/lazy. So let's help at least sustain a minimum decorum of quality. Not just whatever the Chinese want (low-rent devices, awful software, as sold to us dumb 'muricans/yanks and you your'o-pee-ins). After all, this topic is about trying to achieve the greatest possible quality, maybe, hopefully, someday. Not just cheap, easy, lazy, crappy. |
Quote:
We don't really have a culture of your public access (I know that's not the point you're trying to arrive at) so it's kept television broadcast in the hands of the 'big players', it's not a game you can get in to with a modest investment, apart from some seriously low-rent local cable channels that run under a special licence. I seriously digress, and I could be talking utter carp, this is only from tech' documents, what goes out could be dreadful - and almost undoubtedly is. At least you won't be put in prison for owning a television without an annual $220 tax like us Britbongs, then we get to see BBC programs broadcast globally for nix..... Great :D Anyway I've wandered well of topic but thanks for taking the time to reply, I'll post up more of this guide tonight. |
Quote:
You could save 90% of your writings here on digital.faq & videohelp.com and no one have to ask again for best hardware for video capture. |
Site design, images and content © 2002-2024 The Digital FAQ, www.digitalFAQ.com
Forum Software by vBulletin · Copyright © 2024 Jelsoft Enterprises Ltd.