amend 2-GB workaround in GrayscalePdf to use variable '-quality'
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Deriver |
Confirmed
|
Low
|
Unassigned |
Bug Description
Currently, the first uses the default quality level, and if that produces too large a pdf to provide pdf_comp as input, it remakes all the grayscales at quality 70 (typically cutting the file size by about 4).
That has drawbacks:
- time wasted in redoing all the grayscales when output is too large
- default quality probably higher than necessary for good results (and may therefore take longer than necessary)
- 70 may be lower than is needed to shrink output enough to fit
- hasn't happened yet, but potentially failing on the second pass, if 70 is still too high, and giving up
If we can produce a decent estimate of file size under different quality levels, it should be possible to choose an optimal quality level on the first pass: if the lowest quality that yields good results will produce small enough output, use it; otherwise drop quality just far enough to make the output fit.
Changed in deriver: | |
status: | New → Confirmed |
Changed in deriver: | |
importance: | Undecided → Low |