[argyllcms] Re: dispread issues, in an android tablet chroot

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Mon, 03 Sep 2012 09:47:44 +1000

Dan Wilcox wrote:
> OK so after some fussing with gdb last night I tracked it down to /377 aka
> 255 is provided when a file pointer finished in at least this arm
> environment.  So in read_line in pars.c I changed the condition with the
> getch call to check for 255 or -1, as a nasty workaround.

        that looks like a bug. Try changing the "char c;" to "int c;"
declaration at the top of read_line().

> Now I can run dispread and generate a ti3 file but when I use colprof on
> it, I get some very large error margins.  I have turned off the dpms stuff
> the screen would normally do, so it sits at full brightness for the entire
> patch reading but I'm getting about 14 average error.  It is only a medium
> quality shaper/matrix so maybe this is expected?

The .ti3 numbers seem wacky. For instance:

192 8.0770 16.029 64.107 9.9789 2.4932 49.271

A Z of 49 with a Y of 3 ?

Patch 186 has a *a value of -180 !

(ie. try running profcheck -v2 on the resulting profile to locate the outliers).

Try creating a .ti1 with (say) a white patch followed by several of
patch 186 value, and then check the resulting readings for consistency.
(ie. is there a display update delay problem).

Graeme Gill.

Other related posts: