S-VHS JVC VCR as comb filter passthrough for non S-VHS VCR?

[Krazern]

Hi, I have JVC HR-J668 and JVC HR-S7500 (PAL). I've been capturing my tapes using them paired with Panasonic DMR-ES10 or EH-65. I used hauppauge live 2 with ES-10 and Blackmagic Intensity Pro with EH-65 (HDMI route, seems more pleasent as it's luma range isn't limited and double A/D conversion is avoided). I usually like picture the most from J668 (the best physcial state) but as it's not an S-VHS VCR i get a little of dot crawl in my captures. It wasn't bothering me at all but when i put the S7500 between J688 and EH-65 it seems to me it has better comb filter than one in the Panasonic. What do you guys think? Can you see the difference? Is it worth to add S7500 to the capturing chain? Both VCRs were set to EDIT mode with BEST on.

Single frame comparsion (i tried to match histogram values here): https://imgsli.com/MjU4MzYz/0/1

[dellsam34]

Too many conversions, I would not go that route.

[lollo]

HDMI route, seems more pleasent as it's luma range isn't limited and double A/D conversion is avoided

The standard route (ES-10 -> USB Live2) is better than HDMI route (EH65 -> Intensity Pro) in this use case. In the second the details in the whites areas are lost.

frame 580 comparison: https://imgsli.com/MjU4Mzkz

However, although the whites are in the >235 area, there is no clippling in the Infinity histogram:



I assume the problem may come from the known "brigthness problem" in the Panasonic DVD-Recorders.

The same applies comparing J668-S7500-EH65-Intensity.avi

[Krazern]

The standard route (ES-10 -> USB Live2) is better than HDMI route (EH65 -> Intensity Pro) in this use case. In the second the details in the whites areas are lost.

However, although the whites are in the >235 area, there is no clippling in the Infinity histogram:

I assume the problem may come from the known "brigthness problem" in the Panasonic DVD-Recorders.

I don't adjust levels while capturing with intensity- it captures whole range instead limited one in hauppauge where it's needed to be done while capturing to avoid clipping. I capture the hdmi signal uncompressed 4:2:2 10bit with blackmagic media express and then compress it to lossless codec like FFV1. While compressing i also adjust values looking at histogram. This way it's more pleasant to me as i can already see if the values are on point looking at whole capture.

Here i just compressed intensity captures with FFV1 so the files are within the size limit.

I adjusted the levels in single frames comparsion.

[lollo]

USB-Live 2 captures 16-254, not limited range.

Whatever you do, that HDMI capture (Blackmagic only transfers the digital stream, 10-bits are useless) is worse than the classic flow. Changing the "levels" in the BM is useless, it does not "capture" (convert A to D) in this case. There are hints in the doom9 german board about HDMI captures suffering of "brigthness problem"

[Krazern]

The standard route (ES-10 -> USB Live2) is better than HDMI route (EH65 -> Intensity Pro) in this use case. In the second the details in the whites areas are lost.

The details aren't lost, i adjust the brightness and contrast while compressing the uncompressed hdmi stream- after doing that the ES10 capture looks basically the same if I look at potenially lost details in clipped parts. I upload sample adjusted in vdub below.

[Krazern]

There are hints in the doom9 german board about HDMI captures suffering of "brigthness problem"

I think they are more about Panasonic DMR recorders clipping whites with some VCRs instead the HDMI capture itself. I also have JVC HR-J658 and both ES10 or EH65 give me the same result- horribly clipped whites that cannot be corrected while capturing with hauppauge live 2 or after saving HDMI stream from EH65 while S7500 and J668 picture is fine with these.

[oln]

Yeah it's not the HDMI output that's clipping, it happens on the input side. (There is a problem with brightness pumping on the standalone s-video output afaik on some models but that's a separate issue). As you can see in the histograms the same parts are captured, the max white point just ends up at different spots and if it was clipping you would likely just see it clipping at 225-235 or thereabouts with the live-2 instead with those settings. What you are getting from the hdmi output is the same as is converted to analog on the analog outputs.

There are some parameters of the image that has some impact on the gain control/brightness of the image on these dvd-recorders besides just the signal level, possibly sharpness in some way but I don't know what exactly - as it seems to vary a bit from vcr to vcr how much it happens and I didn't have issues the few time I used it with some Hi8 tapes where I had to use a vcr without tbc rather than camcorder.

I guess it's also worth noting that vcr tbcs (or more specifically adc/dacs) can also cause clipping of whites/blacks, newer PAL JVC SVHS decks seem to do it apparently in some cases according to another user here:
https://www.digitalfaq.com/forum/video-capture/10658-luminance-clipping-capture.html

[lollo]

The details aren't lost

Don't you see the missing aerial high voltage cables in HDMI captures? If there were other light areas in the shots there will be more (the white shirt has no fine details)

As you can see in the histograms the same parts are captured, the max white point just ends up at different spots

The histogram of the HDMI capture do not say much, because it is not that of the original capture, but of the post-processing after adjustement, according to OP's later description. That's why I was not able to conclude in my first reply. I suspect that if we look to the histogram of the original capture we find the spikes and the clipping.

I guess it's also worth noting that vcr tbcs (or more specifically adc/dacs) can also cause clipping of whites/blacks

I think that the VCR is out of the equation, because is the same in Live-2 capture (no problems) and HDMI capture (problems).

[Krazern]

Don't you see the missing aerial high voltage cables in HDMI captures? If there were other light areas in the shots there will be more (the white shirt has no fine details)

I definitely can see them missing without brightness and contrast adjustment, but when i change their values while compressing uncompressed video or simply in post processing i can bring them all back. I upload vdub screenshot with visible histrogram and adjusted brightness and contrast made on uncompressed cut.

The histogram of the HDMI capture do not say much, because it is not that of the original capture, but of the post-processing after adjustement, according to OP's later description. That's why I was not able to conclude in my first reply. I suspect that if we look to the histogram of the original capture we find the spikes and the clipping.

The HDMI caputers from first post are without adjustments, i just compressed them with FFV1 without touching brightness or contrast. In 6th reply i uploaded the same clip compressed with FFV1 but with brightness and contrast changed.

I'm not sure if original uncompressed cut would change anythining but here it goes: https://we.tl/t-2oCYXiMQ9Y

I think that the VCR is out of the equation, because is the same in Live-2 capture (no problems) and HDMI capture (problems).

If i wouldnt adjust values while captuiring with hauppauge it would look identical to EH65 HDMI capture.

[lollo]

Don't you see the missing aerial high voltage cables in HDMI captures? If there were other light areas in the shots there will be more (the white shirt has no fine details)

I definitely can see them missing without brightness and contrast adjustment, but when i change their values while compressing uncompressed video or simply in post processing i can bring them all back.

Correct, but partially. And by doing so you are affecting the whole picture, which becomes too dark (uness you use a specific mask based on luma and/or area). Overall, the HDMI picture is less convincing. And it features one missing/drop frame if I am not wrong.

[Krazern]

Overall, the HDMI picture is less convincing. And it features one missing/drop frame if I am not wrong.

Yes, I feel like Intensity HDMI captures are a little more prone to drop frames than hauppauge but it happens rarely and in the end i dont think i have to care about audio and video sync when just capturing HDMI signal (?).

I'd like to return here to the main thought behind my thread- i don't seem to see dot crawl when i use S7500 as passthrough between j668 and eh65. The S7500 alone doesnt handle this tape well so i need to use j668.

[Krazern]

Too many conversions, I would not go that route.

I also thought that before. But in the end isn't it almost the same amount of A/D conversions compared to standard S-Video capture?

Like:
1. J668 scart composite - front composite input on S7500 and s-video out - EH65 A/D conversion - direct digitized signal HDMI capture.
2. J668 scart composite - A/D ES10 conversion on input and D/A conversion on S-Video output - hauppauge A/D conversion on input.

I feel like the S7500 passthrough brings benefits in less crawl dot and maybe better colors(?).

[lollo]

Yes, I feel like Intensity HDMI captures are a little more prone to drop frames than hauppauge but it happens rarely and in the end i dont think i have to care about audio and video sync when just capturing HDMI signal (?)

The a/v asynch can be baked in the HDMI stream while the Panasonic was capturing, so while there no trobules in capturing the already digitized stream, there can be in what creates it. But you can check versus the "classic" workflow and find potential problems (if youjust experiment one or two missing/added frames in a couple 1 hour capturing is not an issue anyhow

I'd like to return here to the main thought behind my thread- i don't seem to see dot crawl when i use S7500 as passthrough between j668 and eh65. The S7500 alone doesnt handle this tape well so i need to use j668.

Yes, we were going a little bit off topic, sorry.

Concerning your main point, I do not see a major difference between J668-EH65-Intensity.avi and J668-S7500-EH65-Intensity.avi, except some more noise on the second, but I did not deeply checked. You are the only one, having a larger test bench opportunities with your hardware and tapes to take a decision.

[oln]

Too many conversions, I would not go that route.

I also thought that before. But in the end isn't it almost the same amount of A/D conversions compared to standard S-Video capture?
.

The Y/C separation in the 7500 is digital so using it will involve an extra A/D conversion compared to connecting the j668 directly to either dvd-recorder it if that's what you mean. That might be fine though using it for Y/C separation in this case works better than using the composite in on the dvd-recorder. You can also experiment with the "comb filter" setting in the menus on the Panasonic dvd recorders, sometimes it may work better with it on or off on videotape sources.

[rgr]

I think that the VCR is out of the equation, because is the same in Live-2 capture (no problems) and HDMI capture (problems).

Panasonic DMR recorded the signal on HDD (MPEG2) from the Panasonic HD640 VCR in the range of 16-235, while from the Toshiba VCR in the range of 16-255 and probably the same with the JVC (all composite).
Of course he marked it as limited.
I'm not writing anything about the HDMI output because I haven't tested it comparatively yet, and I have no idea about the HDMI signal, but I think it's similar.

[lollo]

The VCR is the same in both workflow. It goes to ES10 in one case, and to EH65 in the other case. IMO, the VCR is not responsible of any deviationin the final outcomes.

[rgr]

So why did I have different results on the DMR-EH575/585 after replacing the VCR?

[lollo]

I have no idea , and not aware of the contest...

[rgr]

I have no idea , and not aware of the contest...

But it's interesting nonetheless

[lollo]

So what exactly did you experiment?

[rgr]

I recently captured an NTSC tape using a Sony SLV VCR and a Ferguson DVD recorder (LSI Logic). To my surprise, the video saved in MPEG2 had a range of 0-254.

[rgr]

Don't you see the missing aerial high voltage cables in HDMI captures? If there were other light areas in the shots there will be more (the white shirt has no fine details)

I definitely can see them missing without brightness and contrast adjustment, but when i change their values while compressing uncompressed video or simply in post processing i can bring them all back. I upload vdub screenshot with visible histrogram and adjusted brightness and contrast made on uncompressed cut.

I looked for this thread, but didn't find it -- so I started a new one on this topic.
https://forum.videohelp.com/threads/414437-Panasonic-DMR-HDMI-capture