Automating our home video imports | Hacker News Reader

Automating our home video imports

pierce.dev

86 pointsicyfox41 comments

How I digitized my family VHS tapes:

* I borrowed a good quality VHS player with SCART connector because it sends RGB in separate channels, improving quality considerably. Don't use the single channel composite video.

* Then I bought a cheap SCART to HDMI convertor and used a borrowed HDMI capture card.

* I recorded it with OBS studio and the resulting video looks very good.

So my total costs were about 20$ (for the adapter).

Me too, but I've used a good VHS player with HDMI output. So it's one transformation step less, and maybe the end quality is the best possible (I hope).

This also worked for me. Crucially, the cheap composite capture devices are rubbish and have terrible drivers as well, while the cheap HDMI capture + OBS Just Works.

There is the ultimate solution https://github.com/oyvindln/vhs-decode , but that requires modifying your VCR.

The best solution for "normies" is to get a Panasonic DVD player with HDD recording ability. That thing has a circuit which synchronizes all scanrows, i.e. it completely removes all tearing of the image.

Then proceed with something to digitize the analog output from SCART2.

A couple of models with this feature:

https://www.youtube.com/watch?v=6XW_-16Vo4E

https://www.youtube.com/watch?v=LaH73rhBHbk

This was helpful, thanks for sharing. Would you have a specific model recommendation if cost is not a concern and the use case is long term archival?

With these Panasonics, you either get the right model numbers, or. Tearing or no tearing.

The bigger difference is the VCR (you need one of those, too) and the capture device.

The capture device is a whole thing. See for instance Technology Connections on YouTube.

Depending on what you then capture with, you may need to attenuate the video signal with a potentiometer if the image is too bright and washed out.

Thanks, that's a wonderful rabbit hole! When I worked at a video capture card company in the nineties, our competitive advantage was using the raw capture from the same video capture chip that everyone used but bypassing it's decoder and doing that instead with custom assembly coded digital signal decoding. It's cool to see the same general approach that is open and even closer to the analog tape!

> I borrowed a good quality VHS player with SCART connector because it sends RGB in separate channels

I'd be surprised at that - normally they'd emit Y/C or at best YUV.

Scart can carry RGB and Composite. And good VHS players do RGB.

Never seen one which converts to RGB, must be rare. And then what do you use to digitize from RGB? Most SCART to HDMI converters use the composite video in the SCART connector.

SCART carries RGB but that only comes from DVD players, computers, and other digital sources. At no point in a VHS player is the signal in RGB form - generally not even in ones with digital "trick play" modes.

It doesn't make any sense to move the Y/C to RGB conversion into the VHS player.

> cheap SCART to HDMI convertor

From my understanding this is the "bottleneck" in quality for older systems (at least in gaming consoles), converting Analogue to Digital. Which is why "RetroTink" sells different converters from ~$100 up to $750 (RetroTINK-4K Pro). I've seen a few videos comparing cheap generic USB converters with more expensive upscalers and there is a noticeable difference in image quality

If you have a lot of Video8 or Hi8 tapes to digitise, get a Digital8 camcorder. It will most likely play them back quite happily and emit DV over its Firewire port. Digital8 is just DV on a different tape!

I still use DV/DVCAM tapes because I like shooting with old cameras, and I capture the same way I have for about 25 years when I used a VX2000 to shoot DV for a commercial digital streaming company that did all sorts of training videos.

Cheap crappy PCIe Firewire card (back in the day it was PCI, but no-one has that now), and dvgrab to get a raw DV stream off tape, then ffmpeg -i dvgrab-001.dv -c copy whateveritscalled.avi to rewrap it in something the editing software can read. These days I use DaVinci Resolve on Linux, in the olden days I used Premiere 5 on Windows 2000.

Even back then I used to capture on Linux and then bring it into Windows 2000 because only Linux had reliable Firewire support.

Can vouch for this trick - Digital8 cameras with Firewire is perfect transferring analog Video8 / Hi8. The analog to digital path in the Digital8 cameras is way better than what you can likely scrounge up yourself.

Can be expensive to get but if the camera doesn't break during use you can resell them easily.

I've ended up with three of the damn things, so if you are sufficiently desperate to dub your 8mm tapes that you're prepared to trail all the way up to NE Scotland I'd be happy to oblige.

I bought a Sony TRV120 10+ years ago, back when I was doing this conversion project for the first time. It's built like a tank and still works today.

At the risk of being smited by professional archivists, I'm willing to wager that 99.9% of people can't tell the difference between a $5k archival rig and one of these higher quality camcorders. At this point it really feels like the biggest inhibiter to good quality digitization is the decaying of the tape versus the archival setup.

That said - for anyone with the time, patience, and soldering abilities I would love a more proper A/B test with RF signal capture software encoding. Something like this:

https://rastrillo.ca/digitizing-video8-tapes-with-vhs-decode...

There be dragons - crazy things like taking several captures of the same tape and averaging the frames and such. :-)

I had a similar set of tapes, and ended up collecting a chain of connectors – firewire cable, firewire to thunderbolt2 adapter, thunderbolt2 to usb-c.

Instead of cobbling together an impressive array of tools though, I just got a trial of Final Cut Pro and pulled out everything with that. You can get what I think is a three month trial? Anyway, it was plenty for this one time effort of digitizing old Hi8 tapes.

I think I did end up using Handbrake to take the raws down to a reasonable size to give to family members, but the raw footage and project files I stuck on a couple of 1TB Sandisk drives to keep in physically separate backup locations.

No need for FCP, as iMovie can still capture DV streams from Firewire. I'd expect they both use the same implementation.

However, Apple has removed Firewire support entirely from macOS Tahoe so none of these solutions will work on Mac going forward.

I think it was needed for retaining the original aspect ratio, and capturing in ProRes, but maybe there's a workaround there that I didn't find.

Awesome to see this! I actually wrote PySceneDetect, was great to see it getting some use here. Would you be willing to share what parameters you were using? I'm curious why the accuracy was so low.

PySceneDetect only uses basic heuristic methods right now so it does require some degree of tuning to get things working for certain data sets. Your post inspired me to look into maybe integrating TransNetV2 as a detector in the future!

Nice to see you on here! I used the ContentDetector with a threshold of 27.0 and otherwise default parameters. Realize I could have done a grid sweep to really hone in on a good param range, but because I had only one input video labeled I wanted something that would work well enough out of the box. I imagine this dataset is rather... heterogenous.

If you happen to know a better apriori threshold I would be happy to re-run the analysis and update the chart.

If you're willing, could you try using AdaptiveDetector? It should have better defaults for handling fast camera movements a bit better.

The threshold values themselves can be tuned if you generate a statsfile and plot the result, but that can sometimes be tedious if you have a lot of files (thus the huge interest in methods like TransNetV2). Glad to see the real world applications of those in action. You can always just increase/decrease the threshold by 5-10% depending on if you find it's too sensitive or not sensitive enough as well.

Thanks for the response!

AdaptiveDetector definitely did a better job, will append these new stats to the post:

precision 0.397, recall 0.727, F1 0.513, mean temporal error 0.307 s

I'm still sitting on some 8mm film that my grandpa shot, and even some of the commercials he worked on. Still trying to find a reasonable way to digitize them.

Also multiple boxes of slides, I know you can buy the scanners (I had one). Wish there was a cheaper automated way hahaha

OP definitely used $4,000 worth of expertise plus a decent amount of time, despite automating the process.

It wouldn't have been crazy to spend $4,000 to have the professionals do it, so long as they produced a reasonably equivalent high-quality result.

What is the value for meaning for keeping home photos or videos? Does anyone provide meta tags, detailed information or narration? How is the photo or video meaningful unless you add your detailed memories to it for others to understand?

If the relevant relatives are still around but don't have the ability to play tapes anymore but want to watch them from time to time, you don't need any of that.

Most of my tapes did have pretty detailed narration and date overlays written directly to tape. But even without narration I still had luck doing basic event summarization and facial recognition of family members to build the tags.

Important archival work - much appreciated! However some links on the page are not working. Also, it seems like the author has made a Web app to make conversion easy, but I don't see a repo link or otherwise a way to access it.

Since the webapp is pretty opinionated to my setup (ie. linking against AVFoundation, using MPS for inference, always capturing an image after import) I didn't originally think it would be that useful to open source. Happy to do so - are you looking to get something specific out of it?

Got me thinking of a previous post regarding digitalizing VHS tapes

https://news.ycombinator.com/item?id=24839848

Good lord the misinformation here.

VHS is a composite signal on the tape itself. Composite for sake of this thread means black and white detail plus color information. S-VHS has higher bandwidth in the luma (detail) but the same limited color bandwidth. And there are two audio standards. But there is no "RGB out."

At a minimum you want a device with S-Video out (it keeps the two signals on separate wires). You also need a time base corrector. These come in two forms. One is line-based sometimes built into DVD players. This is how Jason Scott at Internet Archive does it and it's wrong.

The other form of corrector requires a separate box and corrects each frame in full. Many boxes claim to be time base correctors but are not. They are "synchronizers" or amps. Don't buy until you understand the differences.

There are two time sources (not really clocks) in a VCR. The first is physical tape wobbling and stretching over a head that's spinning far faster than seems possible. Line TBC is a tiny buffer that reconstructs the sync of the luma on each line.

The other timing source is the overall signal sync. A proper TBC reconstructs this overall sync on a frame by frame basis and presents something sane to the capture card. Without it you'll drop frames silently, audio falls out of sync, and all the other crap that happens when you try to watch video older than an iPhone. Consumer video capture is total crap and you won't see it until you try to encode, edit, or watch it on a different device. And then you'll be very confused working back to the original problem.

But follow this careful path where you actually capture a clean, proper signal and feed it into even the cheapest Blackmagic box and you're good.

ChatGPT will walk you through this and seems to know more about proper ffmpeg settings than the developers themselves or 30,000 conflicting StackOverflow messages on the topic.

You seem to know what you’re talking about. Could you recommend any devices that fit the requirements you outlined?

VHS-Decode did get a mention in these comments previously but probably beyond the reach (or patience) of even most hackers. But good information on decks and theory:

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

In the (original) spirit of this site, where people get curious and go off and solve impossible problems, I can point to this guy who gets it. He rants better than I do.

https://www.reddit.com/user/TheRealHarrypm

Read on for my rant if you care.

The last useful decks were built 25 years ago. Even then there's acres of bad advice on the internet, specifically around prosumer decks with built-in TBC that isn't sufficient to provide a proper signal to a modern video capture card.

So find any S-VHS deck that actually runs well, S-Video out into a external TBC. Blackmagic owns the cheap-but-great category for capture. Then ffmpeg to suit your storage budget.

Beyond that, things are so screwed up online around preservation technology and knowledge that my advice is to seek the services of a professional if the material is important. Or go ahead and DIY if you're just doing amateur stuff or you're a hoarder trying to justify your conditions. But try to refrain from posting your secret sauce; it likely isn't. Organize and post the videos instead. Most won't, because they know they look like shit and they can't handle the feedback.

I just did a quick Google search and found this on a formal preservation website: "This means that S-VHS tapes are playable in VHS decks but VHS tapes are not playable in S-VHS decks." https://psap.library.illinois.edu/collection-id-guide/videot...

Exactly backwards.

I don't think I've ever read an article about analog media of any type on the internet that didn't have me screaming at the screen.

I appreciate the reply. It does feel like most resources online are just the summary of 3 web searches and boom they’re an expert. Thanks for the links and info.

i do not have videos but i do have audio cassets which needs conversion to digital.

If someone knows a faster and good way pls share.

thanks

I do not have any good advice, only pain.

It turns out that a lot of old tape players have either failed completely or else add excessive background hissing.

Most tapes don't record all the way to the end so you will have to cut it digitally regardless.

I don't think there's going to be a better way than "find high quality player and decent capture device (e.g. Behringer), then press play a lot". It's possible that a truly dedicated process would capture off the tape head and then de-Dolby in software, but ultimately: it's a tape.

If you want more pain, you can correct for wow and flutter after capturing.

https://github.com/HENDRIX-ZT2/pyaudiorestoration

Key IMHO is "Speed matching to hum frequency".

There are also commercial products which does similar things.

https://www.celemony.com/en/capstan

https://forum.audacityteam.org/t/digitizing-tapes/62996/2

Is that webapp shared anywhere? I can’t find a link in the article.

Reply on news.ycombinator.com