Craig Milo Rogers | 21 Nov 06:45 2009
Picon

Possible Test Error

./callback_test

    Zero Order Hold interpolator :
        callback_test    (SRC ratio = 1.0000) ........... ok
        callback_test    (SRC ratio = 0.0990) ........... ok
        callback_test    (SRC ratio = 0.1000) ........... ok
        callback_test    (SRC ratio = 0.3333) ........... ok
        callback_test    (SRC ratio = 0.7890) ........... ok
        callback_test    (SRC ratio = 1.0001) ........... ok
        callback_test    (SRC ratio = 1.9000) ........... ok
        callback_test    (SRC ratio = 3.1000) ........... ok
        callback_test    (SRC ratio = 9.9000) ........... ok
    Linear interpolator :
        callback_test    (SRC ratio = 1.0000) ........... ok
        callback_test    (SRC ratio = 0.0990) ........... ok
        callback_test    (SRC ratio = 0.1000) ........... ok
        callback_test    (SRC ratio = 0.3333) ........... ok
        callback_test    (SRC ratio = 0.7890) ........... ok
        callback_test    (SRC ratio = 1.0001) ........... ok
        callback_test    (SRC ratio = 1.9000) ........... ok
        callback_test    (SRC ratio = 3.1000) ........... ok
        callback_test    (SRC ratio = 9.9000) ........... ok
    Sinc interpolator :
        callback_test    (SRC ratio = 1.0000) ........... ok
        callback_test    (SRC ratio = 0.0990) ........... ok
        callback_test    (SRC ratio = 0.1000) ........... ok
        callback_test    (SRC ratio = 0.3333) ........... ok
        callback_test    (SRC ratio = 0.7890) ........... 

Line 141 : input / output length mismatch.
(Continue reading)

Erik de Castro Lopo | 21 Nov 08:25 2009

Re: Possible Test Error

Craig Milo Rogers wrote:

> 	Is this the correct fix, or is a fix even needed at all
> (perhaps my system is at fault)?

The fix is fine. This one as caused by different CPUs (ie i686 vs x86_64)
having slightly different rounding errors.

Cheers,
Erik
--

-- 
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/
John Emmas | 30 Nov 20:20 2009
Picon

Building libsamplerate on WIN32

Hi, I just downloaded the source for SRC and tried to build it using
MSVC++6.0 (following the instructions that I found here)

http://www.mega-nerd.com/SRC/win32.html

I thought the following observations might be helpful:-

1)  Firstly, in line 4 of the instructions, the lib and dll seem to be
called libsndfile-1 (not libsndfile, as stated).  Also, it seems to be
necessary to include libsndfile-1.def in addition to the two files
specified.

2)  SRC's config.h looks as if it was produced by a differ.  At the bottom,
there's a typical 'diff' line saying something like "<<<<<< diff from left
file".  Sorry if that's not 100% accurate.  I'd already corrected the
problem by the time I posted this.  Anyway, the diff line obvious causes the
compilation to fail.

3)  After I corrected the above, the compilation appeared to be progressing
nicely until it came to src_sinc.c.  At that stage I get this error:-

.\src\high_qual_coeffs.h(327683) : fatal error C1076: compiler limit :
internal heap limit reached; use /Zm to specifiy a higher limit.

How can I find out what the current limit is and/or what should I increase
it to?

Thanks,

John
(Continue reading)

Erik de Castro Lopo | 1 Dec 07:56 2009

Re: Building libsamplerate on WIN32

Hi John,

John Emmas wrote:

> 1)  Firstly, in line 4 of the instructions, the lib and dll seem to be
> called libsndfile-1 (not libsndfile, as stated).  Also, it seems to be
> necessary to include libsndfile-1.def in addition to the two files
> specified.

Thanks I've cleaned that up.

> 2)  SRC's config.h looks as if it was produced by a differ.

Yeah, I've been meaning to release a new fixed version for a while.

> 3)  After I corrected the above, the compilation appeared to be progressing
> nicely until it came to src_sinc.c.  At that stage I get this error:-
> 
> .\src\high_qual_coeffs.h(327683) : fatal error C1076: compiler limit :
> internal heap limit reached; use /Zm to specifiy a higher limit.
> 
> How can I find out what the current limit is and/or what should I increase
> it to?

Well, I'm not a windows guy so I have no clue. I suggest that you start with
a number which is a power of two and if its not high enough, double it until
it is high enough.

Cheers,
Erik
(Continue reading)

John Emmas | 1 Dec 08:52 2009
Picon

Re: Building libsamplerate on WIN32

----- Original Message ----- 
From: "Erik de Castro Lopo" <erikd@...>
>
> John Emmas wrote:
>
>> 3)  After I corrected the above, the compilation appeared to be
>> progressing
>> nicely until it came to src_sinc.c.  At that stage I get this error:-
>>
>> .\src\high_qual_coeffs.h(327683) : fatal error C1076: compiler limit :
>> internal heap limit reached; use /Zm to specifiy a higher limit.
>>
>> How can I find out what the current limit is and/or what should I
>> increase it to?
>
> Well, I'm not a windows guy so I have no clue. I suggest that you start
> with a number which is a power of two and if its not high enough, double
> it until it is high enough.
>
Many thanks Erik,

I found this article at MSDN:-

http://msdn.microsoft.com/en-us/library/aa278580(VS.60).aspx

Evidently, the heap limit is a limit as used by the compiler (i.e. not a
limit for the program being compiled, as I'd assumed).  The limit is to keep
the compiler from using too much memory and Microsoft's "standard" limit is
just 105MB (this being equivalent to /Zm100).  I added the option /Zm200
to your 'Makefile.msvc'.  After doing that, your library then built
(Continue reading)


Gmane