JVC HR-S9600 TBC settings modes?

[Olivier Talouarn]

Quote:

Originally Posted by sanlyn

You're a professional and you don't know how to capture a frame directly from a video? But what do the images tell us?

I will ask you to stay out of this conversation. You're advices are useless; you are insulting and sententious.
I was asked for pictures, i did it quickly. The purpose was not to check a valid pixel size or ratio, we know them, but a blending bug said "impossible".
What would you do with a still, even it it was correct ? Put it in a editing soft ? Silly.

"You might also consider that animation is usually not interlaced but is more often progressive with pulldown.
Oh thanks God ... i didn't knew ...
You are stuck in your world. Being in Pal countries, its a Pal tape, and it could be transcoded from NTSC to pal.
I was about to write it before you come, because i knew someone will go for it (you) : " dont speak about pull down", because even if it was; i would need to keep interlacing to reverse this pulldown. So useless quote again. Technical stuff are not as you want them to be, they are as they are. reality.

This is not a competition.

"If using lossy ProRes422, just be sure the compression is zero or near-zero

Theres is no quality settings in ProRes. It s adaptative. As Dnx or Cuneiform.
Only 4 quality ranges in 422, and a 444 and then brand new raw.
Build for HD, so it s suitable for SD. And easy to fit any common workflow, PC or Mac
I can t force end users to install exotics codecs.

[sanlyn]

Quote:

Originally Posted by Olivier Talouarn

TBC is a Snell&Wilcox TBS24. Input in composite at this time (but in YC as soon the JVC will fe fixed), output in YC. There is a RVB out, but with separate synchro (RGBS) ... Know it can be converted to true RVB, but i try have less possible chain. It have not the NR option. It also will transcode Secam to Pal. Yes, i m in France ... country of secam, scart connector and Concorde ... (also have many secam tapes to digit, need to buy another player) But this problem occurs before it. If it was, it would happen any time. Edit mode woulndt lock to interlaced. TBC also have a "dub-Beta" in, wich will be used soon behind a BVW75 pal). Many kind of tapes ... Some DV too; no DvcPro at this time.

I dont think any (French) Secam player was ever made with an internal TBC ... Can somebody confirm ?

other question. I have severals VHS-C. Spoke with a seller, he told me to avoid cheap converters, which can destroy JVC tapes. Off cours those i have are JVC ... Should i buy a true JVC converter. Not many to sell, but i think i can find some. Or go to a manual swap of tape to a classic enclosure ? (I'd prefer to avoid this ...)

-- merged --

Well; i searched the forum and found the answer ...From Lordsmurf : "VHS-C sucks. Just transplant them into full VHS clamshells."

This is going in circles. Time for a rant:

I guess I must have been overdoing things all these years. And I must have missed reading part of the post I quoted above (this can happen on a laptop, where the display is always so terrible and the keyboard so sensitive and jumpy that you never really know where you've traveled half the time). This discussion looks like a description of processing that I've always avoided, not to mention two images that illustrate why I've avoided them.

The first thing I've avoided is capture to lossy codecs. Yes, I know that ProRes is "near-lossless" and "virtually lossless" and "almost-lossless" and all that. But it is still lossy. A codec is either lossless or it isn't, and with lossy codecs you have no control over what is lost. Since you never see what gets lost, anything you say about how it's "virtually" lossless is nonsense because you wouldn't know what was lost to begin with. DV? DVCpro? It's already admitted that considerable image modification and multiple-stage post-processing are intended, so don't try to sell the virtues of lossy captures. I find that starting with damaged goods from the first step onward is an unappetizing and unprofessional proposition.

This forum has long recommended to use VHS-C adapters in Panasonic players and to avoid using them in JVC and other players. Complaints about VHS-C adapters in JVC machines are not new.

I'm surprised to see someone claim that yadif is too slow, especially since yadif is used by media players, big-name set tops, and TVs for real-time deinterlacing. You will be disappointed with QTGMC and any other motion-compensated or smart-bob process because they are all slower than yadif or simple bobbing. Also surprised to see a pro who doesn't know that VHS is YPbPr and not strictly YUV (I don't think anyone really knows what VHS color and video storage "is" anyway, and the convoluted explanations offered are starting to put me to sleep). Meanwhile the two images posted earlier don't look like video frames, they look like some player's or editor's rendering of a video frame. This can cause an entire universe of variations in the final product, which is why I wonder why a pro is using such a method to make images for analysis. What's going on here?

The tape from which the two images were taken has serious problems or has a tape mastering glitch. The interlaced picture that was posted produces two different images, as expected -- but the top field contains a ghost image of the bottom field, and the bottom field contains chroma remnants from the preceding field. This is not the way interlaced video behaves when it's deinterlaced. Rather, it's the way some forms of telecined video behaves when it's de-interlaced. So, either that tape is not pure interlace, or it uses some form of pulldown, or interlaced fields are out of phase, or there are playback and/or capture problems somewhere in your capture chain. I would suspect your external TBC, your capture card or its software. JVC's primitive dnr is also known to cause similar problems.

Original interlaced image:


Interlaced fields, separated and stacked vertically:


Top field de-interlaced, with ghost image of bottom field imbedded:


Bottom field de-interlaced, with chroma ghosting from previous tip field (look behind the ears):


It goes without saying (but I'll say it anyway, because what seems obvious to one reader might not be so obvious to other readers) that we can't advise on deinterlacing and other cleanup methods using nothing but a cropped still image from sources that aren't lossless or unaltered to begin with. Basically we have no video to help you with.

It's so much easier and exacting to just tell your editor to copy a frame directly to the clipboard with no cropping or other shenanigans required. Even rank beginners know how to do that with free stuff like VirtualDub and Windows Paint. It doesn't take a pro to cut a short length of video in a free app like VirtualDub or even a budget NLE, save it as lossless media with no colorspace conversions or further encoding, and post it in a forum. It's done all the time here and elsewhere.

I wonder, too, why it's so difficult to accept that VHS is interlaced. What in the world do you poor people do with interlaced HDTV broadcasts, retail DVD, and so much interlaced/telecined Bluray in the marketplace? Could I also ask why the "client" is using a defective playback system that can't handle interlace or telecine? What kind of low-tier playback system is that? There also seems to be an unusual information gap about the fact that PC editors and PC screens don't deinterlace the way TV's do. Again, why is TBC hardware and other devices being used that are not optimized or designed primarily for analog VHS, and why does it seem unreasonable to insist on using the correct tools for the task? VHS processing is such a time-consuming chore, where do people find time or patience for re-inventing the wheel?

End of rant.

Quote:

Originally Posted by Olivier Talouarn

Build for HD, so it s suitable for SD.

Decades of professional and advanced amateur experience strongly disagree with your opinion.

[Olivier Talouarn]

"DV? DVCpro? It's already admitted that considerable image modification and multiple-stage post-processing are intended, so don't try to sell the virtues of lossy captures. I find that starting with damaged goods from the first step onward is an unappetizing and unprofessional proposition.

When would you stop to say what i never said ? This text only say "I HAVE ALSO DV tapes". Are you dumb ?

"I wonder why a pro is using such a method to make images for analysis.
NEVER TOLD THIS Was not for "analysis". It was just to dismiss you "it can t happen".
And i chooses this image just as an exemple of what the player can do; "blending".
Let down other artifacts, they are not the subjects.
And keep technical VHS signal theories for yourself, or find me a Player with this components out...

Do you work in real world ? Yes better is better, but in real there are times schedules, hours cost,
and if possible i will go to best, if it s financially possible. If client is pleased with better deinterlace
and ready tp pay for it , he will get it. If it seems to costly, he will have less.
Pro work is not only standards and waveform curve, it s also time and money.

'why it's so difficult to accept that VHS is interlaced. Never said the contrary. Another pure delirium from you

Why is it difficult to accept a VHS with a TBC, DNR is much more complex than what you say ?

"It's so much easier and exacting to just tell your editor to copy a frame directly to the clipboard with no cropping or other shenanigans required. Even rank beginners know how to do that with free stuff like VirtualDub and Windows Paint. It doesn't take a pro to cut a short length of video in a free app like VirtualDub or even a budget NLE, save it as lossless media with no colorspace conversions or further encoding, and post it in a forum. It's done all the time here and elsewhere.

bla bla bra ... It s called 3 children.
Did frame capture very quickly to stop the "it's impossible dogma" .Yes it is. Here are the proof.
Leave us alone, don't come back. Go out in the real work world , you seem to have forgot what it is.

Decades of professional and advanced amateur experience strongly disagree with your opinion.

Oh so all Hollywood industry is misleading ... and all the stuff here about free kind of Jpeg codec is a scam.
Ok buy me a raid array.

"Could I also ask why the "client" is using a defective playback system

This IS THE original question of my post, with a JVC 9600 . But you made it a delirium.

[sanlyn]

As a said at the start, I sensed that your mind is made up, and now I see I was correct. I detect a morbid dread of posting real video as evidence -- or as I assume at this point, you don't really know how to do it. Two still images with no change or alteration in your setup to point to the actual cause of the problem is not a valid test. Your external TBC could be the problem, or your capture device. You didn't even seem to take notice of the ghosting problem seen in your images, so I assume you will accept those errors in your work as well and deliver them to your clients.

I guess you'll proceed as before and get the same results, in which case I don't have the slightest idea what else can be said or done.

[hodgey]

Quote:

Originally Posted by Olivier Talouarn

Yadiff still exist as an OFXplug, but its not GPU, so terribly slow.

Yadif is slow? On a modern machine it should be quite fast, maybe the plugin is badly written or something. QTGMC is muchslower, though gives some pretty amazing results.

[lordsmurf]

Please do not make multiple replies in a row, edit your previous reply to add more thoughts/questions.

ProRes422 most definitely does have compression level settings. Even the reverse-engineered FFMPEG version does as well. My Mac is currently fubar, so I can't pull up any screenshots. But the standard "ProRes422" is fine, HQ would be better. (I forget offhand if HQ allows SD; again, my Mac is off, can't double check.)

Yadif is so fast that it's actually realtime these days.
With QTGMC, don't use slow (which also blurs some), and look instead to the faster settings.

Bad PAL>NTSC or even NTSC>PAL must be eliminated as a variable before even bothering to post samples. After samples are given, especially when using said samples as a means to prove something, claiming PAL/NTSC is suspect. You don't reintroduce new variables midway through a process of elimination, it's bad science. It negates everything, and now you must start over.

BTW, on the sample image, am I the only one that notices that obvious tearing? So now my question is this: Did the Snell TBC cause it? Because it can! Or simply not correct it, meaning (as I already alluded to and stated outright), that TBC is not meant for consumer analog formats. Of course, basic frame level TBCs like Cypress ad DataVideo don't correct tearing, but the S&W makes insane promises.

I think VHS is YCrCb, but YUV is the accepted shorthand.

I also think the sample images look to be screeshots from a player, which isn't good, as players often affect video. Some more than others. VLC can be made transparent, but any MPlayer-based will screw with it.

Again, we need sample clips.

I sometimes cringe when I see sanlyn going into rant mode, but nothing he said above is really out of line. I agree with most of it.

I have never, and will never, accept "they won't pay enough" as a reason to do crap quality work. Professionals do quality work, they don't do whatever gets the quick buck. Working with a client budget is fine, working with so unreasonable a budget that you're forced to churn out garbage is not. That goes for not just video, but most of the creative fields.

I try to always be nice on this forum, and it is in fact the main forum rule, but sometimes even I get pushed with patience. Post a clip, or where all just yapping for nothing.

[Olivier Talouarn]

ProRes422 most definitely does have compression level settings.
NO. There is 4 differents "compression level " version
ProRes Proxy, LT, Standard,HQ. About the same "system" as Avid.. No quality slider.

Will go on with test will the Snell. Was the best choice at this time. Its a mid range product from Snell. S-video inputs a quite rare on their products. Yes would ove to try a DataVideo, but they are impossible to find. W

Yes i want to do quality work. i quite sure they would be pleased with fiels blended movies ... just because they re discover them. No way. But the firs goal is to make those video "visible". They do not ask for ultra pristine result. Too much work and cost, and it's a dead end. I m already far over the initial demand, as i said, if better is possible i must try it. Then pedagogy could come. They also need to learn difference between quick done, and better done, then choose. They decide. Won't have time for files till next week.

I speak about the OFX Yadiff, which code wasn't optimized since years.

[lordsmurf]

Yeah, just post files next week. We'll be here.

[sanlyn]

Quote:

Originally Posted by lordsmurf

I sometimes cringe when I see sanlyn going into rant mode

I cringe, too, when I come back 2 hours later.
Video work can be wildly frustrating when you feel you're not getting anywhere.

In easier circumstances I could have said the same things in a calmer fashion. When I try to convey complex video issues I often stress the importance of patience. In this case I should mind the ancient admonition, "Physician, heal thyself."

Apologies to all.

Quote:

Originally Posted by lordsmurf

I think VHS is YCrCb, but YUV is the accepted shorthand.

YPbPr=analog
YCbCr=digital

[dpalomaki]

I wonder what the history of the material is?
It looks like it could have been CEL artwork to FILM to <?????> to PAL VHS.
But is there something in the <?????>, such as FILM to NTSC to PAL?
The two samples look like adjacent frames, not the same frame, based on the background differences.

[Olivier Talouarn]

Yes , those advertisements are 24 (23,976) converted to 29,97+pulldown, then transcoded to PAL ...
And i know many archive films of WW2, some wicth i edited, which are in the same horrible condition.
Learning avi synth ...

Ok, so, apologies from me too.