Banding

[Phoenix]

I'm using Edius and have an irreplaceable talking head interview in DV in which banding has occurred right at a crucial spot (at the punchline of a joke delivered by a 93 year old) for about 5 seconds. There are about 12 quite wide horizontal bands, in half of which the video continues unaffected, whilst the alternate bands have freeze framed on the last 'good' frame ie last frame before the affected 5 seconds is frozen in alternate bands. The last second or so of the damaged segment, the frozen bands start to pixelate as the video recovers in those bands and then continues unaffected.

Anyone know if there any software out there that can interpolate to replace the frozen bands based on the facial info preceding and following the banding or is this mission impossible. I was thinking perhaps 'deshaker' might work - but probably not enough information there ... maybe some sort of rotoscoping software perhaps.

[PaulTV]

Can your DV camera/deck play at 10th normal speed, if so try it, usually the errored material will play - if successful, digitize this way and speed up damaged segment to normal speed. You may well have to capture this segment manualy using S Video rather than firewire, can't remember if firewire works or not at strange deck speeds, but I suspect no.

This is a long shot but it has worked for me in the past.

Paul

[Phoenix]

Thanks for the tip - I did try playing at 1/10 speed in the deck and it appeared to work.
I had tried capturing from both the camera and DSR-11 with the same banding result in the same place on the tape - so tape was damaged rather than dirty heads etc. Before trying to capture at 1/10 speed, though I ran a head cleaning tape for 10 sec in the deck then tried capture immediately afterward and the banding went from intermittently better to much worse (solid banding). Managed to get a near perfect capture with only a minor glitch after another head clean. Maybe the cleaning lubricant residues on the heads help stabilise the area of tape damage when played immediately afterward ? Phew !

I never re-use tapes and it was only on this rare occasion when I was caught short that I did so ........ never again.

- anyway thanks again for the advice Paul, I'm sure that method will come in handy oneday soon.

[Khoi Pham]

I'm using Edius and have an irreplaceable talking head interview in DV in which banding has occurred right at a crucial spot (at the punchline of a joke delivered by a 93 year old) for about 5 seconds. There are about 12 quite wide horizontal bands, in half of which the video continues unaffected, whilst the alternate bands have freeze framed on the last 'good' frame ie last frame before the affected 5 seconds is frozen in alternate bands. The last second or so of the damaged segment, the frozen bands start to pixelate as the video recovers in those bands and then continues unaffected.

Anyone know if there any software out there that can interpolate to replace the frozen bands based on the facial info preceding and following the banding or is this mission impossible. I was thinking perhaps 'deshaker' might work - but probably not enough information there ... maybe some sort of rotoscoping software perhaps.

If it is a drop out, then you are sol, if the banding was from mistracking, the DSR1800 or DSR2500 I believe has variable tracking head that could fix it.

[robsegaar]

Hi Phoenix,

If there ar grey stipes horizontal then is it a proberly a recording head lineup error.
But if you have a panasonic DVCPRO/DVCAM deck then it has standard errorcorrection and this solve these problems.