4 different IODATA GV-USB2's tested with samples. Line-TBC-Like-Activity?

[Sharc]

Oh, the delicious irony. @Sharc, with his detailed analysis over recent times, has done more than @Lollo or I to promote the GV-USB2 as a good, modern digitiser.

To clarify: These "wiggle comparison tests" alone do not disqualify the GV-USB2 models at all. At least your clips also confirm that the GV-USB2 captures well into the full (.....254) luma range and do not hard clip at Y=235

[VWestlife]

Enough drama. Crack it open and show us if it still contains the Techwell TW9910 chip, or something different. If it's the same, then we're all just arguing about minor performance variations in the same hardware.

[Sharc]

Enough drama. Crack it open and show us if it still contains the Techwell TW9910 chip, or something different. If it's the same, then we're all just arguing about minor performance variations in the same hardware.

TW9910 produced by Techwell, Renesas or Intersil ....... oh well .....

[rgr]

- GV-USB2 in AmaRecTV does not properly report insert/dropped frames. That's because the internal mechanics of broadcast software is about continuity, not signal integrity. Dropping/inserting frames is an expected action of broadcast/streaming software. While AmaRecTV does some reporting, the exact reporting done, and why, is not clearly understood by anyone.

AmarecTV doesn't seem to report dropped frames correctly with any capture device, so it shouldn't even be used for testing.

[Sharc]

What does this have to do with the topic of this thread? Why do so many threads derail badly and end up with a sometimes childish debate of Vdub version xy vs AmarecTV? Everyone make his own experience. Finish.

[lollo]

AmarecTV doesn't seem to report dropped frames correctly with any capture device

False.

[rgr]

What does this have to do with the topic of this thread? Why do so many threads derail badly and end up with a sometimes childish debate of Vdub version xy vs AmarecTV? Everyone make his own experience. Finish.

Since the topic has been raised, I'm not interested in all the VirtualDub vs AmarecTV wars. I correct the claim that AmarecTV drops frames with I-O Data -- apparently, it drops with everything. That's why I ran my own tests, which no one can refute :P (being convinced that AmarecTV will be better)

[lollo]

I correct the claim that AmarecTV drops frames with I-O Data

That claim is from lordsmurf in post #19 is from that charlatan of lordsmurf, who does not know what is talking about, nor provide any fact about it.

apparently, it drops with everything

False.

That's why I ran my own tests, which no one can refute :P

Not confirmed by anyone, while tests from others including myself show the opposite.

[rgr]

Open to other suggestions for specific tests as well as it might be a while before I get to these.

I suggest numbering all the frames. This way you'll know if 10,000 frames will actually result in 10,000 frames.

[Sharc]

Open to other suggestions for specific tests as well as it might be a while before I get to these.

I suggest numbering all the frames. This way you'll know if 10,000 frames will actually result in 10,000 frames.

There are no "frames" recorded on a VHS tape. Just sequences of scanlines and fields and sync info stored in the form of analog waveforms on a magnetic medium/carrier. So how would you "number" this on the tape?
But again, totally off topic.

[lordsmurf]

I suggest numbering all the frames. This way you'll know if 10,000 frames will actually result in 10,000 frames.

There are no "frames" recorded on a VHS tape. Just sequences of scanlines and fields and sync info stored in the form of analog waveforms on a magnetic medium/carrier. So how would you "number" this on the tape?
But again, totally off topic.

Running time codes in display existed 50 years ago.

[Sharc]

I suggest numbering all the frames. This way you'll know if 10,000 frames will actually result in 10,000 frames.

There are no "frames" recorded on a VHS tape. Just sequences of scanlines and fields and sync info stored in the form of analog waveforms on a magnetic medium/carrier. So how would you "number" this on the tape?
But again, totally off topic.

Running time codes in display existed 50 years ago.

Yes, but this is not "numbering the frames" - or I misread rgr's proposal.

Edit: My bad. He (@rgr) referred to aramkolt's planned synthetic test videos/test patterns, rather than some existing tapes.

[lordsmurf]

lordsmurf in post

Can we please just table our past disagreements here?
Please?
We're getting distracted.
I have questions, and I would like your input, too.

Obviously you like the GV-USB2 original hardware. Let's pretend I like it to. Now we agree.

And now we have concrete proof that the "Win11" version has changed. Can we agree? I'm using quotes because we're still not 100% sure that the Win11 designation is the reason for card differences. In fact, can we be sure it's not a fake GV-USB2? We do know those exist, though generally just Easycaps. And IO-Data has disavowed any knowledge of changes.

I hate the sample clip from the OP. But I can still see errors, if I look.
However, instead of these clips here, look at Sharc's sample from another thread
https://forum.videohelp.com/threads/420448-GV-USB2-vs-ATI-TV-Wonder-600-USB#post2796068

That "Win11" version has definite geometrical distortions, and cannot be used without TBCs. At least the old version let you record un-distorted, even if crappy wiggles. I would rather have wiggles than "stable" video that turns straight objects into skewed odd shapes. I hate uncanny valley, unnatural looking video.

Can we agree something has changed, and not for the better?

It's misleading, and a step backwards for capture cards. It's already led to a myth (started on Youtube) about TBC where none exists. Can we agree here too?

I'm really not enjoying this conversation whatsoever, but I think it's important, and we keep getting distracted by off-topic quarrels.

[Sharc]

However, instead of these clips here, look at Sharc's sample from another thread
https://forum.videohelp.com/threads/420448-GV-USB2-vs-ATI-TV-Wonder-600-USB#post2796068

Sidenote only: That nasty clip was originally submitted by @Darryl In Canada. So the credits for it go to him. I just picked that scene to show that things are not always as they appear to be at a first glance - rather than being a proof or personal preference for anything.

[Alwyn]

And on and on we go...

can we be sure it's not a fake GV-USB2?

Of course we can be sure. Aramkolt's got 4 of them, I have 3, and you saw a couple of ads that had GV-USB2 labels on Easycaps. Stop carrying on about it.

And IO-Data has disavowed any knowledge of changes.

You have been scathing over the years of company personnel's description of their products. Why are you suddenly treating IOData's comments as gospel?

That "Win11" version has definite geometrical distortions

For someone who rants and raves about facts and science, what a ridiculous statement. You have no idea what condition the tape is in. The test was done with Aramkolt's' "absolute worst VCR". And if you can't see the straightening (as good as the ES-15) on my test you need your eyes checked; even my half-blind grandma could see the improvement, to borrow some of your catastrophe language.

For you to carry on like a pork chop trying to discredit the "new" version is pretty amateurish.

You're clutching at straws in your continuing crusade against the GV-USB2. By all means present substantiated facts, like you've demanded others do, but you're currently not doing that.

And no, I'm not suggesting people forgo the use of an external stabiliser.

[lollo]

lordsmurf in post

Can we please just table our past disagreements here?
Please?
We're getting distracted.
I have questions, and I would like your input, too.

Sure. Let' start.

Premise 1: You own yourself an IOData GV-USB2, so I do not understand why are you basing yourself on experiments performed by others rather than your own.

Premise 2: I suspect that the ultimate purpose of your request is to belittle the performances of the GV-USB2, as you did so far in all of your posts without posting any evidence; now that we have some experiments, you want to jump to hasty conclusions. I hope to be wrong here

In the last 4 year I bougth at least 10 GV-USB (and Hauppauge USB-Live 2), not all for me, of course, but for people in my fan organization willing to capture VHS/S-VHS of our beloved sci-fi shows, but not only.
I bougth the latest two months ago, a so-called "Windows 11 version", pictures below:



1- You are basing your consideration about quality of the capture in experiments where the input signal is not time base corrected, so not an ideal workflow, leading nowhere for additional conclusions. The goal of the experiments from aramkolt and DarrylInCanda was to understand if GV-USB has TBC capabilities in its last verion. It has not.

2- Extrapolating the behaviour of the GV-USB in the scenario described above, has little sense, but for what is worth in my experiments I do not see a major deviation in quality between the so-called "Windows 10" version (better to call it "GV-USB2" as per IOData documentation) and the so-called "Windows 11" (better to call it "GV-USB2/HQ" as per IOData documentation) version of the GV-USB2 card (framing the active picture apart).
Thereafter a comparison:
comp.mp4

I did not open the cards to check the ICs inside as someone suggested, so my judgements is based on the outcomes of the two cards, not on a definitive visual inspection.

3- The "GV-USB2/HQ" version does not allow to use the latest 115 version, but only 111, 112 and 114 version. That's strange because the last card released does not allow the latest driver to be used. The reason is that the bundled software included is not available to all releases. I am talking about drivers because I experimented the do play a major roles, as in the Hauppauge USB-Live 2 card. Unfortunately.

4- The geometrical distortions you are talking about is present in one experiment in a workflow without a TBC in the VCR. First, we will never capture in that condition, second the card may react well or not well to unstable signal (intra-frame line integrity, or better a missing correction of what you call a "line-TBC"). It is a marginal result, similar or different accordingly to random behaviour.

Conclusions:
- does the GV-USB2/HQ version feature a different hardware compared to GV-USB2 version? No according to the manufactures and to my experiments.
- does the GV-USB2/HQ version add time base correction compared to GV-USB2 version? No according to posted results (and to my experiments)
- does the GV-USB2/HQ version introduced geometrical distortions? If the hardware is unchanged as I suspect, the answer is no. It may be just a sporadic/random result
- is it important if GV-USB2/HQ version introduced geometrical distortions for unstable signals? Assuming that I am wrong, and the GV-USB2/HQ sistematically introduces an unwanted effect, it is really not important anyhow, because we do not recommend to capture without a time base correction. If you look to the ATI USB 600 versus IOData GV-USB2 comparison captures you mentioned, both are crap.

And finally, why on earth you do not run your own test on the USB-600, 710 USB, Live-2 and GV-USB2 cards as I did, and find yourself the truth? And while there, use AmarecTV as well, so you'll stop your non sense about it.

I hope that I have answered all your questions/doubts (as I wrote, I am in a rush lately, so cannot go into details with deep analysis of the videos nor additional specific tests of capture without TBCs). If not just ask

[VWestlife]

I bougth the latest two months ago, a so-called "Windows 11 version", pictures below:

Please crack it open and show us what's inside. With a little care you can do so non-destructively, and then easily snap it back together. (I did it myself, just to confirm that mine contains the standard TW9910 chip.)

[CaptureCraft]

I bougth the latest two months ago, a so-called "Windows 11 version", pictures below:

Please crack it open and show us what's inside. With a little care you can do so non-destructively, and then easily snap it back together. (I did it myself, just to confirm that mine contains the standard TW9910 chip.)

"How dare you bring logic into God's house!"

[lordsmurf]

Premise 1: You own yourself an IOData GV-USB2, so I do not understand why are you basing yourself on experiments performed by others rather than your own.

I don't have this "Win11" version. I bought my current unit several years ago.

Premise 2: I suspect that the ultimate purpose of your request is to belittle the performances of the GV-USB2, as you did so far in all of your posts without posting any evidence; now that we have some experiments, you want to jump to hasty conclusions. I hope to be wrong here

<sigh> Again, let's table all that past disagreement, and focus on this new odd unit.

people in my fan organization willing to capture VHS/S-VHS of our beloved sci-fi shows, but not only.

Side topic: Wait, what? What org? PM me!

I bougth the latest two months ago, a so-called "Windows 11 version", pictures below:
I did not open the cards to check the ICs inside as someone suggested, so my judgements is based on the outcomes of the two cards, not on a definitive visual inspection.

Good, good. Are you able to open it? Let's get some visual confirmation.

1- You are basing your consideration about quality of the capture in experiments where the input signal is not time base corrected, so not an ideal workflow,

Yes. That is the experiment here.

leading nowhere for additional conclusions. The goal of the experiments from aramkolt and DarrylInCanda was to understand if GV-USB has TBC capabilities in its last verion. It has not.

I don't think those are separate. Correct, no TBC. But then what other people think is a TBC clearly is something different. If that sort of of PLL effect? I've never seen anything behave quite like this from a mere PLL. There's got to be further processing somewhere. It creates "wiggle lock", with side effect of geometric distortions.

In one of the posts, you said you knew all about PLL. Okay. Then what's your diagnosis here? johnmeyer weighed in elsewhere, though briefly, and I wish he'd flesh that out more. I'm not unknowledgeable on these things, but I want to hear input from others. I know you're not a dummy, you know stuff, even if we disagree at times (and really only at like 2 things).

2- Extrapolating the behaviour of the GV-USB in the scenario described above, has little sense, but for what is worth in my experiments I do not see a major deviation in quality between the so-called "Windows 10" version (better to call it "GV-USB2" as per IOData documentation) and the so-called "Windows 11" (better to call it "GV-USB2/HQ" as per IOData documentation) version of the GV-USB2 card (framing the active picture apart).
Thereafter a comparison:

Now that myself and Sharc have pointed out the geomtry problems, can you try to see if you see it? Look for it.

3- The "GV-USB2/HQ" version does not allow to use the latest 115 version, but only 111, 112 and 114 version. That's strange because the last card released does not allow the latest driver to be used. The reason is that the bundled software included is not available to all releases. I am talking about drivers because I experimented the do play a major roles, as in the Hauppauge USB-Live 2 card. Unfortunately.

Interesting.

4- The geometrical distortions you are talking about is present in one experiment in a workflow without a TBC in the VCR.

So you are seeing it then?

First, we will never capture in that condition, second the card may react well or not well to unstable signal (intra-frame line integrity, or better a missing correction of what you call a "line-TBC"). It is a marginal result, similar or different accordingly to random behaviour.

Well, yes, it's bad capture practice. But others are doing it, due to false assumption that it has TBC. That's really the entire point of the thread.

- Do at least 2 versions, documented, now exist?
- Does the "Win11" version, which IO-Data seems to disavow, have geometry distortions from a "locked phase" (release/renew pattern). We can worry about proper term later, we're just informally talking here.

Conclusions:
- does the GV-USB2/HQ version feature a different hardware compared to GV-USB2 version? No according to the manufactures and to my experiments.

Well, let's crack open cards, visually verify this. I want to see yours, mine, anybody here.

- does the GV-USB2/HQ version add time base correction compared to GV-USB2 version? No according to posted results (and to my experiments)

Good, same conclusion as the rest of it. And honestly, it was a daft hypothesis, suggesting a sub-$100 card had a TBC.

- does the GV-USB2/HQ version introduced geometrical distortions? If the hardware is unchanged as I suspect, the answer is no. It may be just a sporadic/random result

I think it's too dismissive. Look for errors. I plan to look for a Win11 version card.

- is it important if GV-USB2/HQ version introduced geometrical distortions for unstable signals? Assuming that I am wrong, and the GV-USB2/HQ sistematically introduces an unwanted effect, it is really not important anyhow, because we do not recommend to capture without a time base correction. If you look to the ATI USB 600 versus IOData GV-USB2 comparison captures you mentioned, both are crap.

I fully agree with this. But the problem is other people will say "the card has a TBC, so I don't need anything else now", which is itself the real crap lazy conclusion.

And finally, why on earth you do not run your own test on the USB-600, 710 USB, Live-2 and GV-USB2 cards as I did, and find yourself the truth? And while there, use AmarecTV as well, so you'll stop your non sense about it.

We should PM sometime. It's easy for people to snipe at each other online, and not realize what the person does offline, or elsewhere online.

I hope that I have answered all your questions/doubts (as I wrote, I am in a rush lately, so cannot go into details with deep analysis of the videos nor additional specific tests of capture without TBCs). If not just ask

Same here, busy, too busy. I hope your days go better than mine do lately.

[Sharc]

Well, let's crack open cards, visually verify this. I want to see yours, mine, anybody here.

Just keep in mind it's always the combo of Chip + Driver which produces the result. The Driver/firmware is typically responsible for the active picture positioning within the captured frame, whether the top and bottom half scanlines are correctly shown or masked, what settings one can select from, whether - as I have been showing for another model - there is AGC/auto-contrast action or not being triggered by specific conditions, etc. So I won't crack my old GV-USB2 just to confirm there is a TW9910 in it.

[lordsmurf]

Well, let's crack open cards, visually verify this. I want to see yours, mine, anybody here.

Just keep in mind it's always the combo of Chip + Driver which produces the result. The Driver/firmware is typically responsible for the active picture positioning within the captured frame, whether the top and bottom half scanlines are correctly shown or masked, what settings one can select from, whether - as I have been showing for another model - there is AGC/auto-contrast action or not being triggered by specific conditions, etc. So I won't crack my old GV-USB2 just to confirm there is a TW9910 in it.

Yes, it's a combination of "all of the above".
But knowing the chip, without guessing, really is important in these situations. I've opened dozens of USB capture cards over the years. I only had one break, an ATI 600 USB that was already not working.