I can supply you with few seconds of lossless cap of that clip using the S-Video output with my USB capture device instead of the firewire port and compare them if you like, The intention from that post is to just show one step capture of a D-VHS machine, it wasn't meant to be the best capture method ever.

A most interesting and generous offer! It's rarely the case (if ever) that a lossy -vs- lossless version of the same source is presented, even if one version is incomplete. I think everyone would be interested in a comparison.

I worked on a cleanup of the "Wait" mpg using three different filters. It illustrates how something can be repaired (i.e, improved) but that data loss from lossy originals limits the results and makes the work tougher. It's been an interesting exercise, even if takes only 10 to 12 minutes each to filter 4 different versions. I've been sitting around trying to figure out which looks better and typing up some notes on how data loss interferes with the work and makes it more difficult.

By the way, mastering on the original isn't all that great. There are digital missteps even before that performance became a tape (field=-blended segments, duped frames, telecine on top of interlace, segments with field priority changes, bad chroma ghosting between scenes.....being lossy isn\'t even the beginning.

Below, the not-so-solid black background of the title shot, slightly brightened from the range RGB 16 to 62, showing low-level macroblock noise and bitrate completely wasted on it:
http://www.digitalfaq.com/forum/atta...1&d=1546668864

Chroma damage and field blending:
http://www.digitalfaq.com/forum/atta...1&d=1546669059

Other hortizontal dropouts, and sawtooth edges on gradients:
http://www.digitalfaq.com/forum/atta...1&d=1546669208

Film grain distortion and blocky artifacts:
http://www.digitalfaq.com/forum/atta...1&d=1546669292

Chroma ghosting from previous scenes and frames (red ghost in center of image):
http://www.digitalfaq.com/forum/atta...1&d=1546669405

A lot of that stuff such as chroma ghosting can't be fixed without creating even more damage, and original grain and tape noise imbedded as artifacts is a nightmare. But I managed to clean up some of it and made a short demo video. The attached mp4 compares 5 different quickie, short scenes, with the original frames on the left and the filtered frames on the right of each video frame. The scenes are inverse telecined back to the original 23.976fps film speed, but the demo video is slowed to 10 fps. The first scene is the title shot. Each scene in the comparison demo plays twice. Nothing perfect or exotic, just some quick fixes. http://www.digitalfaq.com/forum/atta...3&d=1546670103

Now all I have to do is decide which filtered complete version I like best, or if I should make a complete video by selecting segments from each. If I can make up my mind I'll encode it and post offsite, as it will be a pretty large file encoded for standard definition BluRay.

Attached is the first minute if the 4-minute "Wait For Love" mpg, cleaned up to at least a more palatable state. Denoising revealed a lot of brute-force and sloppy digital mastering of the film source, although the mpg looks better than a typical YouTube disaster. Perfect cleanup isn't possible or desirable, since repairing 100% of the lossy damage would remove 80% of the video, and who would want that? The encoding artifacts were such that many popular Avisynth filters had ugly side effects or didn't work at all (for some popular chroma cleaners did more harm than good).

The main "workers" in Avisynth were TIVTC (inverse telecine), Biofrost, FluxSmoothST, TemporalSoften, SmoothUV, Dither Tools, a touch of RemoveDirtMC, and some homegrown routines to replace bad dropout pixels with clean pixels borrowed from other parts of the frames. The two VirtualDub workhorses were ColorCamcorderDenoise and NeatVideo, both at 1/3 power. A few visible remnants of some of that frequency-connected vertically rolling debris mentioned earlier are still present but, again, lossy encoding made it more difficult to remove entirely and would have resulted in a denuded video. Much of the mpg's image content consisted of compression artifacts to begin with.

The attached mp4 is suitable for web mounting at 23.9786 fps. The 720x480 frames were populated only in a 704-x480 area, so extra side pixels were removed, the bottom-border head-switching noise was removed and replaced with new borders that centered the image vertically, then 16-bit dithering and Spline36Resize was used to reformat the video to square-pixel 640x480. A standard-definition BluRay version would playback the same way, but 3:2 pulldown would be added for 29.97 playback, Dolby AC3 audio would replace the obsolete and inferior MP2, and the 720x480 format would be retained for BluRay encoding with a 4:3 display aspect ratio.

So the mp4 looks cleaner but it would look much better and would have been a lot easier if it had not been a lossy capture.

There have been times where I transferred something to DVD or MPEG, then later decided to restore it. I'd often have to just re-capture it lossless, trying to restore the MPEG was futility.

I wonder if it is hardwired or can be hacked in firmware.

Nevertheless I now own the JVC HM-DR10000EU. Does anyone know if DVHSTool works with PAL video and how to prepare video to have suitable tx file to record onto tape? Or should I record enything on the VCR, transfer it to PC and make Vegas preset to match the parameters?

EDIT: What am I saying... I cannot record it back, because firewire is input only...

There is no such hack for them, They disappeared quickly and hackers focused on defeating DVD copy protection, They know it's just a waste of resources going after D-VHS.
DVHSTool should work with frame rates for both systems 50Hz and 60Hz.
Recording to a D-VHS deck from any file has a learning curve, You have be able to make a compliant file otherwise the DVHSTool won't record anything or record broken stream.
I went thru that hell recently and actually I was lucky to get help from the guy who worked for JVC on the firewire board for these machines and he made a very efficient commend to use with ffmpeg (he goes by drmpeg @ videohelp). It was a wonderful learning experience, Take a look.

Well, I asked, let's see if it gets answered.

Mind if I ask about a topic within this thread, but redirect it towards my interest?
If there was ever a more descriptive discussion, just send me there, and not burry me with the digital artifact smears here :( Been registered for a long time, sorry to see it's diving into facebook group type discussions.

How you doing Lord Smurf? :) There's a familiar name!

Anyway..... VCR or D-VCR and RGB video at ~15.5hz.
Last I'd searched the forums, the VR-813 was about all I'd seen at the time with RGB output.
DVH-8090,Thomson seems to be a more capable unit overall, abit I don't care about it's digital capabilities, only its ability to output an RGB signal from analog tape.
I can break out from the SCART connector, only need a red, green, blue and ground line. PAL-NTSC doesn't matter.

Am I missing other units? Going rate for one I'm market place is well over €500..
The I/O is an legacy capture card that normally uses an composite of y/c source, and converts to an RBG signal for input to the capture card. There is no audio capture, no TBC or tick marks to align the video post capture with an audio track. Wouldn't matter, max capture frame rate "nears" PAL @ 320*240, and possible full frame rate in lower capture resolutions. Of course quality is important, but not with respect to discussions in this thread. The video is always handled uncompressed, captured to RAM, enough room for about 15k frame count. Like I said above, PAL-NTSC doesn't matter output wise, the machine handles both, but must manage playing of NTSC recorded content. Some specs for the DVH-8090 Thomson show only Euro spec, found a museum site stating NTSC was also available.
Thanks, and hats off to ya'lls knowledge.

Component (RGB) is well capable of reproducing the full spectrum of Y/C (S-Video) so why limit your resolution to 320x240, Sure VHS is a low bandwidth signal but once you expand it using S-Video or RGB why bottle neck it again with such low resolution, If you are capturing VHS from composite (RCA) that should be about right.

Like I said, if you read carefully, it's a capture hardware limitation.
If you really must know, the video capture is done on a 1992 Atari Falcon 030 with 16mhz bus and CPU clock. Why, because I can.
I imagine RGB is used on the capture card as it is also used onscreen within the capture window of the application with no additional processing to do so.
The video content captured will be used in animation and 3d raytraced rendering.
They are not wedding videos.
I'm interested only knowing of tape (VHS) machines capable of RGB output.
PAL output is fine, but must read NTSC material. Quasi SVHS or SVHS is desirable but not mandatory.

Apologies, see under the topic I posted in an incorrect area.

Anyway, as already stated, the VHS/SVHS playback quality of those Thomsons is frankly quite poor. That model is only good for D-VHS... and even then, with caveats: there was a production run with a faulty encoder that tended to 'patternize' the image when recording. (I had a couple of those troubled units; now I have another, less offending, couple.)

Details on this Atari Falcon capture setup? 68k machines (Mac, Amiga, etc) really weren't all that fast for video capture. Most of them compressed YUV video using hardware (Video Toaster Flyer, Avid solutions, etc) and stored data on fast SCSI RAID setups. More consumer oriented cards were very limited in resolution and frame rate (the Video Spigot comes to mind here) due to CPU and IO limitations of the day.

I'll also add that its likely processing YUV internally, and only converting to RGB for the preview window. I own several gen lock cards from the period on multiple platforms. Yes, they output RGB, but its only used for preview as the end product is supposed to be composite or S-video going to tape. Input is strictly via composite video.