JVC VCR Interesting Overscan issue

[Blackout]

hi,

i have two JVC VCRs, a HR-S9500 and a HR-S5700.

Obviously the HR-S9500 is better. But for whatever reason, the HR-S5700 images are SHARPER. maybe its the heads. anyways.

The problem is that the HR-S5700 seems to create a bigger "overscan" blurr bar at the bottom of the picture than the 9500. It is considerably bigger. about 3x as big. When looking at captures from both units, it is obvious that the 5700 overscan blurr is losing more picture. ie when i look at the station watermark, the overscan blurr bar across the bottom has creeped up a *lot* higher towards the station watermark on the screen than the 9500.

i understand all videos create an overscan area but i thought it was the same on all video recorders and was an inherent part of the video tape but this seems to prove otherwise as i am getting condsiderably different overscan amounts for different VCRs. so it must be a machine thing.

So my next question is ....where is the overscan adjustment inside? is it something as simple as video head roller guides adjustment or is it an electronic signal adjustment.

many thanks,
Blackout.

ps i have been a member on Videohelp since 2002 and i have just found this excellent forum for more technical issues. Thank you for the Download Service manual section ! its good to see the ripoff sites charging for these being stamped out

[admin]

Quote:

But for whatever reason, the HR-S5700 images are SHARPER. maybe its the heads.

Yes, heads.

Quote:

the HR-S5700 seems to create a bigger "overscan" blurr bar at the bottom of the picture than the 9500

The noise you're seeing is technically referred to as "head switching noise" and happens in the lower portion of the overscan. The difference in heads and transports is what causes this.

The main difference between the 9500 and 5700 is the dynamic drum, when discussing heads. That alone could account for the differences in noise.

Consumer VCRs generally have poorer performance overall, and it's often readily observed by the size of the head switching noise. (Although I have seen VCRs with huge noise bands, with otherwise stable playback. The low-budget mono 2-head Orion VCRs, for example.)

It's not so much a part of the unique tape as it is of the format itself. VHS VCRs have head switching noise -- all the way to the best machines that exist.

Sometimes you can "fix" this by adjusting the tape roller guides. (See attached image.) However, I wouldn't suggest it unless the noise was leaking well into the non-overscan visible portion of the image. You could just as easily misalign or otherwise damage the VCR by randomly twisting or turning the internals.

When capturing videos, it's best to simply mask that overscan with black. If it appears that the image is off-center, then you can re-center it by cropping the video 2 pixels at a time, then re-padding it with a black matte. Using VirtualDub, for example.

Quote:

ps i have been a member on Videohelp since 2002 and i have just found this excellent forum for more technical issues.

Quite a few video professionals and hobbyists come here, and those same folks are known to haunt Videohelp, Doom9 and a few others. This site was founded in 2002 (under a different name, renamed to digitalFAQ.com in 2004), although it pulls from information that was published as far back as 1999 for earlier online projects.

Glad you found the site, and welcome.

[Blackout]

hi Admin, thank you for your response....

Quote:

Originally Posted by admin

The noise you're seeing is technically referred to as "head switching noise" and happens in the lower portion of the overscan. The difference in heads and transports is what causes this.

The main difference between the 9500 and 5700 is the dynamic drum, when discussing heads. That alone could account for the differences in noise.

um....but the 5700 looks sharper. its the unit WITHOUT the Dynamic Drum. that logic does not make sense?

i dont understand why you have brought the term "noise" into the discussion as something separate from the overscan noise bar. You mention noise within the overscan? the overscan already looks like complete noise! i am talking about the overscan strip at the bottom of the picture only.

Quote:

Originally Posted by admin

It's not so much a part of the unique tape as it is of the format itself. VHS VCRs have head switching noise -- all the way to the best machines that exist.

i understand its part of the VHS format. i accept that. i am not complaining about the fect that there is overscan noise, only the amount for the 5700. it should be the same on all machines surely...

Quote:

Originally Posted by admin

When capturing videos, it's best to simply mask that overscan with black. If it appears that the image is off-center, then you can re-center it by cropping the video 2 pixels at a time, then re-padding it with a black matte. Using VirtualDub, for example.

yes i have been doing this with cookie cutter in Vegas. (zooming in to remove it breaks the interlaced picture and it becomes a mess...wish Vegas would fix that).

but i have to "black out" a lot more of the picture withthe 5700 vcr than the other. that is my problem. i dont like the fact that to get a sharper picture and use the 5700 i end up losing more of the picture data to a black strip across the bottom if i plan on removing the overscan. on the 5700 my black strip ends up having to be bigger.

Quote:

Originally Posted by admin

Glad you found the site, and welcome.

thank you i have attached two pics to hopefully explain the problem. first pic is my 9500. you can see the overscan is small at the bottom. then the 5700 which has considerably more overscan (ie considerably less picture information left).

Kind Regards,
Blackout

[lordsmurf]

Quote:

um....but the 5700 looks sharper. its the unit WITHOUT the Dynamic Drum. that logic does not make sense?

The dynamic drum doesn't necessarily control sharpness. Don't merge otherwise separate concepts -- you'll just confuse yourself. The difference in noise bar could simply be the difference in heads, as proven by the fact that one is a dynamic drum and one is not.

Quote:

i dont understand why you have brought the term "noise" into the discussion as something separate from the overscan noise bar. You mention noise within the overscan? the overscan already looks like complete noise! i am talking about the overscan strip at the bottom of the picture only.

Again, I think you may be merging some concepts here. The "overscan" is a physical area outside the boundaries of the normal viewing area. One reason this area is exists is because it can (and usually does) contain various types of noise -- visual artifacts from closed caption pulses, head switching noise, broadcast pillars, etc.

The noise you have here is "head switching noise", which is a technical term for VCRs.

Quote:

only the amount for the 5700. it should be the same on all machines surely...

Nope. Though not common, professional decks can even have more noise than consumer decks.
For example, in this thread from today -- look at the photos: Panasonic AG-1980 vs. consumer deck...strange...

It's simply an issue of the condition of the unit, the condition of the tape (including the alignment of the machine that made the tape), as well as the head itself. Even machine to machine, on the same model of VCR, this noise can vary.

Quote:

yes i have been doing this with cookie cutter in Vegas. (zooming in to remove it breaks the interlaced picture and it becomes a mess...wish Vegas would fix that).

With Premiere, you use the "clipping" filter to mask with black. It's not the crop filter. I don't use Vegas enough to remember what's what.

Neither Vegas nor Premiere are very good for filtering videos to any advanced degree (excluding color correction), so you'd be best to run it through VirtualDub first, then run it into Vegas for editing second. That's what I have to do for many videos. It's rather ironic that a freeware tool can outperform the $600+ (or whatever it is, I forget) professional NLE.

Quote:

but i have to "black out" a lot more of the picture withthe 5700 vcr than the other. that is my problem. i dont like the fact that to get a sharper picture and use the 5700 i end up losing more of the picture data to a black strip across the bottom if i plan on removing the overscan. on the 5700 my black strip ends up having to be bigger.

Well, assuming this is being created for TV viewing, it's all cropped in the final image anyway. Is this just an academic issue, or is there some specific reason it needs to be smaller -- even if it's long-term cropped from the final viewing version?

Quote:

i have attached two pics to hopefully explain the problem. first pic is my 9500. you can see the overscan is small at the bottom. then the 5700 which has considerably more overscan (ie considerably less picture information left).

That is pretty significant in size, yes --- but I've seen worse. That is a rather old VCR, and there is a chance it has become shifted from perfect alignment over the years. Gravity alone has an affect on VCRs. Even unused, I have machines sitting here that have drifted out of alignment, as time was not kind. I need to get about 4-5 decks repaired, thought I may only fix 2-3 due to budget. (These have drifted too far for my self-repair work.)

[Blackout]

Hi Lordsmurf. Thank you for your time responding. i consider you one of the gurus of VCRs and transferring and it is an honour for me to have your reply. Much respect.

Quote:

Originally Posted by lordsmurf

I have machines sitting here that have drifted out of alignment, as time was not kind. I need to get about 4-5 decks repaired, thought I may only fix 2-3 due to budget. (These have drifted too far for my self-repair work.)

please specify the alignment process you are going to perform to repair this problem. Is it only in the two roller guides next to the head? or is it more

Much appreciated,
Blackout

[admin]

JVC guides can easily be "fingered" -- just turn with your fingers, when a tape is ejected. I do this when I'm in a rush and simply testing (pre- fine-tuning). For more precision, or to do it while playing, you'll need a special JVC guide adjustment tool. I see them on eBay from time to time. I know I saw one last month. You could also Dremel a flathead screwdriver to fit the slot.

Many VCRs are hard to adjust, but not JVC.

Just be aware that you can feather a tape easily if you adjust poorly. So use a test tape, when possible. Of course, that won't be plausible if adjusting for a misaligned recording. Just be careful.

[Blackout]

i have now got 9 VCRs. its quite amazing the difference in quality and playability.

i still dont feel like i have a true "golden child". I think the point about the art of video transfer is that you probably never feel like one unit is truly the best for all jobs.

Generally everyones comments on here hold true for me also. Sony VCRs are crap. Its down to the Pana and JVCs. And the best are the ones with TBC and filters.

A few things i notice:

1:/ As experienced by others, the JVCs (i have a 9500, 9800, 7600AM and 5700) are softer pictures than the Panas. But the noise red and TBC are STRONGER. ie the noise is cleaned up more significantly and when i have an old tape (from the 80s) that the other machines cannot track cleanly to) the JVCs consistently get the best extract off the tape.

2:/ JVC mechanism build quality is consistently crap. Tapes constantly jam, (do not fast forward and then hit play without first hitting reverse for a few seconds, or you get tape streams spitting out) etc.

3:/ JVC audio varies from machine to machine. Sometimes you need to extract video from one take/machine and then get the audio from another.

3:/ Panas (960 and 860) are the SHARPEST. But they are a darker picture from the JVCs. So when you re-adjust the brightness and gain to match, the picture difference is not as significant. But certainly the best for sharpness.

4:/ Where Panas seem to fall down is their NR, which seems to leave a lot more tape noise remaining. A Pana tape cleaned further with NeatVideo plugin afterwards is indeed magnificent.

5:/ i also seem to see an instability in the picture when it is very bright on the Panas...when theres lots of white on the screen. You notice it on the right border, which is not straight but goes all wobbly for those frames. I wander if its a TBC problem. But indeed the TBC is not as "strong" on the Panas. For some really old tapes where the field order seems to *flip* 4 or 5 times every half-hour the pana just falls apart...the picture goes to a blurry smudge. The JVC marches on.

Just observations.

Cheers,
Blackout

[admin]

Quote:

Originally Posted by Blackout

i have now got 9 VCRs. its quite amazing the difference in quality and playability.

Welcome to the club.

Quote:

1:/ As experienced by others, the JVCs ... are softer pictures than the Panas

This is mostly because Panasonic artificially sharpens even at the defeat/zero setting. It's easy to detect, by way is signal ringing -- little halos and echoes in the image. You have to push the slider a bit into Soft on the Panasonic AG-1980s, in order to match true signal. Now from there, the JVC can be softer OR sharper OR the same, as the decks vary a little form model to model.

Quote:

2:/ JVC mechanism build quality is consistently crap. Tapes constantly jam, (do not fast forward and then hit play without first hitting reverse for a few seconds, or you get tape streams spitting out) etc.

If tapes jam that much, you have problems with that(those) specific deck(s). I have 1-2 that are this way, and it's why they're really not used very much. The others are pretty much flawless on tape in/out transport. When the okay decks act up, it generally turns out to be a tape clamshell defect.

Aside from that, your other observations are pretty much right on target.

[Blackout]

hi Admin,



however the NV-HS860 and 960s do not have a "sharpen" slider.

I did try taking a jvc image and sharpen it (using Vegas Sharpen plugin, set to 0 (minimum)), and compare it to the Pana image. The sharpen caused some ringing and halos as you describe. but the Pana image seemed really flat with no ringing at all, for the same sharpness.

Take a look a the images attached.

Image 3 = JVC HR-S7600AM
Image3.jpg

Image 4 = JVC HR-9500 (almost identical to the 7600 image)

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

Image 4b = above image sharpened by Vegas 10 "Sharpen"
Image4.jpg

Image 5 = Pana NV-HS960.

You must be logged in to view this content; either login or register for the forum. The attached screen shots, before/after images, photos and graphics are created/posted for the benefit of site members. And you are invited to join our digital media community.

all vcrs had TBC on and NR filter on.

the Image 5 (Pana) is obviously the sharpest. But not the cleanest!

Cheers,
Blackout

[admin]

The Pana NV-HS960 definitely has ringing. Notice the white line that forms along his chin and down his neck.

Vegas 10 was opposite -- it used black instead of white. Also with a lousy method, as it looks terrible.

In both cases, it's definitely over-sharp.

The 7600AM image is softer than the 9500. The 9500 is pretty much accurate on all counts for VHS source, in terms of sharpness, detail, resolution and color quality. The TBC must be on, too, because chroma noise is suppressed.

The HS960 has a ton of dot crawl present, which creates what looks like grain, which gives a false appearance of detail.

Granulation of an image, to give false perception of detail, is an old photography trick. Slightly fuzzy ISO 3200-shot film was often published because the high grain of that speed of film somewhat hid the soft-focus or blurring issues. I used to shoot 800 Fuji at 3200, as did many others in the 1990s, because of even grain with acceptable color. And even in situations where light was inadequate, resulting in slight blur in action photos, the grain would save you from having an unusable image.

The trick with grain is that it's a mix of darker and lighter pixels (viewed in digital space, at least) which gives a false illusion. The JVC is not showing any less picture, but the Panasonic is simply showing more noise.

The reason you do NOT want the noise, is because there are algorithms that can truly increase detail. Faroujda boxes, for example. Even the resolution booster in the Elite Video BVP4+ is somewhat useful in this area, as are the various detailers like the DR-1000. But those only work if the image is relatively noise free. Otherwise you just enhance the noise mask, and the image looks bad by anybody's account.

[Blackout]

Hi Admin,

thank you for taking the time to respond.

"The HS960 has a ton of dot crawl present, which creates what looks like grain, which gives a false appearance of detail. "

i know about this phenomenon. the grain is not appearing to make the Pana appear sharper. it IS SHARPER. why do i know this? because when i clean up the last image (Pana NV) with NeatVideo (which removes all the noise grain..and with the sharpen slider in the plugin set to ZERO) then the final clean image is still sharper than the equivalent clean JVC image.

in other words, the JVC image without grain = blurrier than the resultant Pana image without grain.

if you can tell a difference between the 7600 and the 9500 you are doing very well. The 9500 should be a bit better (to justify the extra cost of the unit ! hehe) but ill be damned if i can tell the difference. Save the two images to your HD and flick between them with Windows Picture Viewer which swaps them instantly and theres a slight shift in the image position but damn i cant see an improvement.

i am most interested in the detail algorithms you describe. are any of them residing in a plugin or are they hardware only...

Kind Regards,
Blackout

[lordsmurf]

Resolution scaling is best done with dedicated hardware. All HTDVs, for example, include hardware scalers. The output is always the maximum resolution of your TV: 720p or 1080p. Anything less is processed and scaled before being shown.

Faroudja is the brand name upscaler many respect.
Snell & Willcox have quite a few, too.

The Faroudja website (history page) makes for a great read. For example:

Quote:

In the 1980s, the company developed technologies for the deinterlacing of NTSC signals, including motion adaptive processing algorithms. In 1989 Faroudja invented and patented film mode detection, also known as inverse 3:2 pulldown detection. Faroudja was the only company in the world that had the ability to detect the original frames of film within the video stream and reconstruct an accurate image, free of motion artifacts containing full vertical resolution.

To learn more about this kind of tech, look at the ST white papers: http://www.gnss.com/tch_dcdi_overview.phtml
That one is mostly about HDTV resolution + deinterlacing.

MotionDSP Ikena (and to lesser extent, vReveal) have upscaling abilities. Same for dTective. Those programs are about $10k per footprint, and were created for the forensive video community.