JVC BR-S522,S822 decks for 10-bit capture?

[Cingular]

Hello,

These decks and 10 bit capture go against conventional DigitalFAQ wisdom, but I would like to report my findings and see if anyone has any updated information.

I have two of these BR-S522DXU decks and one BR-S822DXU both came from the original owner, acquired in 2011 but they have sat up with only yearly tape use since. Recently I have been testing these decks, and one deck has an audio issue, another has an occasional "41" error, but the common "issue," is a fuzzy/wavy picture when you first turn on the unit but after warming up, for fifteen to thirty minutes the picture stabilizes. If stored for longer before turning on it can take up to four hours for the picture to stabilize. Once the picture stabilizes, these decks, so far, look better than AG-1980, HR-S9600U, Sony SVO-5800/SVP-5600, and am waiting to get the SR-W7U back from repair to see how it compares.

Another interesting point is that when capturing with these decks it removes that wavy line at the bottom of the image. It also looks like you can see the grain in the film, and the picture looks brighter, colors seem better. I need to do some more comparison testing but so far, aside from the 50lb size, and 7-pin S-video connector, I cannot understand why these are not recommended for consumer video?

There is a tracking needle for video that can improve/match the video signal and balanced audio and component output. Component output does appear to soften the colors slightly, I am still testing, but so far I think component output may be better. The BR-S522DXU's are playback units only, so I fail to see that they were made to play back a certain tape, because these decks came from a Wedding photography studio and low light is required, thus creating a need for specialized equipment.

Please remind me again, why these decks are not recommended? The forum does not recommended them because they are usually all worn out or the cards are missing from the decks and that they are optimized for a non-consumer type of video tape. This does not match my findings, except for the worn out part.

I read in the forum that VHS was native S-video, so does that mean the component output changed the color in a negative way? Actually the color is generally bad enough with VHS; I fail to see how component output adds anything negative?

I will save the 10-bit capture, discussion for another thread; I do not want to get this off topic.

[msgohan]

I have a BR-S525U. My complaint with it is that when EDIT mode is used, it seems to oversharpen, adding haloes to edges. The normal picture mode looks even worse, as the processing looks artificial to me.

Orsetto on VideoHelp talks about how the head gap is different than consumer VCRs, causing extra noise when regular tapes are played back.

VHS color storage is closer to S-Video than component. When you capture, you're always creating digital component video, so the combined chroma signal of Y/C needs to be fully separated into two color difference signals somewhere.

One theory would say that S-Video is closer to the tape signal, therefore component output from the machine offers no advantage and could actually be worse, unless your capture card handles S-Video poorly.

The opposing theory would say that because the deck internally separates into component to perform TBC (and possibly its other processing steps) that means component capture is superior: instead of deck separation -> deck recombination -> capture card separation you just have one separation step.

In practice, it doesn't make a difference. You could probably separate and re-combine and re-separate the chroma dozens of times without any visible difference, because the signal is already so degraded. Chroma is the worst part of VHS, by far.

How the capture card handles the two signal types will make a bigger difference.

[Cingular]

I have not used "edit' mode yet. I have not noticed any extra noise on a side by side comparison with uncompressed 10 bit MOV though VLC player, on two Calibrated Sony FW900 monitors and a Sony PVM-14L2. The 522 deck I originally tested had an issue with TBC-2 module which caused an orange-yellow shift in the colors when the TBC is engaged. I have yet had time to capture with the other 522 deck.

The component video is directly tied to the TBC; if I remove TBC-2 of the three TBC cards, even with the TBC off there is no component video. With damage to TBC-2 and the TBC off there is a color shift with component video but not with S-Video. The 522 with working component video, has slightly softer colors so if you saw the color "red" it was not just basic red but a shade of red and the same with the skin-tones, this is on a side by side comparisons on a Sony PV-14LM switching between outputs on a freeze frame. It sounds like the BR-S522DXU internally separates the signal into component to perform the TBC?

The capture card is a Kona LHi, it captures component but it will capture S-Video though “Red, blue,” and composite though “green. “ I will sometimes use a Leitch DPS-475 with audio between the deck and card. I read on the forums that a lot of people use two time base correctors, the one on the deck and an external TBC or frame synch because they do different things? The TBC on the 522 deck seems more than sufficient, I think it might be better to just leave the Leitch out of the mix and capture TBC straight to the Kona card?

The issue I am having with my capture is that several decks, HR-S9600U, and the Sony SVP-5600 and the AG-1980 were capturing with this reddish tint on the skin colors, but somehow the component output on the JVC BR deck was immune to this color issue. When connected directly to a monitor I did not notice any shifting of the color. I think I need to retest and start over. Lately viewing movies on a nothing special large panel LCD the JVC BR looks the best, and the image on the PVM-14LM looks great too. Maybe I am just overly optimistic? Further testing is required.

[msgohan]

I would be interested in seeing some short samples from the properly-functioning machines. The forum supports attachments up to 99MB. RAR compression will crunch down uncompressed pretty well.

My post about the 525 is on VH here.

Quote:

Originally Posted by Cingular

It sounds like the BR-S522DXU internally separates the signal into component to perform the TBC?

I believe so. If you want to try figuring it out from the block diagrams in the service manual, it's available here.

Quote:

I think it might be better to just leave the Leitch out of the mix and capture TBC straight to the Kona card?

If you're not getting dropped frames, sure.

[Cingular]

The first repair tech told me that the wavy lines and slow turn on were capacitor issues, and today another VCR tech confirmed the same thing. So all three of the 22 series JVC's probably need new capacitors. After being run daily for the last few days, it takes about 15 minutes for the fuzzy/wavy lines to disappear, on the "better" unit, and 30 minutes on the other two. I have not noticed any affect on the picture after the units are warmed up, but I need further testing to confirm.

So far, it seems that the "22" series have capacitor issues. The 522's were manufactured December 1996 and the 822 was manufactured Jan 1997. I only used them around once a year since 2011. The Sony SVO and SVP decks do not seem to have this issue. The only capacitor related color issue I noticed is with the TBC-2 card installed in the second 522 unit causing an orange-yellow color shift.

Still, I do like the output, perhaps it should be compared to the SR-W7U? The SR-W7U was manufactured February 2000, so it is only three years newer, and it was recently serviced by JVC. It seems like capacitors could affect any unit at any time but who can judge when the colors are wrong except for a calibrated monitor?

[msgohan]

I was under the impression that JVC wouldn't touch W-VHS models.

[Cingular]

JVC does not work on the W-VHS's. The JVC pro repair department does not normally work on anything over five years old. The tech did a favor for me because I called so many times and did not know where to take the unit for repair. Also, this unit was a one owner and I did not think it really needed any parts. The JVC tech told me that this W-VHS is just a regular VCR with a Digital circuit. He made it sound like any competent VCR repair person could do a repair or adjustment.

[Cingular]

Attached below are two capture comparisons using the movie "Titanic" with the BR-S522DXU, TBC on, PYB Enhance off, video out "normal" to Leitch DPS-475 set to factory defaults SDI out to a Kona LHI capture card to an 8 bit .mov file. The first video is Component out, and the other is S-Video out from the 522 and both are labeled as such. The Leitch TBC was engaged for both videos. The 8 bit mov was imported into Premiere CS5.5 and then exported to an H.264 Youtube SD mp.4.

For component cables were 2ft of Belden 1694a cable, and the S-Video is the OEM S-Video TecNec brand 7-pin s-video cable, the SDI cable is an 8ft Belden 1694A cable straight to the Koha LHI card.

The Component video seems to have better blacks and the skin tones do not have that reddish tint that the S-Video seems to have. What do you think? Do you think the Component output is better? What do you think about the reddish skin tones? Is that a symptom of S-Video output? Notice how that wavy line that is normally present at the bottom of the screen does not exist in the 522's output? Any other thoughts?

[msgohan]

I see the oversharpening halos that I had mentioned with my own unit.

The reddish hue is present on both; the component capture is merely desaturated in comparison. After matching the saturation, the hue is the same.

The clips have been downsized (instead of cropped) from 486 to 480, and then encoded in progressive mode. This destroyed the interlacing structure, and the poor compression quality has smeared away any remaining detail. The same applies to your Hi8 clips in the other thread. It isn't possible to make any real quality determinations.

[Cingular]

I captured both of these clips as an 8 bit .mov file, then I used h.264 Youtube SD in Premiere CS5.5. Exactly what file types should these be rendered to make them a reasonable side to view on this site?

Is there anything I can encode these 8 bit .mov files to make a reasonable quality determination? I tried a few combinations but it is hard to match that original 8 bit file exactly. Maybe next time I should try capturing straight to DV format instead of 8 bit.

The component capture does not look desaturated to me because the blacks look more black, and the skin tones have a lot less red. I am using a Quadro 2000D Graphics card output to 5bnc breakout cables on Belden 1694A to two Sony FW900 monitors recently Calibrated with an X-Rite I1 Display Pro.

[Cingular]

Are you sure that the component video is just "desaturated?" If that is true then why do the blacks look more black? Do you like the Component video or S-Video capture better? It seems like the reddish skin tint is a symptom of S-Video?


Attached is an 11 second straight capture to a DVCPRO50 .mov.

[Cingular]

The better blacks from the Component output on the BR-S522DXU only show up in 8bit capture not DVCPRO-50. In 8-bit capture even in S-Video blacks look more black and the skin tones and colors are different. Is that possible? Perhaps 8 bit capture is much better than I thought, everyone always said that 8-bit was not necessary and 10-bit was overkill but maybe it was because they were talking about amateur home movies?

Attached are DVCPRO50, 8-bit and 10 bit YCbCr Captures on a SR-W7U S-Video to a Leitch DPS-475 SDI to a Kona LHI. Also an 8 bit YCBCr capture on a 522 Component with the same setup.

Please let me know what you think.

[msgohan]

Thanks for the samples. See attachment.

Quote:

Originally Posted by Cingular

The component capture does not look desaturated to me because the blacks look more black, and the skin tones have a lot less red.

Quote:

Originally Posted by Cingular

Are you sure that the component video is just "desaturated?" If that is true then why do the blacks look more black? Do you like the Component video or S-Video capture better? It seems like the reddish skin tint is a symptom of S-Video?

I said "desaturated in comparison." Meaning, the saturation level is much lower than the S-Video capture. I could have also said the S-Video capture is oversaturated in comparison. I wasn't saying that either one is correct; the point is they can be adjusted so that the colors nearly match one another without touching the hue. The skin tint (hue) is equally reddish on both... almost. There is a slight push, but it's dwarfed by the saturation difference.

Black level is separate from saturation.

These are the tweaks I used in the screenshot, quickly trying to match the S-Video to the component: ColorYUV(off_y=-2,gain_y=-3,cont_u=-80,cont_v=-90). Bringing the blacks down slightly, bringing the brights down slightly, and desaturating with a small bias.

The "correct" colors are something else entirely. In this case, the VHS color timing would match the DVD that was produced from the same master, which I included in the attachment screenshots.

The component capture has some sort of diagonal interference pattern in the chroma, though I only noticed it in the "Chroma only" view.

I've only looked at the BR-S522 captures so far, due to time. There's some sort of "grid" pattern across the videos, including the 8-bit capture. Pulling up the SR-W7U 8-bit capture, this artifact doesn't seem to be present. I assume it's caused by the deck's built-in TBC, as my BR-S525 with no TBC card doesn't show that.

[Cingular]

Thanks for explaining about the saturation and adjustments, and chroma.

After re-caibrating monitors, I cannot see much difference between 10 bit, 8 bit YCbCr and DVCPRO50. I used the same settings and captured directly to the Kona LHI capture card using the OEM breakout cable.

I did a test capture straight to the Kona LHI card to see if the Leitch was making any changes and I saw a significant change in the Component Video capture. The change is dramatic. The Leitch DPS appears to be de-saturating the Component output of the BR-S522?

The SR-W7U does seem to be a closer match to the original DVD. I did see the grid pattern, in the chroma. Perhaps that grid pattern is induced by the Leitch DPS because it makes a significant change to the image? Do you see the grid pattern in the straight Kona LHI capture?

Does the Kona LHI straight capture handle the Chroma better?

Based on your screen captures, I did not notice the Grid Pattern in the 522's S-Video, maybe only slightly? The grid pattern is more pronounced in the Component Capture with the Leitch.

Can you measure any affect the Leitch DPS could be having on the 8 bit or DVCPRO50? Is it possible that compressing an 8 or 10 bit file and then uncompressing can have any affect on the images? I would say no, but I have never researched the effect compression and decompression can have on image quality.