VHS capture comparisons blind test

[Mr Chris]

I've been experimenting with my setup and doing some test captures, and thought there was a chance some on this forum may like to play along at home. The source is a relatively noisy but fairly stable off-air VHS recording, and the end result is a Lagarith capture via a Hauppauge USB-Live2. But what hardware have I used in between? Well, in the interests of science and impartiality, I'll leave it a little while, and if there's any interest in this thread, I'll give all the details then.

There follows 10 screenshots - each from a slightly different setup - and attached below that are 10 corresponding video samples for your perusal. (I realise these are big files so this may be overkill, but take your pick.) Each of the samples have been through a "Direct stream copy" in VirtualDub to match their starting point, so they can easily be dropped into the NLE of your choice and compared by eye and with whatever scopes you like. The files are 25fps PAL, 4:3, interlaced (upper field first), YUY2, with no audio.

Which is your favourite? Any and all observations and opinions will be interesting to read, and I'll do my best to answer any questions you may have.

Screenshot A

[Attachment 50136 - Click to enlarge]

Screenshot B

[Attachment 50137 - Click to enlarge]

Screenshot C

[Attachment 50138 - Click to enlarge]

Screenshot D

[Attachment 50139 - Click to enlarge]

Screenshot E

[Attachment 50140 - Click to enlarge]

Screenshot F

[Attachment 50141 - Click to enlarge]

Screenshot G

[Attachment 50142 - Click to enlarge]

Screenshot H

[Attachment 50143 - Click to enlarge]

Screenshot I

[Attachment 50144 - Click to enlarge]

Screenshot J

[Attachment 50145 - Click to enlarge]

[Brad]

Something wrong with one of the heads on the VCR that recorded this? Even-numbered fields look less noisy. Odds play back with more noise, more halos, more rainbows, more wiggle... The alternation during the crowd shot at the end is fantastic.

And a minor thing I noticed: at least with some of the files, at the top of the screen, in the middle, there are a few lines with an artifact that alternates between green and magenta every frame.

Now the meat. Hopefully I didn't typo any of the letters. It was fun, but I should get to bed now, so I can't go any further.

---

A, B, J were played back on VCR1, using the composite output.
- A: heavy rainbows
- B: minimal rainbows, notable 3D comb filter errors (dots in saturated colors and especially along the head-switching noise)
- J: tough to tell comb filtering through the diagonal mess

VCR1 has better dropout compensation than VCR2 (seen at 0:08).

C, F, H: VCR2 adds halos and a sort of diffuse outline
D, E, G: VCR2 sharpening increased; more halos, less diffuse-ness
(VCR1 is not halo-free, mind you.)

Except for A & B, all of the samples are shifted down 2 lines via added black (1 line per field). E & F have Widescreen Signaling (WSS) added to the first half of analog line 23, which I have seen with samples from PAL LG DVD recorders.

A, C, D are direct to capture device from VCR: the head-switching area extends all the way from the left to the right edge, which violates analog blanking.

A, D have wiggles. C has fewer wiggles, so C has VCR TBC on.

Much less rainbowing with C, D compared to A so I will assume that VCR2's S-Video output has been used.

F is VCR TBC & DVD recorder TBC both used.
E is VCR TBC off & DVD recorder TBC used.

H, I are also VCR TBC on, but I can't guess what's in between the VCR and the capture device.

G straightens lines below the head-switch point, so VCR TBC is off and a different TBC is used. Panasonic passthrough?

The frame that's numbered 316 in the others is missing from H, and the complete file is 1 frame shorter.

[Mr Chris]

That's sensationally impressive, vaporeon800 - you're right about practically everything, you've spotted just about all that I wanted to bring up, and you've even raised a couple of points I hadn't noticed! I'll address your post straight under this one, but first the explanation...

I used two tape decks for these samples - one VHS and one SVHS:

VHS = Sharp VCMH704
SVHS = Panasonic NV-HS1000

I also used two DVD recorders for passthrough:

Sony = Sony RDR-GX350
Pana = Panasonic DMR-ES10

Where I list either [Scart|Composite] or [Scart|S-Video], I'm using this style of 'breakout' cable:


[Attachment 50584 - Click to enlarge]

The point really is that the VHS and Sony are consumer-level products which I bought for normal consumer reasons way back, while the SVHS and the Pana are more recent purchases and my attempt to upgrade specifically for capturing. How do they compare? Here are the chains I used for the above samples:

Capture A
VHS - [Scart|Composite] - Hauppauge

Capture B
VHS - [Scart|Composite] - Sony - [S-video] - Hauppauge

Capture C
SVHS (TBC&AI on) - [Scart|S-Video] - Hauppauge

Capture D
SVHS (TBC&AI off) - [Scart|S-Video] - Hauppauge

Capture E
SVHS (TBC&AI off) - [Scart|S-Video] - Sony - [S-video] - Hauppauge

Capture F
SVHS (TBC&AI on) - [Scart|S-Video] - Sony - [S-video] - Hauppauge

Capture G
SVHS (TBC&AI off) - [S-video] - Pana - [Scart|S-Video] - Hauppauge

Capture H
SVHS (TBC&AI on) - [S-video] - Pana - [Scart|S-Video] - Hauppauge

Capture I
SVHS (TBC&AI on) - [Scart|S-Video] - Pana (front) - [Scart|S-Video] - Hauppauge

Capture J
VHS - [Scart|Composite] - Pana (front) - [Scart|S-Video] - Hauppauge

(The AI switch is supposed to be a one-stop-shop for the SVHS's optimal settings - the jury's very much still out on that particular claim - and it overrides certain settings made elsewhere. It's recommended to be used in conjunction with the TBC, so I've basically treated them as one - either both on or both off.)

[Mr Chris]

Something wrong with one of the heads on the VCR that recorded this? Even-numbered fields look less noisy. Odds play back with more noise, more halos, more rainbows, more wiggle... The alternation during the crowd shot at the end is fantastic.

I think you're right - it's part of the recording rather than a playback problem. That wasn't part of the test, though.

And a minor thing I noticed: at least with some of the files, at the top of the screen, in the middle, there are a few lines with an artifact that alternates between green and magenta every frame.

I hadn't even spotted that, and now I can't unsee it! No idea what can cause that - some sort of chroma shift? I'm not going to lose any sleep over it, though.

A, B, J were played back on VCR1, using the composite output.

100% correct.

- A: heavy rainbows

That's the direct capture, so I suppose that makes sense.

- B: minimal rainbows, notable 3D comb filter errors (dots in saturated colors and especially along the head-switching noise)

That's going through the Sony, which to my eyes improves the picture (over a direct capture), but also introduces problems. I think there may be level issues with it in that it dulls the highlights slightly, and going from a very light picture to black usually results in an intermediate navy blue frame. (The latter's not evident from these samples, of course.)

- J: tough to tell comb filtering through the diagonal mess

Yes, I and J are the only samples going through the Pana's front inputs (one composite, one S-video), which are fundamentally broken. I've no idea whether this is a problem with my specific one, or whether the ES10 has this issue in general.

VCR1 has better dropout compensation than VCR2 (seen at 0:08).

YES. This has been a real bugbear of mine, ever since I got the SVHS. In fact, it's almost a deal-breaker - for all the problems the VHS has, the intrusive glitches served up by the SVHS are just about enough for me to go with the former. Am I mad for thinking that way?

C, F, H: VCR2 adds halos and a sort of diffuse outline

Those are the ones with the TBC and AI switched on.

D, E, G: VCR2 sharpening increased; more halos, less diffuse-ness

TBC and AI off, which means that the Sharpness slider kicks in. Now, there's a notch that causes the slider to come to rest in the centre, which led me to assume that was the neutral position (and that's how all of these were captured). Now I'm not so sure.

(VCR1 is not halo-free, mind you.)

I've searched the menus for a sharpness or noise reduction setting, but there isn't one. It's probably set to artificially sharpen slightly. (That said, some of the ghosting may be poor reception when recording. So many variables!)

Except for A & B, all of the samples are shifted down 2 lines via added black (1 line per field).

Those are the direct captures. Looks like both my DVD recorders do the two-line shift.

E & F have Widescreen Signaling (WSS) added to the first half of analog line 23, which I have seen with samples from PAL LG DVD recorders.

Both SVHS through Sony. Strange.

A, C, D are direct to capture device from VCR: the head-switching area extends all the way from the left to the right edge, which violates analog blanking.

Absolutely - well spotted!

A, D have wiggles. C has fewer wiggles, so C has VCR TBC on.

Ditto

Much less rainbowing with C, D compared to A so I will assume that VCR2's S-Video output has been used.

Indeed - always S-video where possible.

F is VCR TBC & DVD recorder TBC both used.
E is VCR TBC off & DVD recorder TBC used.

Yes. Well, it's the Sony DVD recorder, so whether that counts as a TBC or not is up for debate, but in essence you're right.

H, I are also VCR TBC on, but I can't guess what's in between the VCR and the capture device.

It's the legendary ES10!

G straightens lines below the head-switch point, so VCR TBC is off and a different TBC is used. Panasonic passthrough?

There we are.

The frame that's numbered 316 in the others is missing from H, and the complete file is 1 frame shorter.

Yes - well spotted (again). I found this particularly ironic as, on paper, H should be the most stable of all the captures (SVHS, TBC, ES10...), yet it's the only one that skipped a frame.

Thanks for taking the time to download, watch, and reply. Do you find any of the above surprising? I'm still tinkering and trying to get the best setup that I can, so any and all suggestions will be gratefully received and considered. I may even post a couple more samples. You have been warned!

[oln]

That's going through the Sony, which to my eyes improves the picture (over a direct capture), but also introduces problems. I think there may be level issues with it in that it dulls the highlights slightly, and going from a very light picture to black usually results in an intermediate navy blue frame. (The latter's not evident from these samples, of course.)

I have the HDD version (RDR-HX750) of this, I do captures from the HDMI output. Based on the manuals the GX350 seemed to have the same TBC capabilities (same NEC chipset). There are some brightness/level settings that can be adjusted on these to play around with under Picture adjustment, it's set separately for each input. I don't remember the exact settings, but I have changed them a little. To avoid clipping adjust black/white level a bit and enable White AGC which brings down the white level a little if it's too high as the HDMI output is limited to 16-235. Doesn't need as much adjustment for analog outputs. There is also noise reduction on/off and comb filter settings there. My subjective impression is that putting 3D Y/C (comb filter) all the way to the right (Still) worked the best for VHS sources when using the composite in but ymmv.

Yes, I and J are the only samples going through the Pana's front inputs (one composite, one S-video), which are fundamentally broken. I've no idea whether this is a problem with my specific one, or whether the ES10 has this issue in general.

I think it's a common issue with PAL panasonics. My ES10 has several borked (non-working) inputs, which includes working S-Video input. What I've read suggests it's the input/output switcher chip that breaks for whatever reason, possibly due to the dodgy capacitors they used in the power supply that frequency fail (replaced 2 bulgy and leaking ones in mine but it didn't fix the issue).

I have a EH57 as well, which seems have pretty similar tbc capabilities, it has a similar issue and was completely dead until I replaced a bad power supply capacitor that had blown. Thankfully one of the S-Video inputs still work on this one.

One issue I have with both, and I've been told is a problem with other PAL panasonic DVDRs as well is that they clip bright spots, so on tapes where I need them I have to reduce the video signal level before it gets into the recorder if I don't want them to be badly clipped.

YES. This has been a real bugbear of mine, ever since I got the SVHS. In fact, it's almost a deal-breaker - for all the problems the VHS has, the intrusive glitches served up by the SVHS are just about enough for me to go with the former. Am I mad for thinking that way?

I have the same beef with the NV-HS1000 we got here. I did some adjustments that made it a bit better, avoiding the black lines at least, though the DOC is still not great even when comparing with lower-end decks. It seems to be very limited in how many lines of dropout the DOC is active for before turning to noise. It may be that there is something else that can be done to improve it but I haven't found anything so far. Maybe the adjustment will help for you as well, though it's a bit tricky as it's hard to see the video and adjust at the same time (it's supposed to be serviced with some extension cables so the boards can be taken out but have to make do without that.)

Note that there are two revisions of the NV-HS1000 (in addition to variations with different tuners), it seems the newer one (NV-HS1000 EGC/BYP/ECP) improved the DOC a bit among other changes, and the video head drum is different. The one we got here is the older EC variant, since you have the same issues I'm guessing that's what you have as well.

The bad DOC, and tendency for the heads to get clogged makes me not use it very much unless I'm dealing with a pristine tape.

TBC and AI off, which means that the Sharpness slider kicks in. Now, there's a notch that causes the slider to come to rest in the centre, which led me to assume that was the neutral position (and that's how all of these were captured). Now I'm not so sure.

I've found it to be a bit too sharp at the middle, so I turn it town a bit further. AI seems to always be a bit too sharp. Keep in mind VCRs are set up with viewing on a CRT in mind so it's not uncommon for them to over-sharpen a bit. You can also turn off most of the filtering (disables sharpness and AI settings) by changing one of the switches on the front from player to passive. I find it to be a bit sharp and it can get overly noisy on crummy tapes though.

[Mr Chris]

Plenty of food for thought there, oln. Thank you very much.

I have the HDD version (RDR-HX750) of this, I do captures from the HDMI output. Based on the manuals the GX350 seemed to have the same TBC capabilities (same NEC chipset). There are some brightness/level settings that can be adjusted on these to play around with under Picture adjustment, it's set separately for each input. I don't remember the exact settings, but I have changed them a little. To avoid clipping adjust black/white level a bit and enable White AGC which brings down the white level a little if it's too high as the HDMI output is limited to 16-235. Doesn't need as much adjustment for analog outputs. There is also noise reduction on/off and comb filter settings there. My subjective impression is that putting 3D Y/C (comb filter) all the way to the right (Still) worked the best for VHS sources when using the composite in but ymmv.

Well you've just caused me to discover a menu I'd never seen before in the decade or so I've had the Sony, so boy do I feel silly. I'm usually a stickler for settings, too. Nevertheless, from experience, I'd have to describe the TBC(ish) capabilities of the unit as good but unreliable. It stabilises everything brilliantly, apart from when it doesn't. That is to say, everything's great when everything's great, but it doesn't take much of a tape glitch to roll the picture or do a funny little jump. The Pana just seems a lot more reliable.

One issue I have with both, and I've been told is a problem with other PAL panasonic DVDRs as well is that they clip bright spots, so on tapes where I need them I have to reduce the video signal level before it gets into the recorder if I don't want them to be badly clipped.

From my experiments, both of the Pana's S-video outs (including the Scart out there) are right on the edge of clipping, particularly in the reds, but nothing unacceptable. Maybe I'm missing something. On the other hand, the component out seems to be blow the highlights beyond belief.

I have the same beef with the NV-HS1000 we got here. I did some adjustments that made it a bit better, avoiding the black lines at least, though the DOC is still not great even when comparing with lower-end decks. It seems to be very limited in how many lines of dropout the DOC is active for before turning to noise. It may be that there is something else that can be done to improve it but I haven't found anything so far. Maybe the adjustment will help for you as well, though it's a bit tricky as it's hard to see the video and adjust at the same time (it's supposed to be serviced with some extension cables so the boards can be taken out but have to make do without that.)

Thanks for that link - like I said, plenty of food for thought, though my VCR maintenance skills sadly don't go much further than cleaning the heads. I'll put your advice down as "Plan Z".

I've found it to be a bit too sharp at the middle, so I turn it town a bit further. AI seems to always be a bit too sharp. Keep in mind VCRs are set up with viewing on a CRT in mind so it's not uncommon for them to over-sharpen a bit. You can also turn off most of the filtering (disables sharpness and AI settings) by changing one of the switches on the front from player to passive. I find it to be a bit sharp and it can get overly noisy on crummy tapes though.

I've been doing more test captures this weekend, and to prove your point:

Screenshot Q

[Attachment 50675 - Click to enlarge]
VHS - [Scart] - Pana - [Scart|S-Video] - Hauppauge

Screenshot R

[Attachment 50676 - Click to enlarge]
SVHS (TBC and AI off, Sharpness minimum) - [S-video] - Pana - [Scart|S-Video] - Hauppauge

Screenshot S

[Attachment 50677 - Click to enlarge]
SVHS (Passive mode) - [S-video] - Pana - [Scart|S-Video] - Hauppauge

Sample vids below, and you may want to compare with Sample G in the original post, which is the same as Sample R, except the Sharpness is in the central position. Passive mode gives a very good picture, but it's sharpening the tape noise in a way which renders it just about useless for anything other than a pristine tape. And then there's the drop-out compensation... I just can't get past that. I hate myself for this, but I think I favour my admittedly-flawed Sharp VCMH704 consumer-level VHS ahead of the Panasonic NV-HS1000, and it's all down to the DOC. It's quite maddening, really. Am I going crazy?

[oln]

Yeah you pretty much sum up my experience. The video quality is otherwise excellent, but the DOC on the NV-HS1000EC is really not good, even when adjusted to avoid the black stripes. Pretty much all my consumer decks other than one ancient one I have do a better job at it, even the low-end 2 head ones. As it's from the early 90s I wouldn't expect it to be as good as a late 90s early 2000s JVC but I still wonder if there is some design flaw, maybe a sub-par component that gives these models an issue.

I usually put the sharpness a bit below half-way if I use it, but that's not all that often.

The Sony/Pioneer DVRs have that drawback as mentioned, not as reliable as the panasonics. The service menu has a sync level setting, I've wondered if that would have any impact but I don't have a service remote or replacement to test it with.

From my experiments, both of the Pana's S-video outs (including the Scart out there) are right on the edge of clipping, particularly in the reds, but nothing unacceptable. Maybe I'm missing something. On the other hand, the component out seems to be blow the highlights beyond belief.

I've mainly had clipping issues with home tapes, commercial tapes are usually not as bright-heavy. Haven't tested the component outputs, but the HDMI out on the EH57 is especially bad at clipping. I have a home-made RCA volume control together with a splitter to bring down the video levels a little when I need to use the Panas. Older VCRs often have a pot for adjusting video level as well, though they're meant to be adjusted a small number of times. Newer JVCs have the y level in the service menu, but service remotes seem rather rare.

[Mr Chris]

I think I'm slowly coming to the realisation that what I thought was a good well-researched purchase isn't quite the upgrade I was hoping for. I'm not even sure it qualifies as an upgrade at all, to be honest, and that's despite the NV-HS1000 being listed in the much-vaunted DigitalFAQ VCR Buying Guide. Not only listed, in fact, but in bold text as one of the "most respected machines". I understand the features deserve that in theory, but I'm baffled how a deck with such a fundamental flaw could achieve such a ranking - unless later versions really were a huge improvement. Never mind.

I'm strongly considering going on the look-out for one of the JVCs on that list now...

[oln]

Yeah I suspect the newer ECG/ECC/BYP revision (the electronics and drum were altered as you can see in the service manual here) may have fixed some of the dropout issues. As you can see in the thread quasipal had one that didn't have the black streaking issues, but I never got an answer on whether that one was a different revision. I seem to remember them also noting something about a HS-800 (Basically NV-HS1000 without TBC, they have the same SM for the newer revision) having improved DOC compared to their NV-HS1000 models in another thread but I can't find it right now.

[Mr Chris]

Turns out I've been making an incorrect assumption: having just had a closer look at everything, it seems I actually have the BYP version. The plot thickens!

[oln]

Humm, that's interesting, seems my assumption was wrong and both revisions may be suspectible to this.