Capture quality problem: Pinnacle 710, VirtualDub?
 

Hello,

I recently purchased new equipment for my captures, but am experiencing disappointing results regarding video and audio.

Equipment:
PC: Win7 64, 8 GB RAM, SSD
card: Pinnacle HD 710-USB - from LordSmurf
VCR: Philips VR1000 PAL/NTSC (JVC S7600 clone) - from VCRShop
cable: S-Video
VHS tapes: PAL, pristine condition
VirtualDub: no dropped frames, CPU never above 36%

VirtualDub 1.9.11 installed as recommended, and configured according to sanlyn and LordSmurf's inputs. HuffYUV selected.

Pinnacle 710 card installed following LordSmurf's instructions, and Crossbar Thing set to "1: Video SVideo in / 0: Video Decoder Out".

• Problem 1 - VirtualDub (PAL/NTSC)

I'm from Europe, so I set PAL_B (also tried PAL_G) as the format (Video > Capture filter > Video decoder). In Capture pin, I set 25fps, YUY2, 720x576.
I then tried to access the VCR's menus, but VirtualDub continued showing nothing but a black screen.
Eventually I tried changing the format to NTSC_M 720x480, which surprisingly allowed me to see the VCR's menu, although distorted (image 1):

Attachment 19825

I then tried lowering resolution to half (NTSC_M 360x240), and although tiny, the menu was workable (image 2):

Attachment 19826

BTW, all this is in Preview mode. Overlay deforms the screen (image 3):

Attachment 19827

Then I revert to PAL_B 720x576 to play tapes and capture, but having to go back to NTSC_M 360x240 each time I need to see VCR information on the screen.
Anyway, Problem 1 is sort of solved, and I only mention it in case it's a symptom for Problem 2.

• Problem 2 - VirtualDub (no Video source, S-Video, no Audio source)

The picture is ok with the new JVC clone, but does not look significantly better than the one provided by my old Panasonic NV-G45EO, which is much earlier, has only scart ports, and none of these later techs (S-Video, TBC/NR, Picture Modes, auto tracking, D3R).

Despite being well aware that VHS is not HD, I find the captures disappointing. If I use the old conversion method (Panasonic VCR > Sony DVD recorder > DVD rip), I get 720x576 VOB/MPG files that are a fraction in size of these huge 24GB-per-hour 720x576 AVI files, and look pretty much the same (sometimes slightly better, sometimes slightly worse).
Audio capture is significantly worse now, sounds are very muffled when compared to the old DVD rip method.
VirtualDub is set to Raw capture format > uncompressed PCM 48000Hz, stereo, 16-bit.
VCR is set to HiFi, but not surprisingly reverts automatically to Normal because none of the tapes are HiFi.

Something is wrong, I wonder if it's the S-Video cable, Pinnacle card, or software. I assume S-Video is automatically selected, as it's the only image cable coming out of the VCR. However, I see no way of confirming that in VirtualDub, because although the card appears listed in VirtualDub > Device, no Video or Audio sources appear (image 4):

Attachment 19828

Any thoughts? :confused:

Two final photos, showing the greatest improvement I achieved so far with the new method. Is this the best I can hope for?

old method, DVD rip (image 5)
vs.
new equipment, huge AVI (image 6).

Attachment 19829
Attachment 19830

In the lower right of your images, it shows 29.97 NTSC resolution.

Overlay does not work post-XP, only Preview mode.

To hear audio, tick Enable Audio Playback. Verify the timing settings, as the defaults are wrong.

Muffling is from the VCR, which may or may not be performing correctly. JVC audio can be more muffled, however non-HiFi it's generally tweakable at the audio alignment head. (Be careful adjusting that, as it can massively screw up the VCR, and it can then only be correct using scopes!)

VCR menus are usually an odd resolution like 270p which is can confuse a variety of capture cards. Your best bet there is to have a second monitor to see the menus connected to composite output of the VCR that is separate from the capture chain.

MPEG2 at 8+Mbit/s (DVD) honestly isn't too bad in most cases and I feel like that still beats a lot of the modern capture solutions with far less complexity. Lossless if done right should visually be a bit better (10-25% maybe?), but at the tradeoff of much larger file sizes and complexity of the capture chain. While hard drive space is cheap, it is more annoying to archive and manipulate the files when they get that big depending on how many files you work with.

Lastly, you won't get a result that is better than the source material, so the capture can only get so good. VHS_Decode is an interesting comparison where you might be able to squeeze out a little more quality depending on the tape, but then it gets muuuuuuch more complicated time wise, hardware wise, and hard drive space wise.

Also, I'm not sure if this is happening for you or not, but certain image files will get compressed when you upload them here, so the details that you are saying are much different between one screenshot and another might not be as apparent to us if that's happening. Main difference I can see between the two is that the second screenshot has more detail/noise preserved in the pavement which may or may not be desirable to you. The roof of the blue car also has a more gradual gradient of color change from top to bottom on the second image to my eye. All things being equal, I'd go for the second option, but I don't know if I absolutely would if one takes up 8x the hard drive space for the archival copy. Noise/grain kind of gives the appearance that an image has more detail and some people even add grain in post processing to help with that, particularly if uploading to YouTube where YouTube compression can blur away more detail if it doesn't detect motion in certain static/nonmoving areas(or grain changes) between frames in videos as it encodes.

Well look on the bright side, the new method shows two improvements

1. The DVD method cropped a few pixels on the top by masking it with black, the AVI has them uncropped/masked
2. The whites are not clipped (or at least, less clipped) as a bit more detail in the white cars have been recovered.

Can agree with that on the context of mono linear audio (probably the same with Hi-Fi but differences are so small that it should be easy to correct it in post)

Case in point, here are two audio captures from the SP Mode and Mono Audio release of Super Duper Baseball Bloopers by Sports Illustrated. First is my JVC VS-30U and second is my Panasonic AG-1980 from Aramkolt.

That was my point, I can only access the VCR's menus by changing from PAL to NTSC, and then back.

I had tried that, but audio comes out very chopped.

What should I change here then?

Attachment 19833

Right. My VOB/MPG files are 2GB/hour vs. 24GB/hour AVI files.
What I'm thinking is ditching the Pinnacle card completely and see how the new VCR behaves in the DVD rip chain, now that I can upgrade from a scart to an S-Video connection.

Nope, those are the actual captures just like I see them. It's that close. I can find more detail in the tree, pavement, cars, but I think the difference is minimal.

That's true.

Ya, you can get a bit different croppings from different cards. In NTSC land, "full frame" is actually 486 vertical lines. Capture cards usually chose 480 of those internally in hardware (so it isn't typically adjustable which you get to capture). The very top line at least in NTSC is the "half line" with each field containing 242.5 horizontal lines with the very top line being the "half" line. Some DVD recorders will mask the top couple of lines to sneak in closed captioning data if present I believe.

If you do find the DVD recorder method worthwhile and don't want to mess with physical discs, there are a few models of recorder that can instead record the MPEG2 files to an internal hard drive which you can then read with something like ISO Buster. Here's a Toshiba machine I modified to take a SATA drive (SSD in this case) so that the drive can be accessed externally for easy drive connection to the computer with ISO Buster on it which can be done with a SATA to USB adapter.

Attachment 19834

Attachment 19835

An alternative for disc-less MPEG2 captures would be an ATI AIW card directly encoding to MPEG2 and the bitrate there you can go as high as 15Mbit/s with that if I recall correctly. For that to all work, you do need to get ATI Mulimedia Center (MMC) installed and working which can be a little annoying driver-wise sometimes. I like the DVD/HDD recorder in that it is a standalone recording device which may or may not matter to you.

I too notice the small difference between the still frames from:

1. the older Panasonic non SVHS VCR > DVD recorder capture

versus

2. the recommended SVHS VCR > recommended capture card.

The minor clipping in the first looks like a minor gain mismatch between Panasonic VCR and DVD recorder input levels. My guess is with a Proc Amp added a minor gain knob tweak would probably have sorted that out.*

There is nothing wrong with compressed capture files (video or audio) so long as they're not too compressed. It's the result that counts.

*Some HDD DVD recorders have a proc amp built in. My old Pioneer DVR 630H-S has onboard manual tweaking of various input parameters including White Level, White AGC and Black Level. Personal settings can be stored in three presets.
Maybe the OP's Sony DVD recorder also has this in its input menu. It might be worth checking in the Owner's Manual's detailed index.

There is zero evidence of this. All tests done to date show degraded quality vs. quality VCR, TBC, and capture card. This person has a with-TBC JVC S-VHS clone, so quality will be better. Your statement is only true when using thrift-store low-end/garbage VCRs.

Settings are causing the audio problem.
Make these changes:

Attachment 19836

Quality is much less, and color resolution is halved, 4:2:0 vs. 4:2:2. Restoration (filtering, clean-up) becomes far more challenging when starting from that reduced state.

Something else is going on here, and we're all missing it. (Just for fun, also try PAL-I, but I doubt anything will change.)

^ I want to re-visit this comment from the first post.

I don't believe that at all.
Don't show still images, give us clips. The motion will easily betray the non-TBC VCR as having lower quality.

Thanks, that did improve audio playback during capture (much less chopped, but there's still a continuous crackling sound in the background). However, "sync audio to video by resampling the audio" is the only setting that doesn't give me out-of-sync A/V (even if occasionally having people's voices jump to a higher pitch, and then slowly come down to normal).

Pinnacle vs. DVD rip: Yes, to my eyes my latest tests do show that the Pinnacle provides slightly better definition and color than the DVD rip. Since I'm in this for the best quality I can get, I'll go with the Pinnacle and cope with the huges files.

VCR menu: No change with PAL_I. The only time I can access the VCR menu in PAL is right after a capture. Tape stops, and the blue screen appears. Every other time, it's the black screen, and I have to go NTSC and back. But it's no problem really, takes 30 seconds.

Sure, I'll do that a little later on. I'm having a bigger problem right now. All my tests were done with 20-minute files, and I didn't notice big problems. However, I started noticing slight hickups every so often with 1-hour files. It may or may not be a coincidence, but they seem to become more recurring as the capture goes along. Also, all my videos are home movies, so there's always a bit of movement involved, and that also may worsen the hickups. This is stricly a PC/Pinnacle problem, no hickups with the DVD rips.
My PC has 8 GB RAM and plenty of free space in the SSD, so I doubt it's the culprit.

Yes the JVC can be more muffled but here with a tape recorded on a different deck it could have been the Panasonic more muffled.
And you offered the most likely explanation": audio/control head azimuth misalignment.

After having adjusted audio azimuth to a particular tape for digitizing we dont need to be constantly realigning azimuth to a factory azimuth tape.
When digitizing, the only audio azimuth "reference" tape that matters is the tape we happen to be digitizing. We dont need a scope for that, just our ears.

Obviously this is only for people who have the skills to make the adjustment safely and correctly.