Recapturing analog source video files

[diGriz]

I've had a very interesting couple of weeks.

So, for all of those who have captured all those analog sources to some kind of video file (usually .AVI at a low resolution), I can verify you get what you have. I did an initial test run of a VHS converted video file into a low resolution .avi file and performed a 1080i/p real time capture.

The first problem you're going to run into is deciding whether you're going to use the YUV or RGB settings, both of which have drawbacks. Your YUV uncompressed 1920x1080 avi stream is going to eat at least 500GB for an hour, and none of the editing software packages out there (including FCP and AVID) can handle that file size. Reverting to the RGB (M-JPEG) capture settings produces a more compact file size (around 25GB), but the capture will expose any and all artifacts, so plan on doing some extensive video filtering editing work (whether you're using FCP, Adobe, or Avid).

End of the day, the better quality of your analog source makes the difference. I recommend getting your analog video off its existing tapes (I want to meet the idiot who decided VHS was a good video format... since it was designed originally for audio).

[DB83]

Can I ask a question.

What are you smoking ?

[diGriz]

I'll assume that should be translated as "What the hell are you talking about?"

Back in the 70s when VHS was developed for the market, broadcasters intended to use the medium for audio recording since it's a wider magnetic strip. Betacam / Betacam SP was already in place for video recording (and it's still one of the best tape-based sources for video recording- miniDV and DVCAM tapes developed out of its Betacam parent). What they found was VHS was capable of recording video and because of its low cost, it was a cheap method of introducing the consumer to the video market (although VHS suffers the same problems both 8 Track and Reel-to-Reel audio does, which is warping and degregation, after playing over and over again, as well as heat elements). VHS decks were easier to produce, at an affordable price (vs. a broadcast standard Betacam SP desk). Although to be fair, miniDV and DVCAM tapes also are prone to stress and warp with time.

Now let's jump ahead to the present.

The bulk of the work I do during the day is transferring tape source-based video (be it VHS, Hi8, Beta tapes, miniDVs, etc.) people have recorded on and digitizing it to be either burned to a DVD, or in whatever container video file format for hard drive (although largely x/h264 MPEG-4 10/18 based). To use an example, let's say several years back in the late '90s I converted most of what I shot into RGB-based uncompressed .avi files at lower resolution, 23.976 bitrates with .mp3 audio containers (we didn't have FLAC yet).

Older players can support playing various video files (including .avi, MPEG1). Although there are methods of patching media players, which is an entirely different topic. If you were to play one hour of footage you shot on a player into an HDMI capture card at 1080 50/60i 8-bit YUV, the result is an uncompressed .avi file at around 380-390GB.

Now re-capturing your old .avi file using AVI Motion JPEG capture (also known as M-JPEG) at 1080 50/60i 8-bit YUV will result in an uncompressed .avi file at about 24GB.

Software editing programs such as Premiere, FCP, Avid can handle HD content up to a certain GB. Doesn't matter whether you're on a PC-based or MAC system. None of them can handle a 387GB video file (believe me, I've tried).

But in terms of video restoration, which is what this is about, the point is having re-captured lower resolution video as a HD video file, the artifacts (missing pixels) are clearly identifiable, even moreso. However, let's say I have a captured avi stream at 1080i/p from your original 480/360p video .avi file. If I were to apply filters to clean it up best as possible and render it out as a 1280/720p file, for reasons I haven't discovered why yet, I don't see as much artifacts/pixilization occurring.

Make sense?

[sanlyn]

I think that what DB83 is getting at, if somewhat obliquely, is that you're ruining your videos. Never -- I repeat -- never capture those sources to low-resolution lossy formats, then process them as lossy in NLE's that are not designed for proper and thorough analog source cleanup. Period. End of thread.

[diGriz]

You can't ruin a video file.

End of thread.

[jman98]

I'll assume that should be translated as "What the hell are you talking about?"

Back in the 70s when VHS was developed for the market, broadcasters intended to use the medium for audio recording since it's a wider magnetic strip.

I really have my doubts about this. First of all, it's unclear why BROADCASTERS would care about recording audio and not video, or at least that's what your statement seems to imply. It's known that VHS came out of a serious attempt by JVC and others to invent video tape TV recorders that could be sold to consumers. My understanding is that using VHS tape for audio recording had more to do with this being a cheap and practical way to record early digital audio input than the VHS format being designed specifically for this.

Your posts strike me as odd, but perhaps I just don't understand your point. I'm not aware of any capture cards that will allow VHS video to be captured at HD resolutions, but it seems that you've found some way to do it (i know absolutely nothing about DV, so maybe that's involved). Even so, I'm not sure that this is necessarily a good thing. It seems that in a whole lot of verbiage you are claiming that you've found a way to magically transform lossy SD AVI captures into perfect HD captures. I have my doubts about this too.

[sanlyn]

You can't ruin a video file.

End of thread.

You managed to find ways to do it. End of diGriz's videos.

[smrpix]

There was a brief attempt to use VHS for audio, but came AFTER it had been developed for home video. I believe it was called ADAT and was quickly superceded by DA-88. Your information is simply wrong.

edit: looks like jman beat me to this.

Also, the bottleneck for the NLEs is hardware throughput, not software capability. Premiere can handle RED and Alexa 4K files now, given the right platform. I thnk FCPX can now. Avid's current limitation (until next month)is raster size, not bitrate as it can handle multiple streams of 220bit media in real time.

[Cornucopia]

The OP's take on history is biased & revisionist to say the least, but more likely is just plain WRONG!

Back in the '70s, the only professional videotape formats were 2" QUAD and 1" B and 1" C formats, plus the Semi-professional 3/4" U-matic format. BetaMAX is the 1/2" consumer offshoot of U-matic. JVC wanted in on the consumer market also, so they created their own competing format (since they didn't want to license their IP from Sony), called "Video Home System" (aka VHS). All these formats for video-centric, with audio just as an afterthought.

Don't fast-forward to today, because you miss the actual progression of things!

In the (late 70's &) 80's BetaCAM came along, using the same 1/2" cassettes that Betamax uses, but run at ~3x the speed to accommodate a better quality picture (component, full SD). JVC countered with the "M" system (using VHS-style cassettes). One the consumer side, Betamax was upgraded by Sony to SuperBeta (better luminance range, better resolution). Then Sony upgraded BetaCAM, with improved (but backward-compatible) specs in the recording & tape materials, to become BetacamSP. JVC Countered with M2. Sony then added a ProSumer wing to capture the middle ground and created EDBeta ("Extended Definition") as an extension of SuperBeta. JVC Countered with S-VHS. Both used separated Y-C recording for improved luminance range, resolution, and color clarity. There was also the whole Sony video8/Hi-8 line, with no competitor.
Somewhere along the way there, VHS & Betamax/SuperBeta also developed HiFi audio recording (using FM analog audio on a sub-layer to the video).
In NO case did any of those analog formats EVER use digital audio.

Digital audio was originally focused on using DASH reel-to-reel recorders in the 70's & 80's. There was also in the early 80s an adapter device at that time that could be used to convert a stereo digital audio signal to pulses on a video picture so that the video picture could be recorded (usually by U-matic, but also by S-VHS) & later played back & decoded as pristine digital signals (I used this device). Then in the late 80s, DAT (digital audio tape) cassettes came along for professional and prosumer use, and was very successful.

The VHS line did graduate in the '90s to include W-VHS (analog hi-def and RARE) and later D-VHS (digital, compressed data & sd & hd video, and still occasionally seen today). BetacamSP upgraded to Digitbeta & BetacamSX, and on and on...

On the audio front, there were side forays into consumer tape formats (DCC), but by the mid/late 90s consumers were more used to working with disc formats (recordable CD & MiniDisc). In the 90s there were uses of tape for Multitrack digital audio (usually 8 tracks) for prosumer use (esp. in musician & filmmaker circles) in the form of ADAT & DA-88. Those were competing formats, both of which were NOT brief and still survive to some degree today! Since then, even those tape formats have really given way to disc-based (incl. DVDAudio/SuperAudioCD+BD formats) or flashmemory/solidstate-based formats (with up to 24 tracks).

************************************************** *********************

On to capturing VHS (and other SD media) using HD capture methods: This is pointless. Regardless of colorspace/subsampling etc., all you are capturing (beyond a standard SD capture) are additonal noise or upsizing artifacts. Analog SD formats should be captured in SD. Analog HD formats should be captured in HD. Film-based formats are neither SD nor HD, they are ~ED on the low end (with 8mm/Super8) and ~2k or ~4k on the high end (35mm). Won't go into resolution for 65/70mm/IMAX.
Use the right tool for the job. And you are missing out dismissing good tips from Sanlyn...

Scott

[Clockwork]

Excellent write-up, Cornucopia!
Lots of stuff I didn't know!

[lingyi]

@Cornucopia... +1 on the great write up / corrections on the OP’s posts.


A little addition to show some Betamax love. J As I recall, the PCM-F1 was the first commercially available digital recording adapter, circa early ‘80’s. Paired with the portable SL-2000, it was used for some of the first full digital recordings.


@diGriz

“Back in the 70s when VHS was developed for the market, broadcasters intended to use the medium for audio recording since it's a wider magnetic strip. “


??? – As jman98 points out, why would broadcasters be concerned with broadcast audio quality. Stereo broadcasts was years away and they were happily using the far superior 2” and 1” formats for their purposes. Even the ¾” UMatic was marginal broadcast quality at best.


If by “wider magnetic strip” you’re referring to the ½” tape format, audio recording studios were using 1” (8 track) & 2” tapes (16 track) at up to 30 fps with an allocation of ¼” per track, so there would be no increase in fidelity with a 2 track ½” tape.


The addition of Hi-Fi stereo to VHS was a response to Beta Hi-Fi and required adding an additional modulated analog signal under the video track. {edit: Already referenced by Cornucopia"}


“VHS decks were easier to produce, at an affordable price (vs. a broadcast standard Betacam SP desk).”


As Cornucopia states, Betacam SP was a further refinement of Betamax which followed VHS .The original VHS format was never “broadcast standard” quality nor was ever intended to be.

[2Bdecided]

There was also in the early 80s an adapter device at that time that could be used to convert a stereo digital audio signal to pulses on a video picture so that the video picture could be recorded (usually by U-matic, but also by S-VHS) & later played back & decoded as pristine digital signals (I used this device).

Usually Betamax here. Never S-VHS here - it worked just fine with VHS, and pre-dated S-VHS by several years.

It's also a little more important that you imply - the whole reason we have 44.1kHz sampling on CD is to allow the recordings to be made on video tape (6 16-bit samples per usable video line). That's how CDs ended up with such a weird sample rate (it "should" have been 48kHz or 50kHz).

Cheers,
David.