VHS to Digital / TV Broadcast Quality

[Cornucopia]

@ConsumerDV, re: Quad - yes. They pretty much edited via splicing. It is even possible to do with VHS (I have done this). It is easier to do with quad and to a lesser extent with helical-scan 1"B/C because the tape speed is higher, so the slant is much less. On VHS it is a b**ch!. (I only did this because this was the only way to salvage a couple of important but already damaged pieces, as copying wouldn't give a good enough image to survive the process)
The introduction of electronic editing quickly led to the introduction of timecode and again quickly led to the TBC, all within a couple of years in the early 70s (doing this from memory, so the order might not be right).

I agree, consumer capturing CAN SOMETIMES be done without the need for a tbc, but lots of factors have to already be in one's favor (1st generation, clean master tape...), and one's tolerance for inaccuracies shouldn't be too tight.

I think most of us are in agreement about the expected uncompressed SD bitrates (within a certain range), so basically it comes down to whether the combination of USB + PC device can reliably, and continuously sustain the required bitrates necessary for a clean, uninterrupted capture.
For USB 2.0 devices, I have seen this fail at expectations at least as often as I have seen it succeed.
USB 3.0+, SATA, expresscard, PCI-based all no problem. Firewire - fine for DV, and a few other specialty transfers.
But my expectations ARE quite high.

Scott

[ConsumerDV]

A credible help forum should be built on fact and not on fallacy.

I absolutely agree.

After the fact we now have a link that preports to confirm the true formula yet even that fouls the issue by using 24 bit in its examples.

The formula does not use 24 bits. The formula uses "color depth" variable; you substitute an actual value for it to obtain an expression that you can evaluate. For 4:2:2 the effective color depth is 16 bits: 8 bits for luma, and 4 for each chroma component. Here: 16 b × 720 × 480 × 30 s^-1 = 1.65E08 b/s = 165 Mb/s ≈ 158 Mib/s

But pray tell where that mysterious wiki entry that quoted 173 (and I do not give a damn - better to paraphrase Abott and Costello - if this is Mb or MB)

Don't you think that your not giving a damn whether it is Mb or MB contradicts with the idea that a credible help forum should be built on fact and not on fallacy? Here.

I agree, consumer capturing CAN SOMETIMES be done without the need for a tbc, but lots of factors have to already be in one's favor (1st generation, clean master tape...), and one's tolerance for inaccuracies shouldn't be too tight.

I would replace "sometimes" with "often" or "in most cases". Also, when a regular Joe captures a personal Hi-8 tape or a factory-produced VHS movie, this is the 1st generation tape you are talking about.

USB 3.0+, SATA, expresscard, PCI-based all no problem.

I use a device designed for USB 2 running on USB 3 - no problem as well

But my expectations ARE quite high.

Understood. Yet, tons of people upload videos on YouTube that are not correctly deinterlaced and have wrong aspect ratio. This is what these people need help with first and foremost.

EDIT: Re-reading the thread title, maybe I should expound my opinion about TBC-less capture being acceptable in a different thread.

[dellsam34]

@dellsam34, attacking forum members and information sources will not earn you points.

I'm not attacking anyone, you are the one claiming to be Mr. knows it all with a google search diploma, You were called out on your BS numerous times but you still doing it over and over, Even the active threads are not enough for you, you went out of your way and resurrecting a decade old threads to add a useless post.

[dellsam34]

In a word or three many of the replies give no credibilty to the forum and merely serve to massage the individual egos of the contributors (mine included). Or to put it another way, the topic 'died' two years ago and there was no justification in resurecting it.

I gave a perfect example from real life, my own captures that are around 200GB an hour, from there it's easy to figure out how much data we are talking about, But the experts of google search have no purpose on this forum but to look smart, If you ask them to actually capture one hour and then do the math they give you the run around.

[ConsumerDV]

@dellsam34, this is getting boring. I don't care what you think about me or wikipedia or google, so you can skip personal insults. Please clarify these statements or take them back.

60MB/s can barely pass full lossless 4:2:2 AVI even at 8bit.

First of all, I haven't said anything about 173MB/s. I said 173Mb/s.

Just because your wikipedia diploma says 173mb/s, doesn't make it 173 mb/s. It's 173 MB/s.

I also would appreciate if you stopped switching from 8 bits to 10 bits, from USB dongles to SDI - no one is going to capture 10-bit SDI through a USB dongle. EDIT: actually, there are quite a few USB 3 dongles with SDI input, nice!

[dellsam34]

I'm not switching from 8bit to 10bit, The figure I gave you is for 10bit from spec sheets. Yes, you can't capture SDI stream via USB 2.0, You would need USB 3, Thunderbolt or PCIe, SDI to USB adapters wouldn't even show up via USB 2.0.

[lollo]

you went out of your way and resurrecting a decade old threads to add a useless post

It was not useless. We had this pearl of wisdom:

Frame TBC is a later development. Its functionality can be replicated in an NLE.

[DB83]

Well that wiki link that quotes the mysterious 173 actually states "citation needed". So the writer could have plucked that number right out of his a^&£

And here would it not be wiser not to abbreviate but use the full term of Mbit. Then noone should be confused between Mb and MB

[Alwyn]

And here would it not be wiser not to abbreviate but use the full term of Mbit.

No DB, bits and Bytes have been clear for eons.

[lordsmurf]

Frame TBC is a later development. Its functionality can be replicated in an NLE.

NLEs have nothing to do with TBCs, and cannot at all replicate what is done. What you just write is (currently) the dumbest thing I've ever read, as far as what TBC is and does -- quite a feat.

But for someone who wants to capture a dozen of tapes, a TBC is not an essential item,

Nonsense. At the very least, use the ES10-15 for passthrough, and cross your fingers that it will work for all your tapes. Depending on factors, you may be able to skate by. Or not.

USB 2.0 throughput in Hi-Speed mode is 480 Mbit/s, w

Good luck with that. Theoretical only, and not what happens in practice/actuality. USB2 usually is acceptable for SD captures, but there are factors there as well. More modern 2010s systems are generally fine, the 2000s are very suspect. OEM brand pre-made systems like HP, Dell, etc, tended to also not work well, or at all. You must understand that USB runs over the CPU, so you can't have a POS processor and expect results.

frame TBC is needed for consumer capture cards that use the ancient USB 2.0, .

Wait, no. You narrow down USB here, but it applies to any comm method, be it AGP, PCI, PCIe, etc. And both pro and consumer as well. What you are comparing all these to is specific SDI cards, which I (we?) know you use, and like, and that's fine. Like anything else, SDI has drawbacks, maybe benefits.

USB 2.0 is limited to how much data can pass through it so most of the time the capture software is unable to receive the frames on time resulting in out of sync audio to video, Capture cards that are built in frame synchronizer or frame TBC don't suffer from this problem

Hmmmm .... eh. What I think you refer to here is possibly the tendency to drop frames more quickly over USB, as the I/O is somewhat more finite that internal buses. I don't see that as a factor, however, as a bad input signal will drop either way. Just maybe more drops externally. Either way, drops are bad, use a TBC or TBC(ish).

It probably is not a good idea to boast about "just a couple of dropped frames per hour". Doing it right with both line & fullframe TBCs and good decks and capture gear should mean NO dropped frames for days on end, which is what I experienced and expect.
Scott

I never understand that mindset. "It's only partially broke, it's still fine." We all know that sort of mindset of born from being a cheapskate, and nothing else.

Stop showing your ignorance please. It is simple math calculations on primary school level:
720x480x30 fps x24 bits=248 832 000 bits/s
248 832 000/8=31 104 000 Bytes/s
31 104 000/1024=30 375 KBytes/s
30 375/1024=29,66 MB/s=237,28 Mb/s - full uncompressed RGB quality.

Bad math. Bad info all around, actually (RGB, etc). No.

There would be a good percentage of people who would listen to suggestions, who want to do a good job, but don't know where to start. I said "good" because there would be a lesser number that are willing to spend 1000s (even if you can try to resell it) on the "only" setup to use, as espoused by some here. My way or the highway comes to mind.

Is this aimed at me? Because I know you know better. Gear is available at multiple budgets. But budgets can sometimes get stupidly cheap (example: "I have $100 to spend" won't cut it). You can easily run into $K's, but then you can also resell it when the project is done. If you use/buy cheap gear, resale will be much harder, odds are it's yours forever. The TBC is always the most costly item, so you can swap it out for a TBC(ish) type item, such as the ES10/15. But that has downsides, so be aware. Some folks care about quality more than some bucks (which can be returned after resale).

@ConsumerDV, I am not a fan of either Dazzle or AverMedia. Both come from quick-to-market wholly-consumer-focused enterprises. Both are passable for non-critical consumer application, but I would consider both to be unacceptable non-starters for any professional use.
Scott

Dazzle is often unacceptable for consumer use. Why? Because it messes with luma and chroma too much. So badly washed out videos, etc.

A credible help forum

More and more, we're seeing total BS being passed off around here. It sickens me. It wasn't like this 20 years ago, even from the likes of kwag and naked-geek. At least they kept their BS to one or two tropes, not an orgy of misinformation all over the place.

[lollo]

Frame TBC is a later development. Its functionality can be replicated in an NLE.

NLEs have nothing to do with TBCs, and cannot at all replicate what is done. What you just write is (currently) the dumbest thing I've ever read, as far as what TBC is and does -- quite a feat.

As I said, a pearl of wisdom. If our friend jmac698 is still around he can steal the code for his attempt to build a software TBC!

[Pinto007]

Stop showing your ignorance please. It is simple math calculations on primary school level:
720x480x30 fps x24 bits=248 832 000 bits/s
248 832 000/8=31 104 000 Bytes/s
31 104 000/1024=30 375 KBytes/s
30 375/1024=29,66 MB/s=237,28 Mb/s - full uncompressed RGB quality.

Bad math. Bad info all around, actually (RGB, etc). No.

If you claim, that my example is bad math, then explain WHY. Empty words cost nothing.
Think first, before you write something. Even Cornucopia gave a formula, so stop insulting me.

Some technical clarification for this thread...
For fully uncompressed video: H rez * V rez * bitdepth/color primary * framerate = bitrate.

[DB83]

And here would it not be wiser not to abbreviate but use the full term of Mbit.

No DB, bits and Bytes have been clear for eons.

You miss the point. And I do know my bits from my bytes.


But there will be readers looking for assistance who will inevitably be confused with terminology that sails over their head. They deserve a clear, non-complicated, some may say 'hand-held', approach.


And in the current example just stating MB or Mb could easily create the wrong impression.

[ConsumerDV]

It was not useless. We had this pearl of wisdom:

Frame TBC is a later development. Its functionality can be replicated in an NLE.

Wrong choice of words. I apologize profusely. I meant that some issues that frame TBC would take care of, like a rolled or dropped frame, can be fixed in an NLE by replacing individual frames. I did not have anything more serious than several dropped frames per hour, which did not stop a capture, so not having a frame TBC is not a show stopper for me.

Well that wiki link that quotes the mysterious 173 actually states "citation needed". So the writer could have plucked that number right out of his a^&£

Do you really care whether it is 160, 170 or 180? You can calculate the value yourself and add 5%-10% overhead for error correction and signaling, but this makes no difference in the context of 480 Mb/s data pipe.

But for someone who wants to capture a dozen of tapes, a TBC is not an essential item,

Nonsense.

As you know from basic logic, a single counter-example is needed to refute a claim, and since I don't have serious issues running my captures without a TBC, I therefore have disproven the claim that TBC is an essential item. Q.E.D.

Stop showing your ignorance please. It is simple math calculations on primary school level:
720x480x30 fps x24 bits=248 832 000 bits/s
248 832 000/8=31 104 000 Bytes/s
31 104 000/1024=30 375 KBytes/s
30 375/1024=29,66 MB/s=237,28 Mb/s - full uncompressed RGB quality.

Bad math. Bad info all around, actually (RGB, etc). No.

Nothing is wrong here. RGB is non-essential. The result works for 8-bit per component RGB or for 4:4:4 YUV. For 4:2:2 it will be 16 bits/pixel effective bit depth. For 4:2:0 and 4:1:1 the effective bit depth will be 12 bits/pixel. Your dismissive attitude does not help driving your message across.

Dazzle is often unacceptable for consumer use. Why? Because it messes with luma and chroma too much. So badly washed out videos, etc.

Which Dazzle? They had many different models. I use DVC100 and I haven't noticed that it "messes with luma and chroma too much". Even if it did, correcting levels and color is trivial after a video has been captured.

More and more, we're seeing total BS being passed off around here. It sickens me.

You don't say. Why do you keep coming here instead of sharing wisdom on your own forum?

And in the current example just stating MB or Mb could easily create the wrong impression.

Standard SI units and prefixes is common knowledge for anyone who graduated high school, in particular "m" is milli, "M" is mega, "Mi" is mebi, "B" is byte, "b" is bit.

[lollo]

Wrong choice of words. I apologize profusely.

Clear now, no need to apologize, I was just joking following dellsam34 comments

If you have many consecutive dropped frames is not easy to interpolate, also with AviSynth. I have sometime several bad consecutive frames in a segment and the final result after repair is not optimal, even with wonderful MVtools techniques. Sometimes it works, more or less: tmp.avi

edit: here the used code

# interpolate frame GMa_1 94327-94331
# a x     x    x    x    x     b
# a a     a    x    b    b     b
# a a     a    iab  b    b     b
# a a     iaab iab  b    b     b
# a iaaab iaab iab  b    b     b
# a iaaab iaab iab  iabb b     b
# a iaaab iaab iab  iabb iabbb b

	# replace frames 94327 and 94328 with frame 94326 and frames 94330 and 94331 with frame 94332
video_org_rep_zzz=video_org_rep4.trim(0,94326)++video_org_rep4.trim(94326,94326)++video_org_rep4.trim(94326,94326)++video_org_rep4.trim(94329,94329)\
++video_org_rep4.trim(94332,94332)++video_org_rep4.trim(94332,94332)++video_org_rep4.trim(94332,0)
video_org_rep5a=video_org_rep_zzz.interpolate_frame_GMa_1(frame_number=94329)
video_org_rep5b=video_org_rep5a.interpolate_frame_GMa_1(frame_number=94328)
video_org_rep5c=video_org_rep5b.interpolate_frame_GMa_1(frame_number=94327)
video_org_rep5d=video_org_rep5c.interpolate_frame_GMa_1(frame_number=94330)
video_org_rep5=video_org_rep5d.interpolate_frame_GMa_1(frame_number=94331)


function interpolate_frame_GMa_1(clip c, int "frame_number")
{
	# interpolate bad frame
sup = c.MSuper()\
bv = MAnalyse(sup, isb=true, delta=2)
fv = MAnalyse(sup, isb=false, delta=2)
interpolated = MFlowInter(c, sup, bv, fv)

#stackhorizontal(c, interpolated)
#return(interpolated)

	# replace bad frame with interpolated frame
video_repaired = \
	c.trim(0,frame_number-1)++\
	interpolated.trim(frame_number-1,-1)++\
	c.trim(frame_number+1,0)

	# video repaired
return(video_repaired)
}

[lordsmurf]

-example is needed

Dropped frames.
And dropped frames lead of audio desync.

Even if it did, correcting levels and color is trivial after a video has been captured.

Not correct. If values are too far off, it cannot be corrected in post.

Why do you keep coming here instead of sharing wisdom on your own forum?

I've been here for decades. Why are you here?

[ConsumerDV]

Dropped frames. And dropped frames lead of audio desync.

Very few dropped frames and no A/V desync with the DVC100. When capturing with VirtualDub I can see audio sync fluctuating 2-3 ms around the zero value but then returning back, so it does not increase. I captured several 2-hour movies and there was no noticeable desync at the beginning, in the middle or at the end of these movies. Acceptable for someone capturing a dozen tapes at home for personal use. Again, by providing a single counter-example I have refuted your blanket claim.

Even if it did, correcting levels and color is trivial after a video has been captured.

Not correct. If values are too far off, it cannot be corrected in post.

They are not too far off. And since you said that the videos are washed out, it means that black and white levels are not lost. But this has no relation to TBC.

Why do you keep coming here instead of sharing wisdom on your own forum?

I've been here for decades.

So it is an old habit, I see. Since you are here, could you explain what is wrong in the math posted by Pinto007? You said it had "bad info all around". I don't see any glaring errors in it, but I may be missing something, and I am always eager to learn. Share your knowledge, isn't it what you are here for? Thanks.

[lordsmurf]

bad info all around

This sums up most of your info, in most of your posts at this site, for the mere few months you've been here. I'm still amused how, on some topics, you went from deer-in-headlights newbie to instant expert, and others pointed out this ridiculousness. All of your claims in this post vary from wild, to dubious, to laden with weasel words (example: "very few" drops, when it should be zero, or near-zero). What I don't want to see is newbies being suckered by your nonsense.

I apologize profusely. I meant that some issues that frame TBC would take care of, like a rolled or dropped frame, can be fixed in an NLE by replacing individual frames.

That never works, for many reasons. For starters, you have no idea where the drop was. Avisynth can try to detect, but it's far from infallible here. Not everything can be corrected with a magic software "undo" button. Dropped frames are lost data, and at best you can try to dupe data into it. And that, in turn, incurs interlacing issues, etc. Again, messy, and false to claim otherwise.

[ConsumerDV]

Please, explain your "bad info all around" comment that you made in response to Pinto007's numbers. If he made an algebraic error, could you point to it?

As for acceptability of less professional methods, I am not a pro and I am satisfied with results that are less than perfect. I understand that for a pro like you this is not good enough, but this forum is not for pros, is it? This site "has always been more about copying discs (VCD/DVD/BD/etc), downloading videos, using "easy" lesser-quality methods (like "all-in-one" software), open-source software (anti-payware mentality), cheap hardware, etc." So I understand that as both a video professional AND a video hobbyist, that's not what you are about, which is why you have created your own forum for pros like yourself.

[dellsam34]

This forum is for everyone, hobbyist, pro, total newbie, single mom, average joe. Anyone needs help with A/V related stuff can come here and ask questions including ourselves. When someone asks a question the best answer is give them the technical help they need and the options available, you can't decide what is good enough for them, Their expectation and budget is their decision. Just because you don't want to use a TBC, a certain type of VCR or a piece of gear doesn't mean that someone who uses them is a pro. All this has really nothing to do with pro, math, school ... It's experience, practice, and learning from own mistakes so others won't have to make them.