How to archive video for future use?
 

Hello again everyone,

Couple of weeks ago I was finally able to Capture a full length VHS into Avi (HuffYuv) using VC500, Windows XP and LG DVD. Wanting to learn more, I was reading about Color correction, Avisynth filters to improve quality, encoding, clean up etc. The more I sank time into research, the more I understand there's no "Easy" button, and proper restoration will takes a lot of research, experience and time. And sadly - time is something I'm struggling with lately (my first child was born a while ago). So I decided to put this little project on the side for a while, at least until things will settle down here.

Another thing I learned from you guys, is that while I expected VHS converting to be easy in 2020, it's actually getting harder with time. Drivers that doesn't work with Windows 10, hard to find Capture devices and Playback devices etc. There's a good chance one day this XP machine I own will die, or even worse - the VHS I own. So while I do want to focus on this project later on - it feels time is of the essence. So instead of completely dropping the project, I plan to Capture all the tape media I own with the current setup, and save the digital output for later use. In case the older Hardware will die, I will still have the digital files to work with.

So the plan is to capture every VHS tape I own with my current setup (I am following Sanlyn guide for VirtualDub by the way). All videos will be be kept in HuffYuv format.
Yes, I know - that's a LOT of space I will be needing. A 2:30 Video takes around 65G. I guess I will have to invest in a big storage device. But I think it's worth the extra case, to have the ability to improve/clean the content later on.

By the way, Gspot mention I have 8.6 GB of unneeded bytes at the end? is it possible to trim the HuffYuv format further?

I will of-course keep the original tapes and setup. So in case in the future some breakthrough will happen, I will be able to perhaps re-capture the media in a better way.

Is this a valid plan? or am I'm missing here something?

Watching a 65G video file, is a problem. I plan that with every HuffYuvv file, I'll have a smaller compressed file I can watch and stream. I have a question regarding it - but that's will be on another topic for the sake of order.

Have you test to capture to FFV1 lossless codec i use Virtual Dub 2 and capture to FFV1 it works very good for me so i do it

but many here say and recommend to use huffyuv or lagarith and there are some other lossless codec to


or convert huffyuv to FFV1 for archiving takes less storage space

Yes. It is easy to do in Virtual Dub. You set the start point and the end point with the black arrows at the bottom of the preview window. The area that will be "exported" will be shown in blue along the timeline. Then simply set the Video menu to "Direct Stream Copy" and then do a File Save As AVI. You can even set up a batch-type export of multiple sections: after you have set up the segment you want to export, use File>Queue Batch Operation. For example, if you have a 3 hour tape with 5 subjects on it, set the batch operation up for the 5 exports and let it run overnight (although even on my slow old i5-720, Direct Stream Copy runs at about 80 frames per second as it creates the new file).

FFV1 and Lagarith induce dropped frames at capture, so often not good for capture.

When keeping lossless long-term, I generally normal recompress to Lagarith (YUY2>YUY2).

i5 is nice and all, but per-core speeds matter more than overall CPU. Capture only uses 1 core. CPU really isn't as much of a bottleneck as I/O and fragmentation, which is how most Huffyuv lossless frames are dropped. CPU enters the fray with newer codecs, but core is still the bottleneck there.

I've only skimmed this thread, not read, so I'll reply again later. But wanted to quickly leave those comments.

I'm not sure that's what Gspot reports. The start point is the start of the Movie, and the end point is the exact end of it. So I have nothing to Trim, because I want everything the AVI has in it. So I assume Gspot complains about something else?


Ermm. So the initial capture from VHS is Huffyuv, and because there's no frame drops - the re-compress to Lagarith will be the exact same quality but with smaller file stamp? Can I do that with VirtualDub? Would that be the "Fast Recompress" Video settup?

Ofcourse. I will appreciate it. Thanks you lordsmurf.

In a quick no-filter post-capture for myself, I clip off unneeded junk, split to multi files as needed, recompress to Lagarith for later. But more often, I fully run captures through quick overscan masking, etc, and then full process to Lagarith.

Initial captures is Huffyuv, yes.

When you say "clip off unneeded Junk" you basically just remove section of the media your don't care about? Or there's more then that? I wonder why Gspot reports 8GB of unused information.

Oh right. Overscan masking. I completely forgot about it. Is it something I need to do during Capture? or can do it later on the Lagarith file? I probably need to check here about a guide on how to properly Mask it (I know Crop is bad). Maybe it will even save a bit of space, because some of the information is gone.

What this etc includes? I might as well do it now it while preparing the videos from Archiving.

I will check around to how to convert from one format (huffyuv) to Lagarith. I'm assuming you can do it with VirtualDub playback section.

Thank you!

Junk = snow/etc
Masking = post-capture, not during.
etc = CCD for chroma NR, masking, whatever is needed. No 1-size-fits-all.

Those are still things I'm learning. So I guess I will start with Capture to YufHuff, convert to Lagarith (with MODE set to Yuv2 as suggested on Sanlyn's VirtualDub guide) - and later when I'll start working on specific Videos I would like to improve, do masking and other things I don't familiar with yet like CCD or filters.

Thanks again!

I know LS says otherwise, but I have just captured 20 VHS tapes, each one in one go (90gb) with V Dub (1.10.4) and Lagarith and have not had one dropped frame. I have an ancient Intel i5.

I'd try it. If it works ie no dropped frames, you won't have to convert from Huff to Lags.

Regarding your Gspot issue, GSpot has got it wrong. There will be no "no data" at the end of the file. It says the same for a 80gb video file of mine. You can easily check it: just open the file in VDub and drag the cursor along the timeline; you will see that the whole file has video.

To be honest, I didn't check capturing directly to Lagarith - but I don't mind that extra conversion. I'm assuming that if done correctly, it does not hurt quality.

Oh, so yea. Perhaps Gspot just got it wrong.
Thanks!

Best Buy right now has a sale on their 8TB EasyStore drives for $120. So a 65GB file costs about $1 in storage, plus another $1 for the backup copy. I wouldn't bother with converting from HuffYUV to Lagarith or ffv1 422p, you'll be saving only a few pennies per file.

Thank you. I didn't actually tried to convert HuffYUV file into Lagarith, so I'm not sure how much space it saves. I assume that if the delta is indeed not too big, I won't bother.

-- merged --

I had some time today to do some testings. I used two files for that:

- One was 35 Minutes Video: 15.4GB Huffyuv file turn out to be 14.95GB in Lagarith.
- The other was a 2:35 Hours video file: 62.5GB HuffYuv turned to be 63.5GB in Lagarith.

So as traal mentioned. I'm not sure it's worth the extra conversion. Converting the Larger file took 1:15 hours. The interesting part is that it seems like HuffYuv is actually handling Bigger files better than Lagarith. The 2:35 Hours HuffYuv file is 1GB smaller compared to Lagarith.

I didn't tested FFV1, but as for now, it seems like I'll just store the files as HuffYuv.

Lagarith has better/smaller file size, but it has CPU overhead to do it. There's no magic here. The tradeoff to more efficient compression has always been processing to do/undo it. That means both encode and decode requires more CPU. This is why you can't go nuts with x264, long GOPs, etc, because none of the standalone players can handle it.

When it comes to lossless, I keep both types. If my archive is post-capture encodes of any kind (example: unfinished restoration project), then I'd keep Lagarith intermediaries. That's what I was using. If it's just straight capture/archive, I generally store the Huffyuv.

I used to get more bent out of shape about file size, in years past, back when 1 encode would fill a drive. But I got a 16tb Seagate for $385 this summer. A few extra GBs won't even phase me now. It's be easier to buy another 16tb drive, instead of spend time recompressing to save 10-20% file size.

That's why I was surprised when the Hufyuv was actually smaller on the longer tape.

Yep, agree. Space is no longer an issue those day :-)