Well, VHS capturing Mostly cartoons, having some questions:)

[sanlyn]

When you get the new cable post some samples with the old cable and the new cable.

The failure mode with HDMI cables is extremely obvious. You get sparkly pixles with small errors. As errors get worse you get a complete loss of signal. You can't get blurring, ghosting, light noise, contrast loss, color shifts, etc. It's impossible.

http://news.cnet.com/8301-17938_105-20056502-1/why-all-hdmi-cables-are-the-same/
http://reviews.cnet.com/8301-33199_7-57321956-221/why-all-hdmi-cables-are-the-same-part-2/

Yes. I've seen those. Long ago. Similar articles from CNET turned me off to them for technical advice a long time ago. It's easy to angle a "test" when you already know what conclusion you want. Other sources differ.

[zerowalker]

But are there proof?
I would really like to see proof of it;S
as i can´t think of digital decoding differ from wire to wire;S

[jagabo]

Here it is, try to deinterlace it, without getting the Aim (Green) so leak to the left.

Are you talking about the way the green circle is blurred on the right edge? That's in the source.

I do see color blending of the green circle after QTGMC(), the red arrow points at it:



That is an error on the part of QTGMC. The color was too similar to the background so it didn't eliminate it. But that video is progressive. You should use:

AviSource("Leaking.avi") 
AssumeTFF()
SeparateFields()
Trim(1,0)
Weave()

That will be free of errors (except the blurred edge of the green reticule which is in the source).

[sanlyn]

Good point, jagabo.

[zerowalker]

it´s progressive?
Isn´t it interlaced?

and well, some parts in that game are not interlaced, and are progressive, can i use that all the time?
won´t that mess upp with the progressive parts?

(source is S-video)

[sanlyn]

But are there proof?
I would really like to see proof of it;S
as i can´t think of digital decoding differ from wire to wire;S

The proof has nothing to do with numbers. If you don't see or hear any difference, proof or no-proof is moot.

I think I got off-topic anyway. Jagabo gets me going every time we mention "wire". We're both wrong, and we're both right, depending on what we subjectively perceive. Never mind numbers. Trust what you see and hear.

Now, getting back to the thread . . .

[zerowalker]

Just checked, interlaced becomes fine, but progressive becomed interlaced, this makes things very complicated.
As i don´t think there is a way to make it sense when there may be interlaced or not, and i can´t really go through everything.

Btw how can you see when it may be solved with Weave and stuff, as you said it was progressive?

Sanlyn: was more wondering if you had "proof", more like, have you seen one cable that has bad colors or something, while another one has good?
As composite should have the same visual quality normally, hdmi should pretty much be the same, except it´s guarded as Digital i thought.
Would really like to know as knowledge is power

[sanlyn]

Sanlyn: was more wondering if you had "proof", more like, have you seen one cable that has bad colors or something, while another one has good?
As composite should have the same visual quality normally, hdmi should pretty much be the same, except it´s guarded as Digital i thought.
Would really like to know as knowledge is power

Yes, I've seen differences in cables: color, contrast, overall acutance, noise effects, and audio differences. Any metal wire can carry analog (variations in voltage and amplitude) or digital (streams of 0's and 1's) alike. There are people who string thick layers of Ethernet Cat5 wire to their speakers, strictly an analog circuit from start to end, and claim it sounds "terrific" (I thought it sounded tinny and thin). When you get a chance, click and browse some of the review details I linked to earlier at WhatHiFi. Try, for example, some low-rated wire remarks and some remarks with high-rated wire. That will give you some ideas about the kinds of subjective things people notice about images and sound. And as I said: if you don't see or hear any differences, it doesn't matter.

[jagabo]

Just checked, interlaced becomes fine, but progressive becomed interlaced, this makes things very complicated.
As i don´t think there is a way to make it sense when there may be interlaced or not, and i can´t really go through everything.

Then use TFM() instead of SeparateFields().Trim(1,0).Weave(). It will adapt to the progressive/interlace changes.

[zerowalker]

Will very well look into it, sounds interesting
you know any optical audio cable that is good, or you know that it sounds better in your opinion?


Thanks, searched around and only found these Full=false things, and it just made the deinterlacing worse;S


But this make me wonder, when should i use QTGMC() and separate/weave?

EDIT: is there a way to increase the detection, it fails at some occasions.
EDIT 2: Solved it, forgot the TopFieldFirst thing

[jagabo]

The proof has nothing to do with numbers.

Actually, with digital it has 100 percent to do with the numbers. You can verify that the data received at the end of the cable is bit for bit identical to the data that went in. Obvioiusly, you need to have the right equipment to do this.

[jagabo]

Thanks, searched around and only found these Full=false things, and it just made the deinterlacing worse

Did you see post number 399?

https://forum.videohelp.com/threads/335493-Well-VHS-capturing-Mostly-cartoons-having-so...=1#post2138846

[zerowalker]

Yeah i did, that is why i thanked you
As i had found something, but it didn´t do the Weave thing, but instead it used it´s own deinterlacing method.

But still, what is the difference, if weave beats out QTGMC why does it exist?
I guess it only works where it´s progressive being separated or something along the lines, but i guess you or others know more precise

[jagabo]

But still, what is the difference, if weave beats out QTGMC why does it exist?

The underlying video in that sample isn't interlaced it's 30p frames captured out of sync. Ie, instead of pairs of fields coming from the same frame, they are paired with a field from another frame. SeparateFields().Trim(1,0).Weave() pairs the correct fields back together again:

progressive frames: 1 2 3 4 5 6...
broadcast as fields TFF: 1t 1b 2t 2b 3t 3b 4t 4b 5t 5b 6t 6b...
captured out of phase BFF: 1b2t 2b3t 3b4t 4b5t 5b6t 6b7t...
separate fields: 1b 2t 2b 3t 3b 4t 4b 5t 5b 6t 6b 7t...
Trim(1,0): 2t 2b 3t 3b 4t 4b 5t 5b 6t 6b 7t...
Weave(): 2t2b 3t3b 4t4b 5t5b 6t6b 7t...

If the video is already progressive it makes it look interlaced because you are reversing the process.

[sanlyn]

Will very well look into it, sounds interesting
you know any optical audio cable that is good, or you know that it sounds better in your opinion?

Here are 3, actually from a "budget" range. With a friend in upstate New York, I've heard all of them for a brief spell. The QED sounded a tad richer, but all were such clean performers I'd have a hard time choosing. I have an earlier QED, and their budget HDMI (nice!). Acoustic Research also made a well regarded budget optical in both the blue-jacket and gray-jacket line (same price. I wonder why?). But I haven't seen them for a while.

These are UK prices:
http://www.hificables.co.uk/11628/True-Colour-Industries-Coral-Optical-Cable.html
http://www.hificables.co.uk/11739/Atlas-Equator-Fibre-Optics.html
http://www.hificables.co.uk/11538/Qed-Performance-Optical-Digital-Cable.html

The QED company has distributors in Switzerland and elsewhere, stores and mail-order all around Europe -- an old, reliable, quality line from low-budget to very expensive. All three brands are popular. Frankly, I'm just an outlaw (or maybe just an a/v snob), but I prefer coaxial. But many products are optical-only.

Back in 2002 I went nuts and bought a $145 VanDenHul optical + overseas shipping. Prestige brand, costly mistake.

[zerowalker]

But still, what is the difference, if weave beats out QTGMC why does it exist?

The underlying video in that sample isn't interlaced it's progressive frames captured out of sync. Ie, instead of pairs of fields coming from the same frame, they are paired with a field from another frame. SeparateFields().Trim(1,0).Weave() pairs the correct fields back together again:

out of phase: B1+T2 B2+T3 B3+T4 B4+T5 B5+T6
separate fields: B1 T2 B2 T3 B3 T4 B4 T5 B5 T6
Trim(1,0): T2 B2 T3 B3 T4 B4 T5 B5 T6
Weave(): T2+B2 T3+B3 T4+B4 T5+B5 T6...

If the video is already progressive it makes it look interlaced because you are reversing the process.

Ah i see, so it´s not interlacing at all?

But another thing, i thought composite/S-video couldn´t carry progressive?

These are UK prices:
http://www.hificables.co.uk/11628/True-Colour-Industries-Coral-Optical-Cable.html
http://www.hificables.co.uk/11739/Atlas-Equator-Fibre-Optics.html
http://www.hificables.co.uk/11538/Qed-Performance-Optical-Digital-Cable.html

The QED company has distributors in Switzerland and elsewhere, stores and mail-order all around Europe -- an old, reliable, quality line from low-budget to very expensive. All three brands are popular. Frankly, I'm just an outlaw (or maybe just an a/v snob), but I prefer coaxial. But many products are optical-only.

Back in 2002 I went nuts and bought a $145 VanDenHul optical + overseas shipping. Prestige brand, costly mistake.

Well those are very expensive;S
I thought optical didn´t really need shielding as it travels with light?
But i read that some high quality will be able to carry more bandwidth(something with the wave lenght i think).

The problem i have with mine, which costed about 5 dollar, is that it causes jitter from time to time. Not really the worlds end, but starting to become annoying.

Maybe it´s just my card failing to decode the light too, i read it can fail as it hasn´t got any error correction and it pretty fragile.

[sanlyn]

Well those are very expensive;S
I thought optical didn´t really need shielding as it travels with light?
But i read that some high quality will be able to carry more bandwidth(something with the wave lenght i think).

The problem i have with mine, which costed about 5 dollar, is that it causes jitter from time to time. Not really the worlds end, but starting to become annoying.

Maybe it´s just my card failing to decode the light too, i read it can fail as it hasn´t got any error correction and it pretty fragile.

There are no good really-cheap optical cables. Precision manufacturing of optical "tubes" is critical.
But if you'd like last year's QED model, medical grade, extremely popular, and a big improvement:
http://www.hificables.co.uk/12144/Qed-Profile-Optical-Digital-Cable.html
I notice these "Profile" models discounted recently, probably a version change coming next. That's when I usually purchase my own stuff.

[zerowalker]

Understandable, but is there a way to conclude it´s the cable?
Cause what if it´s just the adapters or decoding that makes the issue?

[sanlyn]

But another thing, i thought composite/S-video couldn´t carry progressive?

Not true. Wire is wire, it carries whatever you feed into it. Composite and s-video can carry interlaced and progressive alike, no problem. But because of the circuit design they are limited to standard definition, no hi-def. High Definition can be interlaced or progressive as well. HD uses HDMI or component.

. . .
jagabo, how does a video get dis-organized like that? Sounds as if it went thru multiple steps of some kinky process, or one step to do one thing and then another step that un-did the first step in the wrong way ? ?

[sanlyn]

Understandable, but is there a way to conclude it´s the cable?
Cause what if it´s just the adapters or decoding that makes the issue?

Try using the same cable in another setup.

[zerowalker]

Oh, well, why doesn´t game consoles and similar devices support 480p through composite?


EDIT: Have tried another conole with Optical, same, though i can´t try it out on another decoder as i haven´t got any.

[sanlyn]

480p? That's up to the design of particular consoles, isn't it? Progressive-scan SD DVD players are common, with composite outputs.

Likely jagabo knows that area better than I, as I've watched people work game consoles but somehow it's never tempted me. But it's fascinating the way jagabo analyzed that video's framing setup. I think if I saw it I'd believe I was getting too tired to see straight.

[zerowalker]

thought it always had component for progressive on DVDs aswell.

But yeah, it´s impressive, i just thought of it as interlacing, and use QTGMC;D

[jagabo]

The signal that travels across composite or s-video is always the same and always interlaced -- meaning the picture is transmitted one field at a time, 59.94 fields per second (50 fields per second for PAL), alternating top and bottom fields. There is never any variation from that. But the source video can be progressive (30p with each frame sent as two fields), hard telecined (24p with 3:2 pulldown), fully interlaced (60 different fields per second), or anything else (for example, cartoons may be created at 12 frames per second, each frame duplicated to 24p for film, then 3:2 pulldown applied for video -- or the original 12p could go through 6:4 or 5:5 pulldown to create 60 fields per second).

As to why a video cap can alternate between in phase and out of phase frames, it could be the source device doing it. The device is rendering progressive frames and putting out 60 fields per second from those frames. Most of the time it displays each frame for the duration of two fields. But sometimes it displays a frame for an odd number of fields (most likely the next frame took too long to render so the device output the last frame for an extra field). The odd number of fields causes the phase change:

frame #3 output for 3 fields: 1t 1b 2t 2b 3t 3b 3t 4b 4t 5b 5t
captured TFF: 1t1b 2t2b 3t3b 3t4b 4t5b 5t6b

The first three frames look progressive. The last three frames look interlaced. The next time a source frame is output for an odd number of fields the cap will go back to looking progressive.

This can also happen if the capture device drops or adds a field:

broadcast: 1t 1b 2t 2b 3t 3b 4t 4b 5t 5b 6t 6b
3t missed: 1t1b 2t2b (3t dropped) 3b4t 4b5t 5b6t 6b7t

The first two frames look progressive, after the dropped field they look interlaced. The next time a field is dropped the frames will go back to looking progressive.

[sanlyn]

Thanks for clarifying. That's what I meant to say, just didn't know how to say it correctly.

[zerowalker]

So, if it´s showing top field Even, Bottom Field odd, in 60 Fields per seconds, does that make it, 30fps = progressive?
Cause not sure i get all this, but it´s really interesting, so sorry for not grasping it all!

[jagabo]

So, if it´s showing top field Even, Bottom Field odd, in 60 Fields per seconds, does that make it, 30fps = progressive?

What makes it 30fps progressive is the fact that pairs of fields come from the same progressive source frame.

[manono]

So, if it´s showing top field Even, Bottom Field odd, in 60 Fields per seconds, does that make it, 30fps = progressive?

If the two fields of each field pair are the same (taken from the same point in time) then yes, it's progressive 29.97fps.

If the two fields of each field pair are different (taken from different points in time), then it's interlaced 29.97fps.

Where the video has been captured with the fields out of phase, as in your example, I'm assuming they've already been realigned.

[zerowalker]

Okay, but fields out of phase and interlacing, they are different right?
but looks pretty much the same?

[manono]

When they're out of phase they appear to be interlaced when looking at the frames. But a quick check of the fields shows you there are pairs of identical fields (except for the bobbing up and down). That indicates they're out of phase and that can be easily fixed with a simple TFM and they'll return to being progressive. It's better to field-match than it is to deinterlace, no matter how good the deinterlacer.