What tape format is this? / Can this be simulated with Avisynth?

AVCHDfreak · 22nd May 2015, 08:34

Hi all,
Attached is a screenshot of a frame from a television serial on DVD where severe dropout has occurred. Can someone identify from the dropout what tape format this is likely to be?
(since the original source had dot crawl, my guess is (analog) betacam using composite video signal)
Also can this artifact be simulated in Avisynth?

Thanks in advance.

ChiDragon · 24th May 2015, 14:38

Now that the attachment has been approved...

Analog Betacam can't have digital corruption like that. (Obviously?) Some screenshots of actual Betacam SP dropouts are shown on this webpage about remastering a BBC series (3/4 down the page, heading "DROPOUTS BY THE THOUSAND"). Analog dropouts show up as missing vertical lines or parts of lines. Without dropout compensation they would be black (I think), but usually the playback deck replaces the missing information by repeating a line from the same field, and that's what the link shows.

Your screenshot is 576x432. Is this actually from a DVD, or an encode? It looks like typical download corruption. Even if it really is the VOB file straight from a disc, it may be a bad rip.

This page about the effort to recover a corrupted SpaceX video goes into some detail about why these errors take the shape they do. To recreate the artifact, I think it would be easier to deliberately zero-out or garble portions of a compressed video stream rather than trying to simulate it in Avisynth.

AVCHDfreak · 26th May 2015, 00:27

Quote:

Originally Posted by ChiDragon

Analog Betacam can't have digital corruption like that. (Obviously?)

I thought that was the dropout compensation kicking in (repeating lines but a more severe case). A friend of mine who worked at the station told me that they were using analog tape most of the time because they couldn't afford to make a move to digital completely yet (around 2000-2001).

Quote:

Originally Posted by ChiDragon

Your screenshot is 576x432. Is this actually from a DVD, or an encode? It looks like typical download corruption. Even if it really is the VOB file straight from a disc, it may be a bad rip.

Yes, the capture is from a DVD and was resized, this exact frame also appears at least on another version of the series that was on VCD (they were both international (United States) versions but the domestic broadcast had a clean frame, so I assume the master tape for the international version might be bad).

This image is also said to be a Betacam dropout: (originally from here)

ChiDragon · 26th May 2015, 21:09

That does look similar to your image, but IMO it's different in some important ways. I'm not sure what's going on in the middle section, but the majority of it where it appears stretched vertically is just a single line that's been repeated many times. Your image has two sections like this, but only the luma is repeated. The chroma suffers from some other corruption there. Other portions look block-based.

The resize and overcompression on both screenshots certainly doesn't help clarify things. And this link is a PNG, what the heck...

foxyshadis · 27th May 2015, 00:36

To me it looks like DV dropout with mediocre error correction, not tape dropout. Notice the strong 8x8 edges to the errors (you can even see one good row of blocks beginning about halfway across inside the corruption, if you squint really hard, which is why there are two separate patterns). In fact it could probably be any 8x8 codec. The horrific quality makes it difficult to analyze any further.

Analog dropouts would look like single-line black or white spots with a long tail. Then again, this kind of error correction could conceivably hide that kind of dropout in a similar way, but there would be subtle differences.

A short snip of actual video would be more useful.

AVCHDfreak · 27th May 2015, 04:10

I have seen other discussions at other forums about the studio saying that they used DV for international distribution at the tiem but specifically which one is still of debate.
Can all 8x8 codecs have dropout compensation that repeats previous good lines? How would one simulate this chroma corruption?

ChiDragon · 27th May 2015, 20:38

Quote:

Originally Posted by foxyshadis

you can even see one good row of blocks beginning about halfway across inside the corruption, if you squint really hard, which is why there are two separate patterns

Aha. Wow, good eyes!

22nd May 2015, 08:34	#1 \| Link
AVCHDfreak Registered User Join Date: Dec 2013 Posts: 30	What tape format is this? / Can this be simulated with Avisynth? Hi all, Attached is a screenshot of a frame from a television serial on DVD where severe dropout has occurred. Can someone identify from the dropout what tape format this is likely to be? (since the original source had dot crawl, my guess is (analog) betacam using composite video signal) Also can this artifact be simulated in Avisynth? Thanks in advance. Attached Images

27th May 2015, 00:36	#5 \| Link
foxyshadis Angel of Night Join Date: Nov 2004 Location: Tangled in the silks Posts: 9,562	To me it looks like DV dropout with mediocre error correction, not tape dropout. Notice the strong 8x8 edges to the errors (you can even see one good row of blocks beginning about halfway across inside the corruption, if you squint really hard, which is why there are two separate patterns). In fact it could probably be any 8x8 codec. The horrific quality makes it difficult to analyze any further. Analog dropouts would look like single-line black or white spots with a long tail. Then again, this kind of error correction could conceivably hide that kind of dropout in a similar way, but there would be subtle differences. A short snip of actual video would be more useful. Last edited by foxyshadis; 27th May 2015 at 00:39.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

24th May 2015, 14:38	#2 \| Link
ChiDragon Registered User Join Date: Sep 2005 Location: Vancouver Posts: 600	Now that the attachment has been approved... Analog Betacam can't have digital corruption like that. (Obviously?) Some screenshots of actual Betacam SP dropouts are shown on this webpage about remastering a BBC series (3/4 down the page, heading "DROPOUTS BY THE THOUSAND"). Analog dropouts show up as missing vertical lines or parts of lines. Without dropout compensation they would be black (I think), but usually the playback deck replaces the missing information by repeating a line from the same field, and that's what the link shows. Your screenshot is 576x432. Is this actually from a DVD, or an encode? It looks like typical download corruption. Even if it really is the VOB file straight from a disc, it may be a bad rip. This page about the effort to recover a corrupted SpaceX video goes into some detail about why these errors take the shape they do. To recreate the artifact, I think it would be easier to deliberately zero-out or garble portions of a compressed video stream rather than trying to simulate it in Avisynth.

26th May 2015, 21:09	#4 \| Link
ChiDragon Registered User Join Date: Sep 2005 Location: Vancouver Posts: 600	That does look similar to your image, but IMO it's different in some important ways. I'm not sure what's going on in the middle section, but the majority of it where it appears stretched vertically is just a single line that's been repeated many times. Your image has two sections like this, but only the luma is repeated. The chroma suffers from some other corruption there. Other portions look block-based. The resize and overcompression on both screenshots certainly doesn't help clarify things. And this link is a PNG, what the heck...

27th May 2015, 04:10	#6 \| Link
AVCHDfreak Registered User Join Date: Dec 2013 Posts: 30	I have seen other discussions at other forums about the studio saying that they used DV for international distribution at the tiem but specifically which one is still of debate. Can all 8x8 codecs have dropout compensation that repeats previous good lines? How would one simulate this chroma corruption?