My Grandma Was a Fed – Lessons from Digitizing Hours of Childhood | Hacker News Reader

My Grandma Was a Fed – Lessons from Digitizing Hours of Childhood

sampatt.com

210 pointsSamPatt85 comments

This is so neat! I really think AI turbocharges this kind of personal project way more than it speeds up programming for work:

> I was curious about the possibility of doing this myself, and I asked ChatGPT. Not surprisingly, it knew a lot of the various tapes, file formats, sizes, processing, storage, and after it asked some clarifying questions, it was quite optimistic about me being able to do this myself

Between this, it seems like it helped with so many different parts of the process:

1. Asking for how to do technical things, like transfer video from these old VHS to a newer computer.

2. Writing code for the web portal to host the videos.

3. Writing VLC plugins to help with data entry.

4. Transcribe audio into text.

Similarly, a coworker recently made a website that imitates what Alpha School does to incentivize his own kids to finish their homework all in the span of a weekend, and it's cool to think of the kinds of projects that less or minimally technical people can do with the help of ChatGPT to guide them.

Of course, the debugging techniques and the debugging and problem-solving techniques that you get from being a professional programmer helps a lot with taking what LLMs give you with a grain of salt, and knowing what they're good at and what they're not. But it is a superpower for sure.

> Of course, the debugging techniques and the debugging and problem-solving techniques that you get from being a professional programmer helps a lot with taking what LLMs give you with a grain of salt, and knowing what they're good at and what they're not. But it is a superpower for sure.

I'm really coming around to the idea for the lucky of us (and I'm assuming a lot about the average HN poster) AI really is a force-multiplying tool

I'd love to see the imitation of alphaschool . I would work on my own , and i'd love the inspiration . Any chance you can share ?

What is alpha school?

https://alpha.school/austin/

this is the OG, a few others are springing up. Not sure if i'm just drinking the kool aid, but this type of education just makes complete sense to me.

> I searched for services which offered to digitize Video8 tapes. Most services cost about $20 per tape. Even with discounts for bulk amounts, it would likely have cost about $2k! I considered paying it (how exactly do you value a few hundred hours of childhood video?) but then I noticed how they delivered the videos - a private media hosting solution for 60 days. I knew this would be a huge amount of data, and only giving me and my siblings two months wasn’t sufficient.

I'm not following here. Even if it was several terabytes of video (digitized at high resolution and minimal lossiness for archival purposes), that's plenty of time to download. Especially if you're a developer who can casually spin up a cloud or dedicated server to proxy through if need be? (And $2k sounds reasonable once you start going through "hundreds of hours" at a bare minimum, and again especially if you're a developer with real opportunity cost.)

Also, as far as the video analysis goes, Gemini might've been a better idea?

You don't even need to spin up anything - eg AWS datasync will mirror the remote server for you directly into glacier which can be extremely cheap if you're okay with access-a-file-12-hours-after-you-request-it

I'm not familiar with 'AWS datasync', but isn't the point of Glacier that it's really expensive and slow if you retrieve everything? It sounds like a bad idea for videos he planned to download in order to catalogue them. (He might want to delete them or could safely infer that 'no, neither me nor my siblings will want to watch this on Christmas Day', yes, but he still has to look at them first since he doesn't know what's on them at all.)

You're looking at the technical capability, not the human behavior.

A random non-techie human without particular urgency will not download a hundred old family videos in 60 days. They will watch some of the first one, try to fast forward, stop when it gets boring, and think about downloading them sometime in the future and then forget about it until six months later.

(Except your uncle who is the family historian, who tries to download all of them but runs out of room on day three.)

Yup, you nailed it. This was my exact concern.

As I techie human I’d probably do the same thing, thanks to ADHD

It's also just really really weird.

Every single digitization service I have ever known is a small business of one guy who sends you back DVDs full of files. It's not even a consideration by anyone that they could possibly hold your videos hostage like that.

It wasn't about having the videos held hostage, but about the ease of accessing and viewing them. I didn't want to do DVDs at all - frankly, that's just a bad experience to go through 200+ hours of unlabeled video.

Hi gwern!

My siblings are very much not developers. That's a lot of data for them to download, store, and figure out a way to view.

I was worried they'd just see a list of filenames and not put in the effort. By creating a streaming experience, I thought they'd actually watch them.

You might be correct that Gemini could have helped, I didn't test it, but much of the knowledge of who was in a scene, where it was, and why it would matter is inside my head. I doubt any model could effectively label locations and people over 20 years of video.

As to the opportunity cost - I'm currently looking for work, so mine is undoubtedly lower than yours!

> My siblings are very much not developers. That's a lot of data for them to download, store, and figure out a way to view.

I wasn't suggesting anything about your siblings, but you, who are a developer. I was just talking about the actual download step, not what you did after that. (Obviously you were going to host them somewhere else in some other form. Probably not DVDs but a little quickie website or maybe just a Flash drive with a HTML file index, say, I don't know, lots of options here to make it user-friendly for your siblings on Christmas Day. The hard drive or Flash drive idea has the benefit of LOCKSS, especially if you use up the spare space providing PAR2 FEC.)

> I doubt any model could effectively label locations and people over 20 years of video.

Actually, Gemini is highly promptable with a large context window and a single still image only takes up ~300 tokens IIRC, so I think that you could probably do so! Just include, say, 3 photos of each person over time with a natural language description, and 1 photo of each location, and that might be enough to get back useful labels. Gemini can even do bounding boxes. (Google is quite proud of its vision and video analysis capabilities.) And you can run multiple passes or split up videos etc.

Ah I understand you now. Yes I could have had a service do the digitizing then only done delivery myself. And given the time investment that probably would have been more sound. I don't think I'd do it all myself if I did it again.

I didn't know Gemini models were that capable. I admit I'm still skeptical about this approach though - even if it were capable of accurately labeling people and locations across decades, there's no way it could know when a scene is of personal interest. I kept a running log for each sibling as I was manually doing the labeling, knowing what they'd want to see, which presumably is only possible for me and my siblings to do with any accuracy.

If AI could ever do that then we've definitely hit ASI!

> I kept a running log for each sibling as I was manually doing the labeling, knowing what they'd want to see, which presumably is only possible for me and my siblings to do with any accuracy.

But you could feed that back in! Just write it down. It's all tokens. As you read over descriptions and note down key pieces of family history or per-sibling details, that provides information about better annotating the next video for possible points of interest. And you can chat with the LLM and write down more general principles. It's not like a LLM like Gemini doesn't know an enormous amount about family life and things of sentimental value, and can't make good initial guesses. And when you do this, you still haven't used up more than a small fraction of the context window with these image references and text profiles and principles...

"each MP4 file was likely to be about 3GB. (...) Also, it’s not easy to watch a file like that. You’d need to download the entire file first to even be able to watch it"

I dont think that's true? Even with moov atom at the end of the file, browsers just figure it out, as long as the server supports http range requests. And I suspect the size could be much smaller, given the low resolution of Video8 footage, but maybe author did the reasearch and compression would take too much time.

That depends on the specific MP4 I think. My camera creates MP4s which are completely unwatchable without the ending. I found out the hard way when the battery died while recording my children's school play.

The only way I could save the file was to create a new movie with the exact same camera settings (luckily I hadn't changed anything on the camera) and graft the ending of the newly created mp4 onto the old one using some special utility a hero on the internet had created.

In general, some tools like mplayer or VLC allow playing an MP4 file without an "index", but may require certain CLI arguments.

There are tools which can fix that too (was it mconvert from mplayer or ffmpeg, not sure).

This is not always true. If the 'index' is all that is missing the only thing is that you cannot seek. That's easily fixable. In the case of the files my camera writes a lot more info is only at the end. I can't remember what exactly but at least the resolution and the framerate were only stored at the end. And maybe even codec settings I can't remember.

I looked at it very thoroughly back then (it was a video I did not want to loose). Without grafted on info from a video with identical settings the video could only be played back as 8x8 squares of random noise.

"Once I began to create the metadata I already knew what I wanted to do: create my own streaming platform for our family’s memories."

Immich has become a bit of this for me, my father digitized everything he has, he gave it to me and I dumped it in Immich and spend some time dating stuff. So now I can, with the slide of the finger, drop myself in 1983 and see my own first steps, no audio, video8 quality, still quite magical.

Perhaps normal for all you iCloud and Google Photos fans, but I never wanted that, for me everything was just in folders on harddrives, have been waiting for Immich my whole life I guess ;)

Google Picasa used to be like that before obviously Google had to kill it and create a surveillance version.

Immich is a worthy successor.

Immich is great, after doing this project I installed it for our current family photos and it works well, although I keep hitting a bug where older photos are blurred and never load. I need to investigate that.

> I don’t remember exactly which tape I popped in, except that I saw my mother as she looked when she was around my age now. She passed away in 2013 at 55 years old, and I hadn’t seen her at this age in… well, I don’t know how many years.

Similarly, I found a tape I'd shot in the mid-to-late 80s, probably Christmas 1988. On it was some footage of my dad, who died in 1993 when he was 47.

That's the first time I've seen that tape or indeed heard his voice since he died. He must have been about ten years younger than I am now.

So, the first time I've heard my dad's voice in over 30 years, and the first time his grandson has ever heard him (he recognised his grandad straight away).

Quite a moment, that.

Thanks for the write up. I think you might be happy to hear that ffmpeg has scene detection built in. But you might be unhappy given how much work you did. In my experience it works pretty well.

An upscale pipeline seems like the next job, by the way! you can pull a lot of quality out of those old videos with modern tools. Enjoy.

Thanks for the tip! Part of the issue of automated scene detection was that there isn't any way to get the other metadata, such as who is in the scene or the date.

But I could have probably auto split the entire file at the start then just do that metadata per file. Although I found being able watch the entire video was quite nice.

Yeah I can imagine. I'd want to watch too.

@SamPatt Good post, including the clip of grandma's anecdote.

It would be a good idea to add a final step of burning the videos to M-disc. SSDs and spinning platter drives aren't reliable for long-term storage. You could use a tape drive if the file sizes are too large, but M-disc lasts longer and doesn't require pro hardware to read.

I assume the backup is the copy in R2.

That's better than nothing. Personally, I wouldn't consider it archival storage, so much as the possibility that 20 years from now Cloudflare (or a holding company) pays me $100 compensation for my lost data!

I have a NAS backup as well.

Just fyi, you can just use BDXL discs(get them while you still can!) they use technology identical to M-Discs and should last just as long.

I use M-disc and I am sure the discs will stay safe for a long time. What I worry about are the drives! It seems the business of making drives is not profitable. So companies exit or reduce.

The author managed to find a decades-old Digi8 camcorder in working order.

M-disc is readable by a standard DVD-ROM/Blu-Ray drive.

The industry has manufactured many, many more DVD-ROM/Blu-Ray drives than it ever made Digi8 camcorders, and they have fewer moving parts.

If you're concerned with finding an M-disc burner, I share the same concern.

If you're concerned with finding an M-disc reader, there's less reason to worry than with any other archival media formats.

This is one reason I'd like to see a fully Open Source hardware+firmware optical drive. Probably best to start with CD-ROM, but DVD might also be possible. The optical and mechanical parts seem relatively simple, especially when you're not optimizing for minimum cost or minimum size (meaning you could use the original Philips-style swing-arm mechanism). From what I can tell, the most complicated part is the signal processing, and with modern hardware that looks practical to do in software. I'm not sure how far you could get with home-scale DIY construction, but CDs worked with late 70s technology, so at least that far should be possible.

Thank you, and thanks for the suggestion, I'll look into that.

Neat! I briefly tried digitizing some old VHS tapes for my family. I was just planning on giving them the files, maybe putting it on iCloud, it's a much smaller collection (and I don't have a NAS already!). I did a few, the time investment was the biggest issue, as well as figuring out the right encoding so as to not take up a ton of space, but still be compatible with everyone's (Windows + Mac) native video players while preserving video quality.

On the professional end of digitizing old libraries, I've seen it all. They have cart systems that will automatically load a tape deck, the tape deck rewinds the tape, the system plays the tape while digitizing from the beginning until the end of the tape or end of control track. Once it is done, the tape rewinds, ejects and the robot removes the ejected tape and starts over. This is great for cassette based formats. For reel based formats it still involves an operator threading the tape and changing the reels. Even with the automation, it is a chore.

You could look into one of those VHS to DVD systems. Sure, it's SD MPEG2, but the source was a VHS. At least you're not tying up your computer system to do this. It also means not needing a NAS.

I film my kids with the same camera in 2026. I don't know how to explain it, but it feels completely different. When I film something on a phone I feel awkward but with this camera I could film everyone and it's much more comfortable to all (including adults I mean)

If they're still quite small, put a radio mike on them (or one of the ones you get with recording built-in that you sync up later, I guess, is easier these days) and stand well back and let them play.

That's when you get the good stuff.

A couple of years ago when my son was a chatty, fast-moving toddler, his granny couldn't really follow what he was doing because she's a bit deaf and not as quick on her feet these days. Take him down to the park, mike him up, let him run around, stand well off with a long lens.

Also, because that's a Digital8 camera, it'll output analogue tapes over FireWire as described in the article. It's worth doing this even for Video8 because the output is so much cleaner than with capturing over composite.

Oh, that's unexpected idea. I hope I'll try it one day, thank you.

I use FireWire as well. Spent a month connecting it, that was a devastating part. It turned out, only IEE1394 cards with TI chips works with Sony cameras. And only some cables. But to me, result is worth it.

I'm using a VIA VT6306 based card for mine, cheapest one on mz*n at about a fiver.

It will only detect the camera if the camera is plugged in and turned on when you turn on the PC, and this is consistent across various Sony, JVC, and Panasonic cameras.

In the early 2000s, we had a camera that saved to an internal hard drive. The only way to watch videos was to either copy them to the computer or hook an RCA cable from the camera to the TV. You could also go from VGA to RCA with the right set of cables.

However, everyone did have a DVD player! So I, similar to the author, wrote scripts to take videos, generate DVD isos, and then burn to DVDs.

I learned about message queues (rabbitmq) with that project and had connected a bunch of old laptops with Linux VMs installed.

I never finished the project and nowadays there are a hundred ways to share and stream digital video. I hadn’t anticipated, at the time, that casting videos wirelessly to our TVs would become the norm.

Way too much work :( At least my video collection is mostly miniDV.

Unfortunately, the author spent so much time just ingesting and managing the video that he did not get to the fun stuff you can do in 2026: index, query, and restore it.

Capturing off Video8 is no different to capturing off MiniDV, if you use a Digital8 camera.

I shoot on full-size DVCAM as well as HDV occasionally. Yes, it's 576i or 1080i depending on which you use, but you'd be surprised how good that looks with a decent lens and a bit of care in shooting.

The bad old days of de-interlacing and post-production with avisynth. I've shot and watched more videos since cell phones, thanks to the immensely improved UX.

I use ffmpeg. I had a play with avisynth years ago but it hasn't been updated in about a decade.

Wow, Great story. The title was boring, but I'm procrastinating and decide to read on. It's amazing how LLM is supercharging our capabilities. I was just telling my wife that we need to record more videos of the kids. I use my phone but it get's filled up fast, so this has inspired me to buy a dedicated camera for video recording of the family. Thanks for sharing.

Glad you enjoyed it. Definitely get some good videos.

Non-native speaker here.

I always assumed Fed has something to do with FBI or the Federal Bank. Please, can someone explain me what Fed means in this context?

Thank you in advance

Yes, based on context fro. a quick skim of the content, it sounds like its a jokey reference to working the FBI phone switchboard

Thank you, that's very appreciated!

In the context of my childhood, feds were everywhere:

From the founder of Costco, FedMart was a department store chain and the predecessor to Target

Federal Express Corporation was founded in the early 1970s and, of course, shortened up their name. Don't miss the arrow in their logo!

There were some consumer electronics stores named "Federated" and their commercials featured a mascot named "Fred Rated"

Institutions such as banks sometimes took names like "First Federal" which could give an impression of being a government agency, but probably referred to their organizational structure or nationwide presence. Different from the Federal Reserve and the Federal Reserve Banks.

It can mean "cops" nowadays: https://en.wikipedia.org/wiki/Federales

It's a joke; my grandma worked as a switchboard operator in the 1940s for two weeks. As she says, she "disconnected half of Washington DC."

We often refer to government agents as "feds." I wasn't considering an international audience with my title, apologies!

This is SUCH a great story. Thanks for writing up both the human and technology parts with equal love and depth.

I'm glad you enjoyed it!

I posted this 5 days ago, and it didn't seem like anyone saw it then, so I'm happy folks are enjoying it now.

Not related to the tech bits of this, but I finally got around to watching Aftersun a couple of days ago. It's a great, sad film about somebody watching home video from their childhood and reevaluating what was going on.

I'm in the process of doing something similar but just planning on throwing them on my Immich instance once they're ready (and that lets me share them with other people as well with the Immich account management).

I would absolutely try that first.

Love this story too as many others have said. Thanks for sharing this. Maybe someday I will do the same. Luckily our archive of tapes isn’t so large.

Like the author I've been in the process of archiving family memories since my parents passed away. In my Dad's case his early fascination with super 8mm film gave way to a lifelong quest to own more Kodak slide carousels than any other human. It's sort of an odd place to be, when you're in possession of so much that _seems_ emotionally, viscerally important, but that ordinary people living their everyday lives take very little actual interest in. I scanned all the slides with a Kodak thing I bought on Marketplace. There were more than 6500 of them and the archive has passed 20GB in size. Now I have a stack of plastic kodak carousels taller than I am and thousands of slides that I can't just put in the trash because it feels wrong.

I'm working on loose documents and photos now with a great Epson flatbed scanner that I've had for years. It's something of an obsession for me now: to finish the job and tie their lives and a good chunk of ours off neatly. When I'm done I'll pack it all onto thumb drives and send copies to my siblings and they will look at a few of them and then put the drives in a drawer and that will likely be the end of it. But I will have done my duty to the old folks, even though it took long enough that I became one. It has all reminded me of how I mourned when I lost a hard drive years ago, and then came to the realization that I hadn't looked at any of its contents in years and that if I lived to be 100 probably none of it would ever have been important again anyway. You can't really keep anything, so do your kids a favor and dispose of your stuff while you can :).

So cool! Gave me some motivation to record more things.

That was an amazing read, thanks for sharing!

The AI folks are really getting desperate to be churning out this slop.

A Linux user who’d never installed VLC was weird enough, but the part where they recreate youtube from first principles really strains credulity.

So what?

They achieved their goal of digitising their family videos and allowing their siblings to view them in a coherent way, which they very likely would not have been able to accomplish without the help of AI. Not without like triple the time investment.

They aren't releasing a product here, it's bespoke software which serves the exact purpose they need it to. This is exactly what AI is good for.

Bro, you're paranoid. I'm the author. Look at my blog. I write my stories myself. I wrote a geoguessr article last year that hit the top of HN.

Yes, I've used VLC. But I'm using a fairly new machine and hadn't installed it yet.

Why does it strain credulity to make this myself? I explain it step by step. I enjoy using AI to build stuff myself.

You are so amazing, and you are so lucky to have found this. I read it and cry because I will never have such an experience...

I'm sorry to hear that. I agree. I'm fortunate. I hope that you can enjoy the memories as best you can.

The post is impressive but anyone else feel a bit icky about the comments here? Numbering out the advantages in lists? The phrasing, the exciting exclamation mark(!). This comment section feels like some kind of marketing exercise.

The post has 8 comments at the time of your comment and they look pretty organic for established accounts.

There are now organic looking comments. Maybe it's just the way some people sound.

I asked ChatGPT and it told me it was marketing.

But what is it marketing, though? Buying 40-year-old cameras in car parks? I already do that, no need to sell me on it.

It's the Big Video8 cartel astroturfing.

DVCAM4LYFE

https://ishootdv.cam/

Really? I'm the author and I can't conceive of any way this could be considered marketing. It's a personal project, I'm selling nothing.

But I am looking for work, so hey, if anyone is hiring and they enjoyed the post, then email me.

Maybe you could ask ChatGPT how to get more gigs saying ChatGPT every third paragraph.

I'll have to try that next time

Reply on news.ycombinator.com