Almost every tape i'll be working with is a 'clip show'. And just as you described lordsmurf, these videos almost always use mutliple cameras. I see the same thing with recent dvds as well. Also, since this is a small niche that bloomed out of hte 90's, most of these tapes were recorded and edited by hobbyist and not hollywood professionals.

You each mentioned some errors, but my question is how do i fix them.. and fix them properly for online/progressive viewing? There have been many comments about the problems with the videos, especially with the latest samples i posted. But not so much on how to actually fix or clean up these errors. badly deinterlaced, artifacts, jitter i'm not sure what to do with these terms except add them to my list of google searches.


I read premium captures response last night about the LTBC, which i believe to be TBC or similar? I wasn't paying attention to the VCR for the recordings in the beginning as i thought i had TBC on, but i did some test captures in VirtualDub with the TBC on and off and noticed a difference with the 'interlacing?' lines. As I am a noob, I was playing with any filter that sounded familiar. I used a deinterlacing filter with 'blend fields' selected and it noticeably improved the blur-lines of the TBC capture. Now i have no idea if what is what i should have done, but these are just my experiments. I'll be posting some better samples later today.

I really appreciate all the quick feedback i've received so far. This website is truly a great, and i wish all my hobbies had an online presence as organized as you guys. So I just wanted to say Thanks offering a great place to learn and receive actual helpful feedback...and for dealing with noobs who are in way over their head :)

I think the macroblocks are just due to the fact that it came from DV, they are too big for youtube.

I think everyone's standards are different, but there are some people out there that either don't know how to do a lot of this stuff or don't want to spend more time on something than they will be watching it.

I have never associated the words YouTube and quality together, watchable is a better word.

I agree with all of that, yes. Messed up video just irks me, that's all.

Depends on the moon. :rolleyes:

@pinheadlarry, I can look at those caps today and try to come up with some suggestions later.

Like MPEG, it's BOTH a format and a container. Quicktime is an uncompressed format that fits into a Quicktime MOV container. But unlike MPEG, the Quicktime container can also house other formats like H.264, ProRes422, etc.

I never thought of that -- it may have been deinterlaced before it was even edited back to tape, if pinheadlarry didn't do it by accident in his own workflow. In that case, all you can do is try to undo the jaggy damage, using either (or both) Avisynth anti-alias filters.For example:
Nah, this is not hopeless. Go eat breakfast. :P

It's still overhaul pretty good. Ideal with worse videos all the time. I wish the source was this good! You should see the two sources I'm working on right now. (And you will, when I have time to finish writing that guides AND finish the project.)

You'll never get it perfect, because applying a filter to one clip messes up another, So some level of imperfection will always remain. What you can do is
(1) basic cleanup that applies to the video -- undoing the last generation worth VHS damage, no camera shooting damage
(2) do a good deinterlace with QTGMC in Avisynth

Either
or
I want to see what sanlyn is thinking, in terms of basic cleanup. I'm thinking just a basic 2-3 radius Temporal Cleaner in VirtualDub. Maybe a Chroma Shift in Avisynth, but what I find is the shift changes per clip, meaning it was edited together from non-masters, but copies of copies from home VCRs.

This will help: http://www.digitalfaq.com/forum/news...-avisynth.html

The full guides are currently being worked on, so you're too early for it. You'll just need to ask for help here, and we can try to give it, in terms of software usage.

I knew that would happen. Shame, shame premium! :P

He means "line timebase corrector", "line TBC" -- you almost never see it shorthanded as LTBC. A line TBC is usually part of a S-VHS VCR. If you're using the JVC SR-V10U (right?), then you have on. Just be sure it's turned on in the menu.

Thanks much for the awesome comment! :D

(Added it here: http://www.digitalfaq.com/forum/news...q-members.html too!)

The DV blocks are probably from the fast scene switching. Content like this really needs to be lossless to be perfect to work with (edit, restore, deinterlace/encode for streaming).

The entire lossless > streaming workflow is what I use professionally -- or rather, did use, as it's been 6+ months now since I did any streaming work for studios. With professional software, of course. But I can insert freeware into the same workflow.

Indeed! :mad4:

captured in virtualdub. screenshots taken of raw avi through vlc

No tbc
Attachment 3973

tbc
Attachment 3976

tbc-picturecontrol set to edit
Attachment 3975


ps. just saw lordsmurfs last past. time to start researching :hmm:

In the still, I don't see much difference with/without TBC. But it needs "ChromaShift" in Avisynth, follwoed by "Camcorder Color Denoise" in VirtualDub. It has really ugly color issues. But here's the thing -- is the whole tape like this, or just that one clip?

CCD is Vdub is good tape-wide.
But ChromaShift is not. You want this most, however, if possible.

NOTE: Video gets "sharper" simply by realigning chroma!

some clips look better than others, but for the most part it's similar

here are two small clips

tbc
https://www.dropbox.com/s/kb00yms1jf7u2xn/tbc.avi

no tbc
https://www.dropbox.com/s/7r6xelbybqtaig7/no-tbc.avi

I haven't had time to look over the .avi samples, but the images look cleaner or, at least, color looks better. As for the tbc effects or lack of them: looks like the patched-in shots from the "masters" left ragged edges and fuzzy borders that would have looked better if more care had been taken. Unfortunately, once those problems get recorded into the source, adding a tbc later won't make them go away. But at least it keeps things from getting worse. The "tbc on" shot does have slightly smoother edges in some areas, especially bright-edged objects. The "edit" shot is digitally enhanced but looks OK. Whether or not to use JVC's Edit setting depends on the source tape: it works fine with some tapes, not so great with others.

The chroma bleed in these tapes is really awful.

Well i'm still trying to figure out AviSynth to follow lordsmurfs recommendations, but i have some questions about the capture itself.

Is there a specific resolution i should be capturing in for online viewing? Should i adjust this before capturing or just let the canopus run it's resolution. which i think is 720x480.

I can't get VirtualDub to recognize Huffyuv, so i'll be using Lagarith. No question, just want to get feedback if this isn't the best option behind Huffyuv. (I think someone mentioned huffy x mac problems earlier in this thread.)

Also, i'm going to be capturing in FCP, then i'm opening the raw dv file (mov container) in AviSynth via QTsource. Is this an ok workflow?

I tried to capture using VirtualDub but i had no audio, then WinDV won't recognize my canopus, so i think my only option is FCP.

For the AviSynth headache, I've installed the plugins into the avisynth folder in Programfilesx86, but when i try to type the relevant code into avsPmod i keep getting a "no function named x". I'm sure i'm overlooking something simple, but i'll be spending the day trying to figure this out.

i realize now that i wasn't to specific with the avisynth headache. I think the problem is the .mov doesn't have YV12 colorspace?

ps. i have no idea what that means

I don't use a Mac, so someone familair with running VDub in other workspaces will have to help there.

I don't see that capturing to DV with FCP is going to help. We already know what those captures look like. They won't be lossless in DV, aliasing and other artifacts will be more difficult to clean up, you'll lose some color resolution, and you'll have to go to another lossy re-encode for online or other final output anyway. DV is PC playback only.

Your VirtualDub capture had muted audio, bit it did have an audio channel. Even if it wasn't mute, the sampling rate was too low (it was 96khz). For online viewing -- or for anything else, really -- capture at 640x480. For final output as DVD/AVCHD, etc., you can resize in Avisynth. I don't know the reason for the mute audio, but it could be incorrect settings. Audio capture isn't usually enabled by default. Did you capture with composite cable or s-video?

Your tbc.avi and no-tbc.avi were captured at 20fps, which I don't think is correct. The Colorspace was YUY2, which is preferred. DV capture can be only YV12, and DV uses a somewhat different color matrix than VHS YCbCr.

Use 32-bit Avisynth. If you're using 64-bit win7 it won't work if you install 32-bit Avisynth in 64bit folders or vice-versa. Install only 32-bit filters with 32-bit apps. There are hardly any 64-bit filters. I don't know which Avisynth plugins you refer to.

In VirtualDub, 'audio > enable audio capture' has been selected for each capture. i haven't been able to find a solution yet.

audio input - the only option is 'master volume'

audio source - no options available.

also,i installed a new VD with windows 7 and now Huffyuv is working.

Ouch! Something wrong there. Of course, you have left/right audio cables hooked up somewhere (well, thought I'd ask). I don't know what capture device you're using. If your audio card is a Mac piece, well....that might be it.

That's good news, anyway. Hopefully someone with knowledge of how you're running win7 on a Mac can chime min. Wish I could give more details.

Ye olde vgsamp2.mov (ah, remember that one?) gave me fits but I couldn't just let it sit there. My Achilles heel. Seriously, it's such a wreck. Not your fault, of course. I did what I could with it. Can give some details later if anyone wants, but 90% of the work was in Avisynth, the rest in VirtualDub. Some aliasing still there, lots of smeared chroma remains. I guess you could say the mp4 looks, uh, "better"....for want of a more accurate and less charitable term. What that last shot of the leap over the wall is supposed to look like is anyone's guess. I'm afraid some sort of autowhite or autogain helped do it in. I hate "auto" stuff. Has a mind of its own. Basically I just denoised it. I really wish those guys knew how to deinterlace. Of the few brights that the movie crew didn't totally obliterate, a few stretch toward RGB 255 a bit, but this is PC display anyway. Suggestions welcome. Just don't kick me.

They say anything is possible. Not really.

Windows 7 x64? If so, you were installing the 64-bit DLL, not the 32-bit one. :wink2:

In Windows, in VirtualDub, convert the MOV to AVI.
Then process the AVI in Avisynth, then feed the AVS script to VirtualDub. Save as the processed video.
This is easiest, even if extra steps.

The source capture should be YUY2, not YV12. Where are you seeing YV12?

is this good or bad lol? (i was previously running windows 8)

That may be the easiest way to get to my goal, but what is the best way to ensure quality? Does VirtualDub really produce a better capture compared to FCPx?

For the YV12, i opened my .mov file in virtualdub with QTsource, then when i tried to apply a few of the filters you mentioned i got an error messaging saying it only supported that colorspace. sorry for the unspecific description:smack:

It may be worth a momentary time out to figure out what you expect to get out of the tapes. You mentioned you have a bunch of these and plan to do more. They seem to come from inconsistent sources, which means on the long end you can segment and edit each, which will take a long time, or broadly filter, which will produce results not as good but save you a ton of time.

Since these are commercial videos and you are going to YouTube for final export and capturing DV, I would shoot for better quality than available videos which is very easy to do but thats just me.

At a minimum, I would suggest Camcorder Color Denoise, QTGMC for deinterlacing/noise reduction, and bringing the file back into FCP X for color grading and final export. If you give the same file to ten people, you will most likely get ten different videos back, albeit the best should be fairly similar.

I've been pretty clear with what i want to get out of these tapes. My set up isn't ideal, but i'd like to get the most of my setup. i have a quality over quantity mentality with everything i do..

I assume those are avi filters you mentioned?

I'm still going to work on this audio problem. I'd hate to have to capture in worse quality.

They aren't strictly "AVI" filters. There are Avisynth filters that work with decoded AVI in YUV and RGB, while most NLE's will use RGB filters. I suggest you use Avisynth to convert to RGB if it's necessary. Most NLE's do a sloppy job of it.

I think over three days you have learned a lot from where you started, but I don't think you are accounting for the time factor here. Words like 'best' are often misunderstood. No two VCRs, even from the same make and model, play a tape the same way. Because of this, there is never any absolutely clear reference for how a tape should look, it all comes down to personal preference, something that you can only truly develop once you know enough about video errors, how to correct them, and what effect corrections will have on your original capture.

That's the reason why there are so many suggestions here to drop the Canopus. It is essentially the equivalent of putting a filter on to degrade your video before you start. If you are looking to get them as good as they can look, the Canopus by my estimate cuts the original quality by 25%. Considering how bad VHS is to begin with, you end up turning a bad source into a worse source. And that isn't even considering the time factor.

If you are looking at 30 tapes, you could be looking at roughly 20-45 hours of capture time before even beginning to edit. Since you are still learning about filter options, depending on what you are looking for, continuous time on a single tape could exceed 12 hours. If you work on these two hours a day, you are still looking at three months.

You made some good points premium.

It's not that i'm totally stuck on the canopus, but i don't see many other options...unless i buy a new computer. I'm not aware of any other replacements for my canopus that can provide better results. If you can, please recommend me a device as i'm still within the 30 day return policy at amazon lol :rolleyes:

Time is irrelevant for this. I'm just doing this as a hobby and if i ever need content DVDs aren't nearly as complicated ;)

Also, i think i found a remedy to the no-audio in VirtualDub. I'm going to record video in VirtualDub and record audio with quicktime on my mac.

http://www.digitalfaq.com/forum/myth...html#post32117

This is something you should definitely read. I bring up DV myself a lot because as a Mac user seeing a lot of Mac users praise the Canopus, I came close to buying one myself.

In the end, for stuff I want to put directly to YouTube or commercial tapes, I use the Elgato Game Capture HD. I don't claim its better than uncompressed options, but at high bitrates and 4:2:0 sampling, it retains a high amount of detail with enough bitrate to not choke on noise that editing and quality are fine and look much better than other VHS captures on YouTube I have seen.

Lets say the video quality on YouTube of VHS is a 2. The captures I get with H.264 and some brief editing give me a 7-8. If you want to take commercial tapes to a 10, this is what putting in work for a 10 looks like:

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

FYI: I'm working on a project for Original Trilogy myself, through a long-time member there, but I'm not saying anything else. It's taken years, and I hope to have it done by the end of this one finally.

(If I mentioned what it was, nobody would ever leave me alone! :mad4:)

Are there any comparisons or can lordsmurf chime in about the Elgato Game Capture? the eyeTV? Hauppauge 1212 HD-PVR? anything really..

the company seems to market themselves as providing easy to use hardware, not necessarily quality.

markatisu used the Elgato products quite a bit, back in the day. He was a member here years ago. I saw his work, and the capture looked good. That was an older model, so I can't/won't say much about the current ones. It's worth looking into, at least. If I needed to capture on a Mac, I would start there. I prefer 4:2:0 MPEG captures, at minimum.

EyeTv and 1212 use the same terrible converter. They work OK but by many accounts have a one year lifespan.

https://drive.google.com/file/d/0B8e...it?usp=sharing

I uploaded a short clip. Download the file to watch the original quality, the youtube player has processed the viewable version. It hasn't been deinterlaced or converted from capture, just split via AVIdemux. Looks a little bright but the settings are adjustable.

Again, I am not saying that this is better than an uncompressed capture, but I believe it to be a big step up from DV.

Thanks premiumcapture for your input. I went ahead and did more research last night and it seems like the game capture is favored over the canopus (atleast @ digitalfaq). I went ahead and ordered the Elgato Game Capture.

Before i send back the canopus, i'll capture some clips to reference the quality of the canopus vs the game capture.

Once you get it, test it against the canopus before you decide to return it just to make sure, but I am pretty sure you will be happier with it. If you keep it, get the retro adapter for S-Video ($10 on the Elgato site).

If you end up going the uncompressed route later on, you can always get a $40 capture device, but the game capture will let you do everything in Mac if you want it to, plus you will be able to capture anything over HDMI/Component/etc...

It has its own software, no VDub support for capture, but you can still bring the files in to VirtualDub after capturing. Files run something like 6GB/hour at most at max bitrate.

What capture device is $40 and produces uncompressed?

Perhaps i'm misunderstanding terminology, but isn't uncompressed better?

Uncompressed is better but more drive space. The ATI 600 USB that LS uses or the VC500 that I use for important stuff are roughly $30-$50 new or used. Not sure if you're using an iMac or MBP, but my MBP with parallels can handle the VC500 for uncompressed:

http://www.amazon.com/Diamond-VC500-...keywords=vc500

Note that there is a Mac version but it will not work with VirtualDub; the Mac software it comes with uses terrible compression. I own this one and it works.

The difference between the canopus, game capture hd, and the vc500/ati 600 is chroma subsampling and intraframe compression. DV is the worst at 4:1:1 and 20 year old compression techniques, the Game Capture HD works at 4:2:0 (like a dvd or iTunes movie) and applies compression through the newer H.264, which is a significant upgrade from DV, and finally a compatible USB capture device for Windows that captures 4:2:2 with no intraframe compression.

The VC500 at least I know will work in Parallels on a MBP with Mavericks. I suggested the Game Capture HD since you started with an all Mac workflow, which is definitely not a bad choice, but for getting the absolute best out of every tape you need one of those instead.

wow I always assumed those little USB adapters were shit.

do you have any comparisons between the vc500 and the game capture?

also I was recommended something that is $100 more and produces a worse capture? what's up with that..

not a bad assumption, most of them are. I dont have any raw vc500 caps to share at the moment, but the decent usb sticks with virtualdub will basically copy everything it sees output from the VCR. Different USBs use different comb filters and have other subtleties, but the VC500 is the last acceptable usb capture device IMO that is still produced new.

the game capture HD is $100 more, but these two devices dont work the same way. the usb stick offloads all the input for the PC to process, which means poor computers/drives will not cap right, while the elgato has a built in encoder that does all the capturing outside of the computer and simply copies the postprocessed data. the device was made to capture HDMI gaming, but is compatible with older game systems for youtube uploads, and because of its HDMI build, it can capture at bitrates that will make the actual capture very closely resemble the uncompressed copy at much better data rates. it beats out all the vbr MPEG-2 captures I have seen. If you had no PC access, I wouldnt hesitate to recommend it, I have used it for a while for various tapes and have been happy with the results.

You will always get more out of an uncompressed capture, but capturing with the Elgato is a prick in quality compared to DV.

Whatever you decide to do, a timebase corrector is recommended. I use the AVT-8710 most of the time but I do have a TBC-1000 that needs refreshing.

Neither the AVT-8710 nor the TBC-1000 will do very much to correct analog line timing errors. Each of those tbc's adds undesirable effects that are unnecessary unless some full-frame related problems exist with the source. I believe the O.P. is using a JVC with built-in line tbc. What would the full-frame tbc's be for?

This appears to be a pretty active thread right now.
So please read this: http://www.digitalfaq.com/forum/news...ead-reply.html -- we need feedback on it ASAP.
Thanks.

FYI: Just read that last reply -- and it's correct. External TBCs do little for "line timing" (horizontal jitter). You need internal TBCs in a VCR, or a passthrough deck like the Panasonic DMR-ES10/15/20/25.

I have a JVC SR-V10u, which has TBC.

i'm more interested in the usb device now. I have a decently upgraded iMac, so it should be able to run fine.

What did you mean by this?


I believe we have similar workstations so what is your workflow with these devices? Can they capture in VirtualDub via Parallels?

I realize this. My only thoughts are that since these are commercial tapes, Macrovision might come into play and cause capture problems.

With the Elgato you could in practice stay in Mac 100%. You can edit the MP4s in VirtualDub if you want but it would let you edit them if you wanted.

The difference between high bitrate mp4 and uncompressed is much smaller than uncompressed 4:2:2 vs. DV. Since your ultimate goal is to go to YouTube anyways, much of what you lose in the initial capture you will end up losing in the end for upload anyways. It is always better to start with a higher quality source, but I would argue for your purposes the difference is negligible. If you wanted fully uncompressed tape backups, the VC500 would let you do that but the Elgato would not.

For this, I would suggest capturing first, filtering, final touches/color correction in FCP X and export from there. I would personally filter in Windows but everything else in Mac will offer good results.

One little nitpick with the Elgato - it will try to automatically deinterlace after every capture. The workaround is to go to the folder and delete all the other files other than the actual .mp4. It does this because it assumes the video is for youtube. A small annoyance, but one to be aware of.

Most of my audience will come from youtube, yes. but i still want a file to offer for download. I assumed before this could be the same file. But just want to say that this won't just be for youtube. No one has spent the time to archive these tapes before, and since i'm building an archive for this niche i might as well try and get the best copies.. i can achieve.


For my workflow. I'd like to capture in VirtualDub as it gave me better results than FCPx. From there i'd like to filter and correct the capture with AviSynth / VirtualDub then export the final copy from there.

Is this possible with the Vc500? the Elgato Gamecapture HD?

it's a bit ridiculous that i was recommended to buy a whole new set up but not give a $30 device a try...

here's a youtube comparing the VC 500 vs Elgato Game Capture HD. Yes i understand it's youtube and all the difference factors involved but it looks to be in favor of the Game Capture.

http://www.youtube.com/watch?v=QAYOwZ1H-q8


PS here is a comparison for the lols (not real)
http://www.youtube.com/watch?v=HLzuU60Tyik

Most people here don't use Mac, and if you run a setup only in OS X, the Elgato is hands down the best option. Even with Parallels, much of the typical capture hardware used isn't compatible with either its drivers or the Mac ports. The VC500 is the only one I have been happy with, but I honestly think no one knew/tried it.

Also, I did mention it on page 1 :)

Very true premium. Just since no one else chimed in i didn't know if they were ok devices. But as i started researching these different devices and working with a Mac, i keep finding you in random digitalfaq threads ;)

here is the canopus vs the game capture as promised. I'm not sure if the game capture's quality is 'better' but it's definitely different. Looking forward to hearing your thoughts on this.

Here are some screenshots. One source is from the elgato game capture using the 'best' settings to capture. The canopus was captured in virtualdub with uncompressed avi. Screenshots taken in VLC

Canopus
Attachment 3997

Elgato
Attachment 3999

Canopus-2
Attachment 3998

Elgato-2
Attachment 4000


Here are two small clips via dropbox.

https://www.dropbox.com/s/rfhof94rxs9co3j/caponus.avi
https://www.dropbox.com/s/kkgs8s4829ce9zb/elgato.mp4

I touched on it briefly before, but in order to avoid the Elgato from deinterlacing, you need to close the program after you hit stop recording, go into the file folder, and delete everything except for the largest file (MP4). Did you do that for this recording?