Archiving hundreds of home videos: advice?
 

Hi all.

I've read over several threads here and there looks to be a plethora of great info. I just wanted to get some specific input about my needs before I pull the trigger on spending more money. My goal is to archive a large collection of home movies, preserving or improving the quality of them if possible. My budget is limited, but if there is a drastic improvement in quality by spending a little more I would like to know that.

First, here is a description of the media I wish to archive:

~200 8mm tapes spanning the years 1987-2003. It is possible some of these are Hi8/Digital8, but not positive.

- An undetermined number of VHS tapes spanning the years 1985-1987

I have a Sony Digital8 Camcorder, model DCR-TRV740. I had been using the Firewire port on the camcorder to capture to uncompressed AVI files on the computer. This was working great (~20 tapes or so captured) until the firewire port on my camcorder stopped working. I then puchased the Elgato capture card on amazon: http://amzn.com/B0029U2YSA.

So far I have captured 2 tapes with the Elgato. In both cases, huge chunks of the video were missing from the resulting captured video. A second attempt to capture the tape resulted in the full video being captured. I do not have an external TBC (from reading these threads, perhaps this was the only issue).

I just built a new computer for gaming and video editing (part of why my additional budget is limited). Here it is: http://pcpartpicker.com/p/WmrfMp (I also added a Firewire PCI-E card before the firewire on my camcorder stopped working).

My question for all of you is simply: what capturing setup should I use to archive all these home videos?

From my browsing of the other threads on here it sounds like I will definitely need an external TBC. I am looking at the AVT-8710 TBC: http://www.bhphotovideo.com/c/produc...Time_Base.html

Beyond that, I'm thinking I should return the Elgato capture device and get something that will allow me to capture uncompressed files again, probably using VirtualDub. Is there a capture card that I can get for my new PC setup that would give me a good capture workflow? Once I finish capturing the 8mm tapes, I'll probably have to check out the VCR buyer's guide on here and grab one of those as well.

I've seen a lot of comments that the best capture setup is actually an old Windows XP rig with an ATI AIW card. I would basically have to hunt down old parts and build that computer from scratch, which would probably get pretty pricey. Is there a capture card I can get for my new PC that would yield equal/similar results? If I were to bite the bullet and build a WinXP capture machine (assuming I can find the parts), what benefits does that give me?

Thanks for reading this, and I appreciate any advice at all that anyone is willing to provide. I've been hosting the videos I've captured so far on a Plex media server, and my family is already cherishing the few videos I've gotten so far.

The captures you did over Firewire were not uncompressed; Firewire from your camcorder is DV-compressed.

The Elgato can do lossless capture in VirtualDub.

I tried using Elgato with VirtualDub. VD recognizes the elgato device, but I couldn't successfully capture anything.

Assuming I successfully troubleshoot that issue and actually get the Elgato to capture w/VirtualDub, do I just need to buy TBC and then I'm good to go? Or should I still replace Elgato with something else?

Thanks

UPDATE: Ok I figured out that it was selecting Elgato as the audio device that was keeping VirtualDub from capturing. So now I can capture using the Elgato with VirtualDub, and I was able to get the Huffy codec installed on my Win10 x64 machine.

So, should I stick with this setup and add a TBC and I'm good to go? What about a sound card? the onboard audio on my mobo is supposedly pretty good, but idk any better. If I will get better results using some other device/card to capture than the Elgato, I can return it.

Also, I've noticed that the lighting on a lot of these videos is pretty dark. Is this something that can be fixed in a way that still looks good? If so, would that need to be a capture setting, or could it be done in post-processing after capturing a lossless file?

Your camcorder is built in line TBC and DNR noise reduction system and that's all what you need, External TBC's are expensive and most likely it will add nothing to your video, I have the exact same model that you are looking for and didn't fix a tape that has frame jump.
You don't need camcorder firewire port just use a USB capture device like the one you already have.

For DV or D8, correct. :congrats:

The digital origin signal (be it native DV/D8, or converted analog) can be modulated out over analog, but that analog signal is usually clean. Sometimes digital vs. analog out isn't as big a difference as some would think. I often transfer DV over s-video. The original content isn't really 720x480, as consumer cameras couldn't resolve that well. And they were noisy, so any added analog noise won''t be noticed.

(For analog, not correct. Dropped frames are still an issue. That's what the external TBC is for. Remember: Internal to clean the picture, external to clean the signal.)

The bigger concern is converting analog to DV with a camera. That's rarely any good.

And TBCs in cameras often are not true TBCs. So you actually drop frames in the camera, before the modulated analog is output again. Or the digital DV file is transferred over Firewire. Remember that playback and analog capture is done as one step with DV/D8 cameras. That's rarely a good thing.

Ok so just so I'm understanding you correctly, are you saying that my camera as the video source is not great? Or was that simply a confirmation that I do want to get the external TBC? I am definitely getting loads of dropped frames in VirtualDub during my captures right now. Thanks for the replies.

I'm saying the cameras is the weak link right now. I'd hate to use one to capture 10 tapes, much less 100 or more.

I'm also saying that TBC is usually not needed between a DV camera and computer, be it via analog capture or via firewire transfer. The digital or analog output signal is DV, not true analog. It was not analog the moment that the camera touched it.

Then again, all models differ some. There is a potential that this camera is playing Video8 and Hi8 without the DV conversion step. And that would be preferable. And in that case, a TBC would solve your problem. Dropped frames are most often caused by lack of TBC on a bad video signal.

However, bad VirtualDub settings, or capturing with sound preview, is another. So let's confirm your VirtualDub settings.

Hm, interesting. So when you say the camera is the weak link, is it weak enough to make a noticeable difference in quality? If so, what realistic alternative is there for capturing these 8mm tapes? I've seen 8mm to VHS adapters that would let me play them via VCR. Those seem terrible to me (just a gut feeling), but would that be preferable? I don't have a VCR yet but I'd get a good quality one.

Regarding VirtualDub settings, I did have audio on during capture so maybe that's my problem. I know there are settings guides on here so I will double check the recommendations vs what I have (thought I'd done that) and report back here with specifics if it's still dropping frames. Thanks!

I think Lordsmurf means if your camera dropping frames then you either get another one, get a fancy deck or buy an external TBC, assuming you are playing analog tapes with your camcorder, for digital tapes you don't need any of that just transfer native DV.
And by the way there is no such 8mm to VHS adapter.

I have an external TBC the AVT-8710, It just introduces artifacts to my video I don't trust it anymore so I took it off my workflow, Maybe because I don't have dropped frames to begin with, I got it because I have few VHS tapes that have wavy magnetic tape edge was hoping to compensate for the damaged synchro track on the tape but it did nothing to fix the problem.

To just transfer DV for digital tapes, wouldn't my cam'a firewire have to work? Can you give an example of a fancy deck? Thanks

Yes for Digital8 tapes you have to have a working iLink port on your camcorder. Before you buy another camcorder or deck see if dropped frames are Vdub issue post your setting and follow up with Lordsmurf or anyone who is well familiar with Vdub. D8/Hi8/8mm decks range from $500 to $1500 not an option if you are on a limited budget.

Troubleshooting VirtualDub dropped frames: first confirm that you can capture with 0 drops if you disable audio capture entirely and capture from a stable source (DVD player, if possible).

Back with an update. So I believe I'm following the settings suggestions as best as I can from here: http://www.digitalfaq.com/guides/vid...virtualdub.htm

Some details that might be relevant:

- Under Device, I have "Elgato Video Capture (DirectShow)" selected.

- I have Video set to Overlay, and the Format option is greyed out.

- Video source is Video SVideo (I have S-video running from camcorder to elgato)

- For video compression, I have Huffyuv v2.1.1 set to predict median (best) and predict gradient (best)

- For audio source, I have "Capture Device" selected. There is an option for "Analog Audio In (Elgato Video Capture)", but strangely if I select this the capture immediately fails. I do get audio on the captured videos using "Capture Device" though.

- For audio compression, I have <No compression (PCM)>

- In the bottom right, I set audio to 48k/16/s and video to 29.97fps.


I confirmed that I had previously had "Enable audio playback" turned on, so I turned that off. It seemed like this had done the trick and it was working great, I captured the first 2.5 minutes or so with 0 dropped frames (it had previously been a pretty constant stream of several dropped frames per second). Then, the dropped frames count jumped to 3332 and stayed there. I let it capture another 15-20 seconds and then stopped it to check on the captured file. When it got to the part where the frames dropped, the audio in the video sounded like the tape was being eaten. I ran back the tape on the camcorder and played it there again and it still plays perfectly.

So... what gives?

And I saw the suggestion to capture a known stable source with the Elgato. I'm worknig on that. Working on finding a device that has the right outputs to test (S-video, RCA...)

Thanks for the help!

EDIT: Ok I just found a DVD player that had one of my wife's workout dvds in it. That should do the trick as a capture test. So far it's 2 minutes in and no dropped frames.

I'll go update that guide this week. :smack:

15 minutes in and 0 dropped frames from the DVD player.

So what does that mean? Bad camcorder? Camcorder okay and I need an external TBC?

I'm even browsing the web and stuff and the DVD capture is going like a champ. FWIW I was not doing anything whatsoever in the background while capturing from the camcorder. But my system should handle it fine even if I did.

OK I did a brief audit of the 8mm tapes I have and have some additional info. The tapes actually span 1987-2007. It appears to me that most/all of the tapes from 1987-2000 are simply labeled '8' or 'Video8'. In the early 2000's, there starts to be a smattering of tapes labeled "Hi8". Beginning in 2005, there are some tapes labeled Digital8, but not all of them.

Furthermore, there are settings on the camcorder itself that can be tweaked if that would help. Here is a link to my camera's user manual for reference: DCR-TRV740 Manual

Here are the VCR settings on my camcorder that I can tweak:

The format for this will be [Feature]: [options] (currently selected option)

HiFi Sound: Stereo, 1, 2 (Stereo)

TBC: On, Off (On)

DNR: On, Off (On)

Audio Mix: Slider from ST1 to ST2 (Centered between them)

A/V->DV Out: On, Off (On)

PB Mode: Auto, Hi8/8 (Auto)

ETC options

Display: LCD, V-Out/LCD (LCD)



So, it looks like maybe I could set A/V->DV Out to off and have it play back completely analog with no DV transcoding going on within the camcorder? But then if I did that I would need an external TBC (or maybe I do regardless?) What would that setting do when I get to the Hi8/Digital8 tapes if I have it set to off?

Thanks again for taking the time to read this stuff and help me. Getting these videos backed up and available digitally will really mean a lot to my family. We just lost my (maternal) grandmother, and my mother especially appreciates having all these memories on video.

You need DV to be ON for Digital8 tapes only, Hi8 and 8mm you can turn DV off, Try capturing with DV off and post back.
Also if you are not using S-Video cable try using one that may help.

Yeah I am already using S-Video. And it's a brand new monoprice s-video cable that I just ordered from amazon.

I will try capturing again with DV off. What about PB Mode? Should I take that off Auto and set it to Hi8/8?

All that A/V->DV Out does is change the signal direction of the analog connections. The analog ports on these units serve a dual purpose since there aren't separate inputs and outputs. When you turn the function on and use the standby or record mode, they become inputs instead of outputs. When playing back a tape, they will always be used as outputs.

The only downside of using PB Mode Auto that I have seen is that you can lose some frames or perhaps a second of video at the start of a new recording as the player attempts to decide whether it needs to interpret the signal as analog or digital.

For the capture troubleshooting, I was thinking more like a 2 hour capture to be sure. After that, do the same thing but enable audio capture from the DVD player. If that succeeds, try an 8mm capture with audio capture disabled.

So, interestingly enough, I just captured a full 8mm tape (same one as before) with 0 dropped frames. There were 10 inserted frames, but anecdotally it appears as though it just inserts a couple here or there between scenes on the tape. An 11th inserted frame occurred at the end of the tape.

I will post Vdub stats below. The only differences are that I turned A/V->DV to off and I changed PB Mode to Hi8/8. So far for all of my captures I have captured audio, including this one. The problem I was having earlier was that I was also enabling audio playback during the capture. So, what do we think about those findings?

Here are the VirtualDub stats for the capture:
http://www.digitalfaq.com/forum/imag.../DIXOQFV-1.png

The file came out to be about 60 GB. That's twice the size of the DV captures I'd been getting over firewire. And that's apparently with some amount of compression from Huffy. I wouldn't say the video quality is noticeably better than the DV captures I've done. In fact, with my limited sample size I'd be inclined to say it might be slightly worse, but obviously that's just one tape.

Is there anything I can do about low-light conditions? Here is a screenshot of what I am talking about:
http://www.digitalfaq.com/forum/imag.../7LS6jFu-1.png

I also don't know what the preferred method of taking a screencap of a video is. I just used VLC player to do it. Anyway, is there anything I can do at this point to improve on that? Would I need to change the playback brightness settings on the camcorder prior to capture, or is it something that could be done either within the capture software or after the fact with Adobe Premiere or DaVinci Resolve?

See, You just saved $200 for yourself for an external TBC, As for DV quality, it is impossible that DV lossy compression is any were near losseless compression. You can tweak the brightness a little bit while capturing with Vdub but you have to be careful, Once it is captured it cannot be reversed.

Ok so I would have to recapture the tape with different brightness settings in vdub? That's interesting. I would have guessed that anything that could be done within the capturing software could be done later in premiere/resolve.

Are there any standard recommended tweaks for the vdub brightness, or do you just leave it at the default unless there is a specific need for tweaking? And is it generally a custom tweak for every tape captured, or do you generally land on a good setting that works for most tapes recorded on that cam?

YSuper-black or super-white luma or color values that are out-of-range in YUV during capture or other processing are clipped at the dark and/or bright end. Clipped means "destroyed". You'll never see those darkest or brightest details again, just as there's no dark detail to retrieve in the image you posted earlier. During color correction later, it's impossible to attain snappy, full dynamic range with clipped visual data that's destroyed during capture. I'd also advise that you depend on Avisynth, not Adobe, for consistent and correct color conversions between YUV and RGB and post-processing repairs.

Every tape has its own requirements, although corrections for invalid video recorded and input levels tends to encompass a general range of variations. Some tapes are nightmares and require strong correction. All corrections require that you use a histogram to see what's happening and to see what you're doing. If you don't know the basics of using histograms and other similar 'scopes (Adobe Premiere has about 8 of them, if I recall, which you paid for), you won't be able to see what you're doing. That's especially true when using an uncalibrated PC monitor.

VirtualDub capture has a luminance histogam that can be used fort analyzing input before starting a capture. Adjustments are made with the "Levels..." filter group, which are usually a simple hook-in to your capture device's proc amp controls. Using VirtualDub's histograms has been mentioned in several tutorials and has been posted in many threads in this forum. The most recent summary and explanation was on 4/23/2016 in post #18 at http://www.digitalfaq.com/forum/vide...html#post43646. In case you missed that thread, part of it is quoted below:

Wow, awesome post! Thanks for that explanation and for quoting the other thread. You're right, I hadn't seen it.

And yeah, I am actually using my projector as my primary computer "monitor" right now, so using VirtualDub's histogram would definitely be in my best interest. I'll go play around with those settings and try re-capturing the same tape and see how it goes.

Thanks again.

So I recaptured the same tape again after tweaking the brightness/contrast to get the red edges out of the histogram in VirtualDub. It was really difficult because every new scene location was different, so I basically corrected the worst offender and then the rest of the scenes were pretty much clear of the red bars as well.

I probably didn't do a great job, but I figured I'd post some screencaps for feedback on my first attempt at this.

First, here are the levels I ended up with.
http://www.digitalfaq.com/forum/imag.../5DYMwYN-1.png

Here are the stats for the capture:
http://www.digitalfaq.com/forum/imag.../mMhvW3M-1.png


Here is a cap from the first capture (no color changes):
http://www.digitalfaq.com/forum/imag.../ObUJS0i-1.png

Second capture:
http://www.digitalfaq.com/forum/imag.../kCPbOaI-1.png

First capture:
http://www.digitalfaq.com/forum/imag.../fjtL3ou-1.png

Second capture:
http://www.digitalfaq.com/forum/imag.../vbSBm3P-1.png

Worst offending scene first capture:
http://www.digitalfaq.com/forum/imag.../XlWegdh-1.png

Second capture:
http://www.digitalfaq.com/forum/imag.../3KANkW2-1.png


So what do you think? Is it any better at all? Did I go too far, and does it seem like there's a third option somewhere in between that would be better than all the screencaps above? I spent probably an hour or more going through the various scenes on the tape and tweaking brightness/contrast to get the red edges out of the histogram. If I end up with noticeably better results, the time spent doing it is worth it to me. Plus I'll get a lot faster as I improve at it.

Also, I haven't paid for Adobe. I have used it in the past and am familiar with its workflow, so I've sort of been keeping it in my back pocket as an option if I can't find a free/cheaper alternative that is as good (or better). Otherwise, I plan on just subscribing to the premiere cloud a month here and there as needed.

I also just bought the entire AVS4YOU software suite. It's apparently normally priced at $199 and is marked down 70% off this month. I found a code that knocks another 20-30% off that price so I ended up getting the suite for $40. I figured it was worth it to jump on that and then decide later if any of the programs will be useful to me. I've used their video converter in the past and at the moment I'm not aware of a better alternative.

I've also installed DaVinci Resolve, but have not poked around much or really used it yet. Basically my point is, I can use whatever you recommend because I don't already have a sunk cost into much of anything when it comes to software.

Thanks again for the continued help everyone.

All images show seriously underexposed video, as you probably know. Making multiple captures of problematic originals is par for the course, so join the club (at least membership in this club is free!). As far as I can see, the better images are within a valid levels range, such as they are given the circumstances.

It's possible to get some detail from underexposure, but it has to be done with the original YUV colorspace in Avisynth before doing anything else. The caution here is that underexposure, whether video or pictures, involves noise you won't believe until you see it raised to visible levels. Denoising in that same colorspace is another part of the process. Don't expect perfect results. I've been this sort of thing with my sister's camera antics and other forum posts.

The images have been converted to RGB, so the only way to work is with an unaltered sample of the original YUV capture. You can make an unaltered short sample of several seconds in VirtualDub by cutting a portion of video and saving it as AVI in "direct stream copy" mode. Samples can be posted here.

The most direct way of making images of frame captures is with VirtualDub's top menu toolbar: "Video..." -> "copy source frame to clipboard". Open an image app and create a new image or frame, paste from the clipboard, and save the image as-is.

Ok so to attempt to improve the underexposure, I would take my second capture file (that I adjusted brightness/contrast for) and load that into Avisynth?

Ok so to get sample clips and comparison images, I need to queue the tape on my camera up to the specific scenes and then recreate the same brightness/contrast settings and capture clips of that and again with the default settings? Not complaining, just confirming that's the correct pipeline.

Do you have any guides you've done that would help me work on the underexposure with Avisynth?

Thanks!

To apply filters and fixes, you'll end up doing some in Avisynth, some in VirtualDub. That's for later. NLE's are OK for cuts and joins, audio, timeline, color, but restoration and repair is not what NLE's are desired for.

Not necessary. You already have a capture or two. Open a captured avi in VirtualDub and make your desired cut with VDub's navigation tools in the lower left of VDub's window. Put VDub into "direct stream copy" mode, then save the cut as a sample avi.

What is used depends on specific problems. The best anyone can do with what you've shown so far is to try a few things and see what's possible.