Workflow for VHS capture?

[Davis]

Just looking for some feedback. I've always captured VHS and other SD video as a fully uncompressed file. I would then edit and clean it up and adjust audio in my editing software. Then I render the timeline to uncompressed AVI and render that to MPEG2 for DVD using TMPGEnc Mastering Works 7 on another PC. That has always worked great and I don't think you could achieve any higher quality. My capture computer system may be failing and I will have to go a different route. I have new drives coming for the RAID so I'm hoping that takes care of the problem. If not, I'm looking at some other equipment but the best capture it has is MPEG2 I-frame only at 50megabit rate. I can do everything else the same as before. I really don't think there would be any difference in the quality capturing the video in that format at that high of data rate, especially SD. Any opinions welcome.

Thanks,
Marty

[traal]

You can use the HuffYUV codec in your workflow to keep bit rates down to around 60-70 Mbps and it's completely lossless unlike MPEG2 I-frame or MJPEG.

[Davis]

That would be great but that one is unavailable in the equipment I was looking at. I was actually looking at an old Tricaster Broadcast unit. For VHS and other SD formats I don't think you would see any difference capturing uncompressed versus MPEG2 I-frame only at 50mbps data rate. And hopefully the new drives I have coming will fix the problem I have with my old system. It freezes sometimes and then I lose the video I was capturing. I have replaced the memory so it's not that causing it. It could very well be a drive though, they have had a lot of use and sometimes testing them they would fail and then again not. So I will find out for sure when I get the new ones Friday.

[Eric-Jan]

I guess you have no problems having a stable video signal to capture from ? what's your setup ?
resolution wise MPEG2 is no problem, it only makes large files in the end….
I used to use ProRes422LT to capture with an Intensity Shuttle and an old MacBookPro, (thunderbolt2)
which for me worked fine, but now i use a BMD converter analog to SDI with a BMD recorder, which does the same,
and post can be done on any computer from the (flash) SD cards.
This way all video levels are under control the right way, using s-video (or HDMI/component video when available)
BMD will be more affordable then Tricaster, and will have good quality also.

[Davis]

Yes, I should have a stable signal to capture from. I'm using an AJA FS100 TBC which is full frame. Going into it with the S-Video from tape machines and capture the component out signal. It has worked great ever since I have been using it. It will even output SDI, but the current VT computer system I'm using doesn't have SDI input. That's why I capture using the component out of the TBC. I then clean up the video and adjust the audio levels and render out the timeline to Uncompressed AVI 4:2:2. That is what I then encode to MPEG2 for the DVD and author. I use TMPGenc Mastering Work 7 and Authoring Works 6 for that. Everything has worked great for quite awhile doing it like this until the capture computer starting freezing sometimes.

[Eric-Jan]

For how much can you get that Tricaster ? otherwise i would say check - out the BlackMagic Design hardware,
Ive got the Video Assist monitor/recorder from BMD which has SDI and HDMI in and outputs, The model i have is pre 3G, and not for sale anymore, so i guess a lot of people have bought or upgraded to the 3G model, and there should be 2nd hand (pre 3G models) available, 720x576i or 720x480i recordings are made from 526i or 525i the Video Assist will also be a converter to HDMI at the same time.
I am using the mini converter Analog to SDI straight on the RCA YUV component video output from my DMR-ES35V combo, without any frame drops previously i captured with a Intensity Shuttle, which worked also fine,
Using a JVC vcr with build-in TBC did not work for me that way.

[Davis]

Thanks for the reply. The BMD Video Assist does sound like it would do the job. The five-inch model is about $525 I think. The Apple ProRes quality for SD video should be fine. I can get a Tricaster Studio for around $200. It and the Broadcast version have been real workhorses as the saying goes. I've been using Newtek equipment for 20 years or more. I really wish it recorded in a higher quality than MPEG2 I-frame only at 50Mbps, but for VHS and other SD video, you can probably see no difference at all between that and capturing uncompressed. I would still render the timeline to uncompressed before encoding it to standard MPEG2 for DVD. So essentially, I should lose no quality at that point.

Thanks,
Marty

[Eric-Jan]

The Video Assist you found is the 3G model, that one is still in production, just like the 6G, and 12G versions,
The Video Assist will do all the ProRes422 grade codecs, in any resolution up to FullHD

[lordsmurf]

Be aware that, for SD video, especially consumer formats, especially VHS, Blackmagic gear can, and often does, drop frames, even with TBCs in use. It is very unstable analog input.

[Eric-Jan]

Quote:

Originally Posted by lordsmurf

Be aware that, for SD video, especially consumer formats, especially VHS, Blackmagic gear can, and often does, drop frames, even with TBCs in use. It is very unstable analog input.

No problems for VHS > PAL 625i without TBC, in my case…. using only BMD hardware, no visible or audible inperfections other than being it VHS… using ProRes422 which is not a consumer format. VHS output is very unstable, BMD video input is rocksteady

[lordsmurf]

Quote:

Originally Posted by Eric-Jan

No problems for VHS > PAL 625i without TBC, in my case….

There's always somebody that claims "worked for me!", but sometimes there are vastly more than claim otherwise.

In this specific instance, even the manufacturer knows it's not accurate, and will tell you outright that a TBC is required for consumer analog source ingest. Those BM cards simply fail without it, all sorts of issues trying to stubbornly/cheaply bareback/rawdog the video signal.

There are items that literally catch fire, burn down houses, kill people, recalled by the manufacturer, company sued into oblivion. At yet, even in those cases, "worked for me!"

[Eric-Jan]

Quote:

Originally Posted by lordsmurf

There's always somebody that claims "worked for me!", but sometimes there are vastly more than claim otherwise.

In this specific instance, even the manufacturer knows it's not accurate, and will tell you outright that a TBC is required for consumer analog source ingest. Those BM cards simply fail without it, all sorts of issues trying to stubbornly/cheaply bareback/rawdog the video signal.

There are items that literally catch fire, burn down houses, kill people, recalled by the manufacturer, company sued into oblivion. At yet, even in those cases, "worked for me!"

So now i have to call my local fire department to check on my BMD gear, for fire hazard ?
That's a great example you give for video capture, no car mechanics, or cooks, this time.
I try to stay serious when i'm talking about video in general, everybody has his own level of budget, or effort to put in it.

[lordsmurf]

Quote:

Originally Posted by Eric-Jan

So now i have to call my local fire department to check on my BMD gear, for fire hazard ?

Maybe?

Too many people think video is a TV, or Netflix, or a DVD, or whatever. That's the the viewing devices, the delivery formats. So they have near-zero understanding of concepts. You have to break it down into relatable chunks. So screwdrivers, hamburgers, flammable junk ... whatever works.

Go back to posts I made 15 years ago. I really did want to like Blackmagic when they came on the scene. But it's just not.

[lollo2]

Concerning the BlackMagic cards, on my side I also experienced the well known "black frames" when restoring a capture for a spanish friend.

If you look to the video on the right you'll notice them. I asked him to re-capture with its Pinnacle 710 USB within the same workflow, but had no news since.

compareX_DBZ_test_org_rest.mp4

[lordsmurf]

Quote:

Originally Posted by lollo2

Attachment 16455

As an off-topic note, classic Japanese anime isn't as smooth as modern Flash-type animation. That clip is a bit too smooth. You have to resist over-filter tendencies. It becomes a real struggle when you have more damaged footage, and filtering to restore also destroys the native grain too much. As an 80s toon collector, I run into this especially on Robotech (Macross, etc) rarities native only to VHS, most native only to Japan.

[lollo2]

Quote:

Originally Posted by lordsmurf

As an off-topic note, classic Japanese anime isn't as smooth as modern Flash-type animation. That clip is a bit too smooth. You have to resist over-filter tendencies. It becomes a real struggle when you have more damaged footage, and filtering to restore also destroys the native grain too much. As an 80s toon collector, I run into this especially on Robotech (Macross, etc) rarities native only to VHS, most native only to Japan.

Thanks for the comment.

I am not expert on anime, but the DVD releases I checked for DBZ were quite clean.
I think that what has been removed with the restoration is the "dancing" grain introduced by the VHS, not the details. The filtering used does a pure temporale denoise, nothing spatial. The "smooth" look is because there were little details to start with.
The rest of the processing was to fix shifted fields, black frames and remove scratches.

On the other hand, if the original "look" was a bit grainy compared to the DVD release I used as reference, featuring "natural noise" different than what has been introduced by the media, it can be easily added with specific filter to introduce a "80s toon look".

[hodgey]

Quote:

Originally Posted by Eric-Jan

No problems for VHS > PAL 625i without TBC, in my case…. using only BMD hardware, no visible or audible inperfections other than being it VHS… using ProRes422 which is not a consumer format. VHS output is very unstable, BMD video input is rocksteady

If you are talking about using ES35V out that digitizes the video and outputs a stable signal so the blackmagic box don't have to deal with it, it's not the same as connecting to the average VCR which tends to cause black frames galore.

I have yet to see any evidence of the blackmagic devices dropping frames on a stable input though.

[lordsmurf]

Quote:

Originally Posted by lollo2

I am not expert on anime, but the DVD releases I checked for DBZ were quite clean.

Slightly off-topic again, and yet good info to know...

I'm not a DBZ fan, but it's a common complaint among many anime prior to the late 90s.

Read this: https://www.animenewsnetwork.com/ans...-04-17/.145777

Quote:

The video nerds among us have been abuzz about the new Dragon Ball Z boxed set, and not in a good way. While fans have been holding out hope for a Blu-ray release in the show's native 4x3 aspect ratio (the previous release was cropped to widescreen), the previews released so far gave the impression of a smudgy, overly-cleaned image with colors that seem a little off. And as fans are wont to do, they're making their displeasure known. Loudly.

Something like Big O was clean, while something like the original Slayers, DBZ, and Tylor was not.

Long lived shows like Pokemon straddled this era, with earlier iterations/seasons being a wee more grainy detailed, and latter being slick and Flash-like.

Cel animation was drawn, yes, but it was later slow filmed.

Also, the animation was clean(ish), but the backgrounds were painted. You can easily lose details on the non-active animation elements by over filtering.

Ironically, in practice, VHS tended to better retain these details than digital, especially compressed DVDs, even with a lower bandwidth of resolution, color-under, etc.

Low budget DVDs (or anything digital, be it Blu-ray or streaming) is/was especially compressed, and most animation is/was disrespected as "kid stuff" that gets/got the ultra-budget treatment. And yes, sadly, that included not-cheap Funimation DVDs, which were still low budget productions. Budget for them, not us!

MPEG for DVD-Video was relatively compressed, and you have to pre-clean to maintain a good bitrate allocation. That's actually true of all compression. That's why, for example, we want to mask (NOT CROP!) busy overscan crap, to reduce the bitrate needs of the image edges to near-zero. Another way to reduce bitrate needs is to compress then color and detail information. For DVDs, that allows no blocks, but at a cost of obliterating fine detail, when compresed to heavily. H.264 went the mush route to avoid blocks, but the loss is still very present.

Compression and filtering is very much an art. Some treat it like fine art, some like finger paintings.

Anyway, for on-topic...

A point to be made here is that VHS can hold lots of details and quality. It's your job to properly extract it. No make excuses about "VHS quality" (aka lazy crap conversion), but to extract the quality that truly is on the tape. And no, we don't need vhs-decode for that ... we just need quality extraction hardware and methods.

[lollo2]

Quote:

Read this https://www.animenewsnetwork.com/ans...-04-17/.145777

Thanks, it confirm my suspect that the DVD/BR reference has been cleaned too much.

Quote:

Ironically, in practice, VHS tended to better retain these details than digital, especially compressed DVDs, even with a lower bandwidth of resolution, color-under, etc.

Yes, correct. The filtering prior to a DVD release sometimes is too invasive: edge enhacement to give a false perception of sharp picture and aggressive
noise reduction to help MPEG2 encoder.
I remember when we provided rare material in the hand of us collectors for the italian DVD release of a sci-fi series. The final result on the market was
a bit of a disappointment for the above reasons.

Quote:

A point to be made here is that VHS can hold lots of details and quality. It's your job to properly extract it. No make excuses about "VHS quality" (aka lazy crap conversion), but to extract the quality that truly is on the tape

Absolutely correct, an excellent summary of what we are here for...

About the capture I worked on, I think it was not that good, and once more, the removed noise during restoration was the dancing noise introduced,
not the "natural" look of the video master.

Quote:

And no, we don't need vhs-decode for that ... we just need quality extraction hardware and methods.

I am positive about vhs-decode, as soon as the idea "use whatever VCR as player" will be revised. In the meantime, we can have equal/superior results
with our classic workflows.

[lordsmurf]

Quote:

Originally Posted by lollo2

I am positive about vhs-decode, as soon as the idea "use whatever VCR as player" will be revised. In the meantime, we can have equal/superior results with our classic workflows.

vhs-decode actually has multiple issues, but most can likely be resolved with an intelligent ingest appliance that hardware processes.

The real problem is psychological, certain folks involved in the project. Egos, miserly attitudes, too libertarian. (Libertarian = insisting on autonomy, not wanting any help from "the other".)

For example, the "software only" notion is backwards tech thinking in the 2020s. This isn't a phone app, but a complex digital tool.

And rather than build on what exists now (ie, workflows we all use now), most insist on scrapping all of it, both the good and the bad. This "blank slate" approach is not working, and likely never will.

Add in the cheapskate factor of using random hardware, and the whole project is an inferior waste of time.

It's disappointing, pathetic even, but it is what it is.

Some devs are reasonable about having quality discussion, but more are not. You can almost read the exasperation in their words at times.