consider increasing probesize and Encoder did not produce proper pts?

[lixee]

I’ve just set up my dockers and everything seems to work so far, but I’m curious about some error messages I get:

Consuming /consume/Picture_003.jpg ** Processing: /tmp/paperless/paperless-ide2qb2n/convert.png 492x699 pixels, 8 bits/pixel, grayscale Input IDAT size = 63016 bytes Input file size = 63085 bytes

Trying: zc = 9 zm = 9 zs = 0 f = 0 IDAT size = 59897 Selecting parameters: zc = 9 zm = 9 zs = 1 f = 0 IDAT size = 59809

Output file: /tmp/paperless/paperless-ide2qb2n/optipng.png

Output IDAT size = 59809 bytes (3207 bytes decrease) Output file size = 59866 bytes (3219 bytes = 5.10% decrease)

Processing sheet #1: /tmp/paperless/paperless-ide2qb2n/convert-0000.pnm -> /tmp/paperless/paperless-ide2qb2n/convert-0000.unpaper.pnm [pgm_pipe @ 0x55b2bce30cc0] Stream #0: not enough frames to estimate rate; consider increasing probesize [image2 @ 0x55b2bce32100] Using AVStream.codec to pass codec parameters to muxers is deprecated, use AVStream.codecpar instead. [image2 @ 0x55b2bce32100] Encoder did not produce proper pts, making some up. OCRing the document Parsing for eng Detected document date 2007-03-12T00:00:00+01:00 based on string 12/03/07

d Document 20070312000000: Picture_003 consumption finished

As I said, in the end it is working, but I’m wondering whether I can tweak some settings, since the log tell me to „consider increasing probesize“, which doesn’t show up for every document I’ve tried so far. But in all cases I’ve got „ Encoder did not produce proper pts, making some up.“.

[tanderson1992]

I'm here because I'm looking for information on the same issues. Apparently the Stream #0 error is from ffmpeg, which either unpaper or paperless calls. The answers I'm finding online are to add something like "-probesize 10M" to the ffmpeg command line. I don't think that's something we can influence directly in paperless, but I'm not certain about it. Still searching for the other errors as well.

EDIT: As far as I can tell all 3 errors are part of the unpaper run according to the issues I've found at https://github.com/unpaper/unpaper/issues?q=is%3Aissue+is%3Aopen+probesize. That is, people running unpaper are getting the same output, meaning it's unlikely paperless has anything to do with the errors.