Proc amp? Fix levels at input to Canpus ADVC110

[2Bdecided]

I'm capturing PAL S-VHS.

The levels are as follows:

sync voltage = 258mV (should be 300mV)
peak white = 702mV (OK)

The Canopus ADVC110 would be perfectly happy with peak white at 700mV if the sync pulses were the correct amplitude.

Unfortunately, it sees the sync pulses are low, assumes the entire video needs amplifying, and in doing so clips the peak whites.

I guess I can either restore the sync pulse voltage, or attenuate the video signal.

Any suggestions for how I might achieve either on an S-video connection?

Cheers,
David.

[2Bdecided]

No one?

[2Bdecided]

I tried buffering the video through a Panasonic DVD recorder, hoping it would fix the sync levels.

It did, but it AGC'd the video up to match, keeping it "too bright".

Can no one suggest a proc amp that can help with this situation?

Cheers,
David.

[2Bdecided]

So no one here uses a proc amp?

[filmboss80]

You didn't say what your final captured file is. If it is DV-AVI, you can use a software pro amp. On occasion (usually after the footage is captured to harddrive) I have used JohnnyMalaria's nice little Enosoft DV processor (with waveform monitor and vectorscope in the program) to correct a few errant DV files. There is a free version of it. If you use it, I advise that you go under the Configurations and deselect "Disable seeking/navigation." (I'm not sure what that is, but it sometimes gives me edge artifacts when it is selected.) One other thing: be patient in waiting for responses to your inquiry. Some of us live in different time zones, and are not able to monitor this site 24/7.

[schematic2]

AVT-8710 is a low priced unit that may do the job....

[2Bdecided]

Thank you schematic2 - I'll see what the local price is (already got a TBC in the VCR though - hope they play nicely together).

filmboss80 - it's clipping before A/D - no processing in the digital domain will fix that. I think waiting a month for a reply was hardly impatient, whatever time zone you're reading in! I know it's bad form to reply to your own posts, but I know people here use these things, and assumed that thsoe that do hadn't happened to spot this thread last time around.

Cheers,
David.

[edDV]

Do you get the same problem on a different S-VHS player? Sync pulse height issues are usually a recording source issue. A Pro S-VHS deck will have an internal playback TBC and Proc amp. If this is a single tape issue, you can save money paying a dub house to handle the capture to DV format.

If this is a volume issue, you may need to trade up to an ADVC-300 (internal line TBC+ProcAmp) or higher. It may give higher quality than an upstream A/D D/A ProcAmp.

[2Bdecided]

It's the tapes themselves. They're the same on two S-VHS machines I've tried, they're all the same (all 100 hours from that camcorder, whether originals or dubs!), and both machines play other tapes back correctly.

You can't fully disable the filtering in the ADVC300, which put me off.

I'm hoping to find something quite simple to solve the problem - both to save money, and to avoid adding more problems - the full works including A/D and D/A seems way over the top for a simple analogue level adjustment.

Cheers,
David.

[edDV]

It's the tapes themselves. They're the same on two S-VHS machines I've tried, they're all the same (all 100 hours from that camcorder, whether originals or dubs!), and both machines play other tapes back correctly.

You can't fully disable the filtering in the ADVC300, which put me off.

I'm hoping to find something quite simple to solve the problem - both to save money, and to avoid adding more problems - the full works including A/D and D/A seems way over the top for a simple analogue level adjustment.

Cheers,
David.

You can adjust levels in software (even in real time with Enosoft) but if the short sync pulse is causing white clipping in the ADVC, you can't correct that in software.

Can you post a frame cap showing the problem? VLC screencap works well for this. I can load the frame into the Vegas scopes for analysis. This is an ADVC-100 cap from a normal VCR.

[edDV]

PS: Something isn't adding up here because most video equipment references the sync back porch rather than sync tip. This prevents influence from sync amplitude. Sync pulses often get replaced in the broadcast chain.

[2Bdecided]

edDV,

Thanks for the kind offer to take a look in Vegas. I value a second opinion!

Here are a couple of frames:
http://www.david.robinson.org/video/255clip.avi

I've used the following AVIsynth script to look at the levels, and try to put them where I want (still with clipping, obviously!). It interleaves original and "fixed" for comparison.

avisource("255clip.avi",pixel_type="yv12") # Cedocida DV decoder

b=last.coloryuv(gain_y=-25)

interleave(last,b)

histogram(mode="levels")

converttorgb(matrix="Rec601")

I think I see clipping at 255 in YUV. It's mild, but you can see (both from the hystogram, and from comparing the analogue signal with the digitised version) that there's some clipping.

My understanding is that most video recording devices apply an automatic gain control circuit to the incoming signal to restore the sync amplitude to 0.3V, hence usually ensuring the rest of the signal is in-range too. The back porch is used for black (or in NTSC, blank) level clamping.

Cheers,
David.

[2Bdecided]

Any thoughts?

[edDV]

Sorry for missing this. It has been a busy week.

Here is the direct analysis of your 255clip.avi clip. It appears to follow normal 16-255 consumer DV camcorder template but whites appear to have clipped at some point and black level is elevated. This may have been camcorder exposure or post processing.

[edDV]

Here I used Vegas levels filters to linearly pull whites down to 235 and place blacks around 16. This would need to be viewed on a digital TV or 0-100IRE analog to match levels. Computer display is 0-255.

[2Bdecided]

Thank you - those plots are very useful.

Cheers,
David.

[2Bdecided]

I borrowed a simple video amplifier with a straight forward signal gain adjustment. I verified on a scope that it was doing exactly what it said on the box: adjusting the gain of the entire video signal.

It has no useful effect when capturing into the ADVC110. You can see a transitory effect if you change the gain rapidly, but the ADVC110 quickly adapts to counteract the change, giving an identical result over a very wide range of gain settings.

I'm fairly sure the ADVC110 is measuring the sync pulse amplitude, and AGC-ing the entire signal to bring it back to where it thinks it should be.


edDV spotted an important fact though - the blacks are elevated too, and simply bringing the whole lot down linearly works quite well. This suggests a "brightness" control (as usually implemented) rather than a gain/contrast control, might do a reasonable job - I think I know where to find one, and will report back if it works. However, looking at how rare the digital clipping is, and the fact that a simple linear shift won't introduce contouring, I might be happy doing it in the digital domain post-capture, just as edDV did - I need to try a few more tapes and find worst case examples.

Thanks for your help edDV.

Cheers,
David.