Re: [Lilypond-auto] Issue 3533 in lilypond: Patch: Quantize a number of measurements to 1/200in for better rasterization

[lilypond]

Updates:
        Labels: -Patch-review Patch-needs_work

Comment #11 on issue 3533 by address@hidden: Patch: Quantize a number of  
measurements to 1/200in for better rasterization
http://code.google.com/p/lilypond/issues/detail?id=3533

> What is the improvement on the regression-test PNG output?

More uniform appearance of the staves.

> But stems, for example, are 1.3 times as wide as a staff-line, so stems  
> do not seem to benefit.

That would indeed quite reduce the benefits of the patch, making for a much  
worse overall tradeoff.  The question is whether this can usefully be  
addressed, and if so, where.

> The patch sets the same staff-line thickness for all staff-sizes up to  
> 25-pt

And that seems excessive.  One should possibly have a degradation of  
quantization: if quantization to 200dpi yields results that are off by more  
than 30%, successively try quantizing to 300dpi and then to 600dpi.

As the patch affects all resolutions but leads to improvements only for a  
common set of low resolutions, the net results should be acceptable as a  
set of styles even if the style elements change in discrete steps.

But if the relation of stem thicknesses to line thicknesses (and other  
elements) get thrown off too much, the cost is too high.  There will  
obviously be a number of other places where thicknesses would benefit from  
some sort of quantization, but if the overall effect which is not  
restricted to low-resolution devices gets disproportionate, it is no longer  
the lesser of two evils.

--
You received this message because this project is configured to send all  
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings