#1  
02-08-2019, 06:01 PM
hodgey hodgey is online now
Free Member
 
Join Date: Dec 2017
Location: Norway
Posts: 1,472
Thanked 404 Times in 344 Posts
I thought some of you may be interested in this.

As have been mentioned before on these forums, the people behind the domesday86 project have been working on decoding the rf signal off laserdiscs for a while now. I've restarted work on using the software to do the same with VHS (and maybe other tape formats eventually). It's still very much early stage work in progress, though I got some interesting results already.

frame_pal_source_1.jpg

https://github.com/happycube/ld-decode/issues/16


Reply With Quote
The following users thank hodgey for this useful post: sevarre (03-09-2020), wimvs (02-10-2019)
Someday, 12:01 PM
admin's Avatar
Ads / Sponsors
 
Join Date: ∞
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
02-08-2019, 06:20 PM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 12,161
Thanked 2,246 Times in 1,929 Posts
I don't know what value this has. Explain.

Offhand, I have to wonder if access to the "full" signal might allow software-based timebase correction. Other than that (far-fetched?) idea, I'm not sure why it would be done.

For LD, isn't it essentially allowed LD ripping? (True ripping, not "ripping" misused to mean a capture.)

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
  #3  
02-09-2019, 08:31 PM
hodgey hodgey is online now
Free Member
 
Join Date: Dec 2017
Location: Norway
Posts: 1,472
Thanked 404 Times in 344 Posts
Software-based timebase correction is indeed one advantage of this, as the decoder will has access to the same signal as a VCR or capture device would. The ld-decode software already does correct the horizontal sync to a degree, but as you can see it's a bit inaccurate on the VHS capture decode as of now. You can also do other stuff like dropout compensation (which is partly implemented for the laserdiscs) as well.

Doing it in software means it doesn't have the limitations of a decoder chip that has to do all this with very limited memory space within a very short time, and it can be tweaked to suit each individual capture in post. Also it means that you are not as reliant on needing the top tier SVHS decks with all the processing bells and whistles, as you are bypassing the player processing entirely.

As for the reasoning behind the project for capturing laserdiscs, you can see the FAQ on their website. I suppose you could use the term ripping, The Domesday discs, and many other laserdiscs utilized some of the unique features of the format to do more than just straight video playback. There is also the fact that Laserdiscs don't store luma and chroma separately like VHS (and other tape formats) so having excellent comb filtering is very crucial, which further complicate capturing the "traditional" way.
Reply With Quote
  #4  
02-10-2019, 01:36 PM
latreche34 latreche34 is offline
Free Member
 
Join Date: Dec 2015
Location: USA
Posts: 2,446
Thanked 428 Times in 396 Posts
Actually the VHS one is more important than laser disc, very few laser discs didn't make it to DVD or Blu-ray so not much interest there when you can just get the HD version of the movie. But for home videos this sounds like a good ideas with great potential for better quality. Especially if the process involves one digitization only vs 3 with the traditional workflow assuming the line and full frame TBC's are used.

Last edited by latreche34; 02-10-2019 at 01:57 PM.
Reply With Quote
  #5  
02-10-2019, 04:32 PM
jwillis84 jwillis84 is offline
Free Member
 
Join Date: Aug 2017
Location: College Station, TX
Posts: 797
Thanked 210 Times in 170 Posts
This looks like software defined radio in a FPGA chip, driven completely by a firmware upload to the FPGA on boot up. Though I suppose we're getting to the point where CPUs have enough cores and speed to do it by pinning a task off to one part of one core while the rest of the operating system runs on the rest of the CPU.

The trick then becomes one of a proper signal conditioner and an Analog to Digital converter.. which can also be defined in silicon by software.. or as a stock add-on in many system on chip with bundled "peripherals" like the CPU in a cell phone. A CCD camera chip is after all a very simple ADC with many many converters in row and column arrays. Its not like the old days when you needed an analog circuit tied to a photodiode to generate a single "bit" of data.. now you have acres of bits and a flood of digital data to work with in software.
Reply With Quote
  #6  
02-14-2019, 03:03 PM
captainvic captainvic is offline
Premium Member
 
Join Date: Sep 2018
Posts: 43
Thanked 6 Times in 5 Posts
Quote:
Originally Posted by latreche34 View Post
Actually the VHS one is more important than laser disc, very few laser discs didn't make it to DVD or Blu-ray so not much interest there when you can just get the HD version of the movie. But for home videos this sounds like a good ideas with great potential for better quality. Especially if the process involves one digitization only vs 3 with the traditional workflow assuming the line and full frame TBC's are used.
Yes, this process could be very important for VHS sources, but to say "very few laser discs didn't make it to DVD or Blu-ray" is not wholly accurate. Approximately 1/3 of my laserdisc collection has yet to be released on DVD or BluRay, probably because most of my laserdiscs are not big-name Hollywood movies.

This process could be important for a variety of analog sources. Thanks for sharing! I find the Domesday Duplicator project to be fascinating.
Reply With Quote
  #7  
03-09-2020, 04:49 PM
Infrid Infrid is offline
Free Member
 
Join Date: May 2012
Posts: 4
Thanked 0 Times in 0 Posts
Quote:
I don't know what value this has. Explain.

Offhand, I have to wonder if access to the "full" signal might allow software-based timebase correction. Other than that (far-fetched?) idea, I'm not sure why it would be done.
Don't underestimate the importance of such project, the ability to sample the waveform, save it and process (or reprocess) later is such a huge advancement that can revolutionise how we preserve analogue footage.

A decent VRC will be required, but you can avoid the purchase of an expensive TBC, outdated capture card and complicated driver/capture software setup.

With an high sampling frequency and bit depth you can be more precise on how the colour is filtered and decoded. Not to mention the software time base correction, to me it's a godsend.

The ability to re-process the sampled footage when a new version of the decoder is out, it's simply amazing. A new denoise algorithm is out? just replay the file! A more precise TBC? replay!

If I can sample all the signals from my VHS collection, I won't worry too much about rotting and degradation of the media.

I captured lots of VHS using my limited equipment, dropped frames, A/V sync problems, unstable image , etc etc. I could solve those problems with an investment of 1-2k, but I don't want to spent that money on the equipment.

Projects like this can cut down the cost of digitalising your VHS collection, I am looking forward to this. I really do.
Reply With Quote
  #8  
03-09-2020, 09:13 PM
sevarre sevarre is offline
Premium Member
 
Join Date: Dec 2016
Location: Minneapolis
Posts: 31
Thanked 9 Times in 6 Posts
Yes I am also following this closely. The documentation on the LD decode project is nice and gives a nice tutorial on building your own DAQ system (think it uses a 40 Mega Samples / second ADC chip). I am checking out some old used NI products that are marketed as "digitizers" but which are basically a DAQ with a couple GS/s and see if I can use one of those instead of building the the thing on the LD Decode project (although that looks like a fun/reasonably easy task given the documentation).

I don't know anything about the RF signal that comes off the tape head, though. Have to research that. I know all about the NTSC composite video signal, but I am guessing the signal coming off the head amp is not simply just that signal. I believe this project is particularly useful for laserdisc because they can sometimes contain auxiliary information that is not simply audiovisual (like how teletext is contained in the vertical blanking interval, maybe?).

The question I have for VHS is whether or not software can extract the audiovisual information from the raw RF head amp signal as well as the electronics in the AG1980P can.
Reply With Quote
  #9  
03-09-2020, 09:48 PM
NJRoadfan NJRoadfan is offline
Premium Member
 
Join Date: Sep 2010
Posts: 1,135
Thanked 351 Times in 287 Posts
This PDF gives an overview of how the signal is stored on tape: https://www.av-iq.com/avcat/images/d...20-%204611.pdf
Reply With Quote
The following users thank NJRoadfan for this useful post: sevarre (03-13-2020)
  #10  
03-10-2020, 12:57 AM
SFtheGreat SFtheGreat is offline
Free Member
 
Join Date: Sep 2015
Posts: 174
Thanked 4 Times in 4 Posts
Quote:
Originally Posted by sevarre View Post
I don't know anything about the RF signal that comes off the tape head, though. Have to research that. I know all about the NTSC composite video signal, but I am guessing the signal coming off the head amp is not simply just that signal. I believe this project is particularly useful for laserdisc because they can sometimes contain auxiliary information that is not simply audiovisual (like how teletext is contained in the vertical blanking interval, maybe?).
NYSC VHS also had closed captions, for years I wondered how to play my canadian Star Trek on a PAL VCR and see the text.

Also, the ability to tap into the directs signal off the tape before the electronics interpret it according to being either PAL or NTSC would mean that any system will be able to read any system, as long as cotrol track and tracking are fine, simply grab the RF before it is converted to CVBS and audio signal.

And regarding audio, the Hi-Fi stereo needs to be interpreted from the wave then.
Reply With Quote
  #11  
12-21-2020, 03:06 PM
Kaos-Industries Kaos-Industries is offline
Free Member
 
Join Date: Dec 2020
Posts: 13
Thanked 0 Times in 0 Posts
What's the latest on this? Dedicated TBC devices are ridiculously expensive and if this something like this matures it would really democratise the process of good-quality VHS video capture.
Reply With Quote
  #12  
12-22-2020, 04:55 AM
lingyi lingyi is offline
Premium Member
 
Join Date: May 2011
Posts: 320
Thanked 125 Times in 101 Posts
Most current progress is discussed here: https://forum.videohelp.com/threads/...nals%29/page13

Lots of impressive progress in the last year and a half, but probably still years away from a practical kit.
Reply With Quote
The following users thank lingyi for this useful post: Kaos-Industries (12-22-2020)
  #13  
12-22-2020, 04:57 AM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 12,161
Thanked 2,246 Times in 1,929 Posts
Quote:
Originally Posted by Kaos-Industries View Post
What's the latest on this? Dedicated TBC devices are ridiculously expensive and if this something like this matures it would really democratise the process of good-quality VHS video capture.
The RF project has no overlap with TBC.

Yes, this project is neat. But I can say the same about lots of failed projects, lots of vaporware. I want to encourage developers, but at the same time casual users need a bucket of cold water thrown on them. Something like this won't be coming soon -- and maybe never. This particular project still has special hardware requirements, so don't start thinking it's some sort of magic software-only solution. It's not.

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
  #14  
12-22-2020, 02:28 PM
latreche34 latreche34 is offline
Free Member
 
Join Date: Dec 2015
Location: USA
Posts: 2,446
Thanked 428 Times in 396 Posts
I started to think that way too, There was this guy on youtube that was trying to make a software PCM encoder to record digital audio on a VHS tape as video, so I responded to him and told him that this is already out there and gave him the github link so he said oh ok cool didn't know that. He seem to be knowledgeable, So I told him instead of wasting time on PCM software why don't you come join the VHS decode project team, he came over and gave it a little push but now he is gone (so far).

There are two or three active members currently working on the project but the progress is dragging.

Even the Laser decode project is not a ready product yet, you would have to have the skills to build and run such hardware and software.
Reply With Quote
  #15  
12-22-2020, 03:12 PM
hodgey hodgey is online now
Free Member
 
Join Date: Dec 2017
Location: Norway
Posts: 1,472
Thanked 404 Times in 344 Posts
Yeah things are progressing, though very slowly at the moment, though hopefully I can spend a bit more time on implementing stuff after christmas this year.
Reply With Quote
  #16  
07-25-2022, 12:38 PM
latreche34 latreche34 is offline
Free Member
 
Join Date: Dec 2015
Location: USA
Posts: 2,446
Thanked 428 Times in 396 Posts
We abused another thread by going way off the rail so I though this is the right thread to discuss the matter.
I want to respond to harrypm from his last post on that thread:

Quote:
Originally Posted by harrypm View Post
However both lordsmurf & latreche34 are 10-13 months out of the loop the VideoHelp Thread is a legacy userbase mostly from the LaserDisc side and is not using current documentation or methods.
No, I follow the progress of some members work, Besides if you think that thread is outdated why don't you update it with your progress.

Quote:
Originally Posted by harrypm View Post
Hell, not even a single mention about CVBS decode which does what a traditional TBC setup does use that baseband signal out the back to a digital file no need for the original source signal no need for control over the demodulation process just run whatever you get though the TBC code and that's that, wait ain't this like apples to oranges huh.
Quite few actually, but why decode CVBS when I can get better captures over S-Video using the conventional UYV capture.

Quote:
Originally Posted by harrypm View Post
Anyone can pick up CX Cards and install CXADC on any cheap desktop and capture this data off any colour under tape format VCR with test points.

That is the 1:1 copy, saved today to be decoded today or tomorrow.

Tape decoding is after the RF of the medium is captured and saved, not during the decoding process.

That's a hard reality.
No, not a reality at all, Just because the hardware is affordable, doesn't mean anyone can connect it, set it up and knows how to troubleshoot it using Linux and script commends. So just because I'm a mechanical engineer I expect everyone to fully understand how mechanical processes work or design a product? that's absurd.

Quote:
Originally Posted by harrypm View Post
Then categorically prove it with one simple act RF capture your "deserves the best" tapes today, decode them on this week's release, then re-decode them in 1-5 years or more or sit on them and let them demagnetise in the background of solar flairs increasing seeing is believing so go see for yourself how far its come
I said willing to support the project financially, I don't have time for tweaking and fiddling with Linux or even Windows if a version exists. And no I don't have tapes sitting to demagnetize.

Quote:
Originally Posted by harrypm View Post
Decode wins today with conventional alongside for hardware reference, It's allowing one-and-done archival today, to the consumer it's the affordable option to the equipment scalping community its death, to the restorative and archival communities it's the first analogue toolset foundation with a value equivalent of FFMPEG in the digital realm.
Wining what? there is a handful of people are tinkering with said gadget and we can count them if we try, We are talking about the masses who don't have knowledge with electronics, software troubleshooting and scripting, Linux ...etc As much as I want to see it winning I don't see how few guys working on an experiment is winning. By that analogy you know who's winning too?
Digital Audio Tape recovery using a DDS drive, Working for me so it must be winning. No, the masses are still using expensive DAT home decks and digital cable.
VHS PCM recovery and recording, Working for me so it must be winning. No, Archivists still using expensive Sony PCM adapters.

I can go on and on but this shows that you have no grasp of reality, just because some of us have some ways of doing things differently doesn't make them a marketable product that everyone can buy and use.
Reply With Quote
  #17  
07-25-2022, 03:22 PM
RobustReviews RobustReviews is offline
Free Member
 
Join Date: Oct 2020
Location: London - UK
Posts: 569
Thanked 86 Times in 75 Posts
Quote:
Originally Posted by latreche34 View Post
Quite few actually, but why decode CVBS when I can get better captures over S-Video using the conventional UYV capture.
A big point it's important not to overlook here is that not all formats offered any sort of YC output. Video2000 and PAL Betamax (I'm not opening that one again if it does exist, it's vanishingly rare) are two examples.

Part of the issue with VHS-Decode I personally think is the name, it's a lot more than just VHS.

V2000 was never released in NTSC markets, but it did take a chunk of the PAL market so that's one of the cases where CVBS is useful. It had died by 1984, so well before YC outputs were adopted on PAL VHS, also the PAL markets had N1500/N1700/SVC etc which were abandoned before VHS was widely adopted.

The Western Europe PAL markets have videotapes (although, I fully admit, uncommon) that exist well before VHS & Betamax. Technically the Japanese formats were the 2nd/3rd generation home recording formats in this market. That's counting Telcan as the first domestic system (1963) -> The Philips/Grundig N' systems (1972) -> Betamax/VHS/V2000 (1977 onwards).

So, yes, very high quality CVBS capture is appreciated maybe more here.
Reply With Quote
  #18  
07-25-2022, 06:28 PM
latreche34 latreche34 is offline
Free Member
 
Join Date: Dec 2015
Location: USA
Posts: 2,446
Thanked 428 Times in 396 Posts
Quote:
Originally Posted by RobustReviews View Post
A big point it's important not to overlook here is that not all formats offered any sort of YC output. Video2000 and PAL Betamax (I'm not opening that one again if it does exist, it's vanishingly rare) are two examples.
Isn't it what laser-decode is for?
Reply With Quote
  #19  
07-27-2022, 06:58 AM
lordsmurf's Avatar
lordsmurf lordsmurf is offline
Site Staff | Video
 
Join Date: Dec 2002
Posts: 12,161
Thanked 2,246 Times in 1,929 Posts
Quote:
Originally Posted by latreche34 View Post
No, not a reality at all, Just because the hardware is affordable, doesn't mean anyone can connect it, set it up and knows how to troubleshoot it using Linux and script commends. So just because I'm a mechanical engineer I expect everyone to fully understand how mechanical processes work or design a product? that's absurd.
This needs to be quoted, because it's true. You can buy sports car kits to retrofit a classic car as DIY project, but it's for a tiny niche. Most people that want a sports car just buy it. Or don't buy it. They don't DIY it. This is really comparable to buying quality hardware, vs. essentially making it.

Quote:
And no I don't have tapes sitting to demagnetize.
These tapes myths always piss me off. Yes, tape longevity is a concerns, certain issues are real. But there is far more FUD than not. Stuff like "tapes sitting demagnetizing" is cow plop.

Quote:
We are talking about the masses who don't have knowledge with electronics, software troubleshooting and scripting, Linux ...etc
Simply refer to how hard it seems to be to get some users to use superior freeware Avisynth, vs. paying for vastly inferior software like Topaz, NeatVideo, etc. Add a hardware layer on top of that, and good luck. I don't see even serious users wanting to venture beyond Avisynth, which has already tested their patience.

Quote:
I can go on and on but this shows that you have no grasp of reality, just because some of us have some ways of doing things differently doesn't make them a marketable product that everyone can buy and use.
Again, realism. I'm a realist. Not a cheerleader. I don't have my head in the clouds (or in darker danker places). Techs, IT, has a significant problem communicating effectively with the masses, even minority audiences that are amenable to the topic. This is why so much tech utterly fails, even if superior. But also realizing that many think their non-superior (just different) tech is better than it actually is. Or will be bigger than the market supports.

Quote:
Originally Posted by RobustReviews View Post
A big point it's important not to overlook here is that not all formats offered any sort of YC output. Video2000 and PAL Betamax (I'm not opening that one again if it does exist, it's vanishingly rare) are two examples.
This is something I mentioned years ago now, and is a great point. I'm glad somebody else sees it. (I'd actually completely forgotten about this, it gets lost with all the VHS talk.) VHS has both unique and non-unique properties and concerns. The unique attributes being committed to the project really does take away from formats that have no good hardware, with different issues to contend with. Leave VHS alone, the already have (probably) the best gear already, and that ever will be. The most vhs-decode offers is truer sharpness to the signal.

The main thrust (at least to many, most) seems to be cheapness, not quality. However, that cheapness carries negative economics and ROI, so it's really dumb. The more realist devs of the project are not doing this for cheapness. As I mentioned elsewhere, until the projects shakes out the cheap/stupid, it will flounder. I've seen this before. It may need to fork before it gets mature. Or if. Cheapness will create a "good enough" mantra (ie, not good, still many issues), while the quality seekers will grow disgruntled. Fork.

Quote:
(although, I fully admit, uncommon)
This is a reason that certain statements like "the restoration community" from certain devs are laughably absurd. The actual community is not worried about VHS whatsoever. Most of us have vast collections of gear, and the methods from a decade ago still apply. Our main worry is keeping the gear working properly (mostly doing timely proper maintenance, learning some repair), as well as tape issues like oxide shedding. Not squeezing out a small % of sharpness, with a negative time ROI to crap around with cobbled setups.

But we are worried about rarer formats. We'll do whatever is necessary to salvage. Again, it's very much like handling old undeveloped films. It's be costly, very manual. And not something we'd mess with if paths of lesser resistance were available (aka, traditional VHS transfer workflows).

Quote:
So, yes, very high quality
Sometimes people accuse me of having overly high standards (which is mostly false, I only have a baseline standard, somewhat binary, crap vs. not crap). Using a viable vhs-decode would be an example of overly high quality, in attempting to eek out a tiny sliver of more sharpness from a tape. I have certain tapes that I transfer, and re-transfer, in the hopes of better quality each time. But I must say, the curve always gets larger. vhs-decode doesn't have much to offer, and in fact would mostly be worse, aside from sharpness. That's the problem. It needs to actually be fully superior, not trade issues.

Quote:
Originally Posted by latreche34 View Post
Isn't it what laser-decode is for?
ld-decode is for lasersdiscs
cvbs-decode is a composite variant, like vhs-decode for Y/C. And it's really rough right now, not really even as usable as vhs-decode is.

BTW, the "cvbs" is a giveaway that it's yet another very PAL heavy project. That acronym was rarely used in North America, but more heavily in Europe and Asia. While that's great for PAL users, there are enough differences between PAL and NTSC that NTSC is getting less attention. Sometimes discussing NTSC is talking to wall, in various discussions online.

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
  #20  
07-27-2022, 01:54 PM
latreche34 latreche34 is offline
Free Member
 
Join Date: Dec 2015
Location: USA
Posts: 2,446
Thanked 428 Times in 396 Posts
I was told that laser-decode can capture any composite in nature format, I understand how CVBS decode works, I was just making a point that if ld-decode is ready why bother with CVBS decode. Anyway, I've seen CVBS decodes samples and I can assure you a good FPGA composite capture device like those from Grass Valley or the previous model of Sing-Mai can do a better job, the comb filtering in those devices is so advanced that even implementing any similar technique in software would not yield that much more of quality difference, remember composite is a hot mess.

I would argue that vhs-decode (not CVBS decode) is the best method for CVBS only formats, because it avoids the composite stage all together. Not so much for Y-C formats, the difference in quality from Y-C out vs RF is not worth the hassle right now until it is stable and finalized as a working product.

https://www.youtube.com/user/latoak34/videos
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
SDI signal to multiple monitors? ottoj Videography: Cameras, TVs and Players 1 12-13-2017 06:26 PM
All of a sudden, I get no video signal? jobeis Capture, Record, Transfer 2 01-09-2017 02:11 PM
Difficulty playing (decoding?) MP4 download Reading Bug Edit Video, Audio 5 07-20-2014 06:05 AM
No TV signal under MMC 9.14 wit ATI AIW X1900 Mejnour Capture, Record, Transfer 13 02-12-2014 09:22 AM
Do I lose PQ advantage switching signal from s-video to RCA in signal path ? Mejnour Project Planning, Workflows 4 02-01-2012 05:12 AM

Thread Tools



 
All times are GMT -5. The time now is 06:14 PM