John,

Here are the results of my working back through GBS console versions in
seach of the origin of the 48000sr problem. Unfortunately, I'm not sure I
came up with much.

The error message (below) shows up in console GBS9 (Windows) and is
duplicated in GBS10 (and run aborts, back to command line).

In GBS6-8 the situation with respect to which values of ksmps work and which
don't at sr=48000, is exactly the same. The only *difference* is that in all
realtime runs, as soon as the performance starts, there is an immediate jump
back to the command line, and no messages are printed (quite unlike
GBS9-10). So, *no info*. OTOH the expected ksmps values that work under
GBS9-10 also run/sound correctly here, and the ones that don't under GBS9-10
also don't under GBS6-8.

Under flCsoundGBS6 everything works fine.

Under flCsound 0.3 everything works fine.

There is no Windows console .exe earlier than GBS6 to test (at Sourceforge).

Again, I've the feeling that this leaves little to go on, but perhaps it's
useful info for developers. (I hope so.)

As always, I'm always happy to test anything.

Art Hunkins

----- Original Message -----
From: "Art Hunkins" 
To: 
Sent: Sunday, March 20, 2005 12:55 PM
Subject: Re: [Csnd] GBS 48000sr Problem


> John,
>
> Here is the actual error message generated (sorry I didn't quote it
> correctly earlier):
>
> orch compiler:
> 25 lines read
> sr = 48000.000000, kr = 480.000000, ksmps = 100.000000
> error:  inconsistent sr, kr, ksmps
> 1 syntax errors in orchestra.  compilation invalid
> inactive allocs returned to freespace
> end of score.              overall amps:      0.0
>            overall samples out of range:        0
> 0 errors in performance
>
> This is using the same test .csd I sent earlier; it does *not* specify kr
> (csound calculates it itself).
>
> I will methodically go back and find the latest GBS that does *not*
> demonstrate this error, and report back to you.
>
> Art Hunkins
> ----- Original Message -----
> From: "John D. Ramsdell" 
> To: 
> Cc: "Art Hunkins" 
> Sent: Sunday, March 20, 2005 9:03 AM
> Subject: Re: [Csnd] GBS 48000sr Problem
>
>
> > "Art Hunkins"  writes:
> >
> > > I've a bit of new information relative to the console *Windows* GBS10
> > > sr=48000 issue. (Under certain conditions it gives an "illegal sr
rate"
> > > error and stops.)
> >
> > Is I said before, I don't know where to look in the code for this
> > problem.  I couldn't even find the string "illegal sr rate" in the
> > sources or the strings files.  It's particularly troubling that it
> > works with flCsound but not with console csound.  They share the same
> > code used for rendering the input, and were compiled using the same
> > compiler.  At runtime, the csound engine will most likely be loaded
> > into different portions of memory, but why that should matter, I
> > cannot say.
> >
> > At this point, I can only think of one path forward.  I suggest
> > testing old versions of the code, both old GBS versions, and the ones
> > built by John ffitch for Windows from sources in the repository used
> > to build the GBS distribution.
> >
> > The Csound developers have been very good about maintaining a log of
> > changes, I've been tagging whenever I release, so if we know the
> > release at which this stopped working, we should be able to figure out
> > what change caused it.
> >
> > John
> >
> > >
> > > It seems that at certain values for ksmps it runs - at others, it
> doesn't.
> > >
> > > At the following values for ksmps, it runs:
> > > 5, 10, 20, 30, 40, 80, 240, 250, 500, 1000
> > >
> > > At these values, it aborts with the above message:
> > > 29, 31, 35, 50, 81, 85, 90, 99, 100, 101, 200, 220, 230, 235, 239, 241
> > >
> > > With flCsound, *all* values for ksmps work fine. (Also, in console
GBS,
> no
> > > problems at all with sr=44100.)
> > >
> > > With Victor's CS4 console version (Windows), MyCsound4, all values
work
> > > fine.
> > >
> > > I ran all tests with the attached simple .csd. Please note that in all
> cases
> > > kr was not specified (in the hope possibly of ruling out "rounding"
> > > problems?)
> > >
> > > Art Hunkins
> > > Perhaps someone may get an idea from this info.
> > >
> > > Art Hunkins
> > --
> > Send bugs reports to csound-bugs@cs.bath.ac.uk
> >               (or to http://www.cs.bath.ac.uk/cgi-bin/csound )
> > To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I made the following changes in an attempt to fix Art's 48000sr
problem.  At least it can't hurt, right?

Index: csound/oload.c
===================================================================
RCS file: /cvsroot/csound/csound/csound/oload.c,v
retrieving revision 1.6
retrieving revision 1.7
diff -u -r1.6 -r1.7
--- csound/oload.c	7 Feb 2005 07:18:22 -0000	1.6
+++ csound/oload.c	21 Mar 2005 23:37:08 -0000	1.7
@@ -561,7 +561,7 @@
     if ((nn = init0()) > 0)                             /* run instr 0 inits */
       die(Str(X_828,"header init errors"));
     if ((ensmps != (MYFLT) ksmps) ||
-        (gblspace[0]/gblspace[1] != gblspace[2])) {     /* & chk consistency */
+        (gblspace[0] != gblspace[1]*gblspace[2])) {     /* & chk consistency */
       printf("sr = %f, kr = %f, ksmps = %f\n",
              gblspace[0], gblspace[1], gblspace[2]);
       die(Str(X_903,"inconsistent sr, kr, ksmps"));             /*   one more time   */
Index: csound/otran.c
===================================================================
RCS file: /cvsroot/csound/csound/csound/otran.c,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- csound/otran.c	9 Jul 2004 09:31:50 -0000	1.2
+++ csound/otran.c	21 Mar 2005 23:37:08 -0000	1.3
@@ -548,7 +548,7 @@
     }
                                 /* That deals with missing values */
                                 /* However we do need ksmps to be integer */
-    if (tran_sr / tran_kr != tran_ksmps) {
+    if (tran_sr != tran_kr * tran_ksmps) {
         printf("sr = %f, kr = %f, ksmps = %f\n",
                tran_sr, tran_kr, tran_ksmps);
         synterr(Str(X_903,"inconsistent sr, kr, ksmps"));

"Art Hunkins"  writes:

> John,
> 
> Here are the results of my working back through GBS console versions in
> seach of the origin of the 48000sr problem. Unfortunately, I'm not sure I
> came up with much.
> 
> The error message (below) shows up in console GBS9 (Windows) and is
> duplicated in GBS10 (and run aborts, back to command line).
> 
> In GBS6-8 the situation with respect to which values of ksmps work and which
> don't at sr=48000, is exactly the same. The only *difference* is that in all
> realtime runs, as soon as the performance starts, there is an immediate jump
> back to the command line, and no messages are printed (quite unlike
> GBS9-10). So, *no info*. OTOH the expected ksmps values that work under
> GBS9-10 also run/sound correctly here, and the ones that don't under GBS9-10
> also don't under GBS6-8.
> 
> Under flCsoundGBS6 everything works fine.
> 
> Under flCsound 0.3 everything works fine.
> 
> There is no Windows console .exe earlier than GBS6 to test (at Sourceforge).
> 
> Again, I've the feeling that this leaves little to go on, but perhaps it's
> useful info for developers. (I hope so.)
> 
> As always, I'm always happy to test anything.
> 
> Art Hunkins
> 
> ----- Original Message -----
> From: "Art Hunkins" 
> To: 
> Sent: Sunday, March 20, 2005 12:55 PM
> Subject: Re: [Csnd] GBS 48000sr Problem
> 
> 
> > John,
> >
> > Here is the actual error message generated (sorry I didn't quote it
> > correctly earlier):
> >
> > orch compiler:
> > 25 lines read
> > sr = 48000.000000, kr = 480.000000, ksmps = 100.000000
> > error:  inconsistent sr, kr, ksmps
> > 1 syntax errors in orchestra.  compilation invalid
> > inactive allocs returned to freespace
> > end of score.              overall amps:      0.0
> >            overall samples out of range:        0
> > 0 errors in performance
> >
> > This is using the same test .csd I sent earlier; it does *not* specify kr
> > (csound calculates it itself).
> >
> > I will methodically go back and find the latest GBS that does *not*
> > demonstrate this error, and report back to you.
> >
> > Art Hunkins
> > ----- Original Message -----
> > From: "John D. Ramsdell" 
> > To: 
> > Cc: "Art Hunkins" 
> > Sent: Sunday, March 20, 2005 9:03 AM
> > Subject: Re: [Csnd] GBS 48000sr Problem
> >
> >
> > > "Art Hunkins"  writes:
> > >
> > > > I've a bit of new information relative to the console *Windows* GBS10
> > > > sr=48000 issue. (Under certain conditions it gives an "illegal sr
> rate"
> > > > error and stops.)
> > >
> > > Is I said before, I don't know where to look in the code for this
> > > problem.  I couldn't even find the string "illegal sr rate" in the
> > > sources or the strings files.  It's particularly troubling that it
> > > works with flCsound but not with console csound.  They share the same
> > > code used for rendering the input, and were compiled using the same
> > > compiler.  At runtime, the csound engine will most likely be loaded
> > > into different portions of memory, but why that should matter, I
> > > cannot say.
> > >
> > > At this point, I can only think of one path forward.  I suggest
> > > testing old versions of the code, both old GBS versions, and the ones
> > > built by John ffitch for Windows from sources in the repository used
> > > to build the GBS distribution.
> > >
> > > The Csound developers have been very good about maintaining a log of
> > > changes, I've been tagging whenever I release, so if we know the
> > > release at which this stopped working, we should be able to figure out
> > > what change caused it.
> > >
> > > John
> > >
> > > >
> > > > It seems that at certain values for ksmps it runs - at others, it
> > doesn't.
> > > >
> > > > At the following values for ksmps, it runs:
> > > > 5, 10, 20, 30, 40, 80, 240, 250, 500, 1000
> > > >
> > > > At these values, it aborts with the above message:
> > > > 29, 31, 35, 50, 81, 85, 90, 99, 100, 101, 200, 220, 230, 235, 239, 241
> > > >
> > > > With flCsound, *all* values for ksmps work fine. (Also, in console
> GBS,
> > no
> > > > problems at all with sr=44100.)
> > > >
> > > > With Victor's CS4 console version (Windows), MyCsound4, all values
> work
> > > > fine.
> > > >
> > > > I ran all tests with the attached simple .csd. Please note that in all
> > cases
> > > > kr was not specified (in the hope possibly of ruling out "rounding"
> > > > problems?)
> > > >
> > > > Art Hunkins
> > > > Perhaps someone may get an idea from this info.
> > > >
> > > > Art Hunkins
> > > --
> > > Send bugs reports to csound-bugs@cs.bath.ac.uk
> > >               (or to http://www.cs.bath.ac.uk/cgi-bin/csound )
> > > To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
> >
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net