[CELT-dev] V0.8.0 Problems

Mike Hooper Hooper mihooper at bellsouth.net
Wed Jul 7 05:39:10 PDT 2010


Jean-Marc,

I commented out the fir function in my dsplib which seems to have fixed the 
conflict with the fir fucntion in plc.

One last issue. The compiler can't seem to find "dsp_fft16x16t.h". Where should 
it be? In the past I have omitted c64_fft from the compile to eliminate this 
problem. Is it needed?

thx
MikeH




________________________________
From: Jean-Marc Valin <jean-marc.valin at usherbrooke.ca>
To: Mike Hooper Hooper <mihooper at bellsouth.net>
Cc: celt-dev at xiph.org
Sent: Tue, July 6, 2010 8:38:05 PM
Subject: Re: [CELT-dev] V0.8.0 Problems

Hi Mike,

On 10-07-06 08:10 PM, Mike Hooper Hooper wrote:
> Now on to a rather simple fix. In vq.c my compiler complains about
> #ifdef FIXED_POINT
> int k = celt_ilog2(E)>>1;
> #endif
>
> being placed after an executable statement. Apparently this is OK for
> C++ but not C. I moved this up just after E=EPSILON and my compiler quit
> complaining.

Just checked in a fix for that.

> Next, there appears to be a problem between my dsplib.h and plc.h and
> plc.c. I get ther following error messages:
>
> In plc.c
>
> "declaration is incompatible with "ushort fir(DATA *, DATA *, DATA *,
> DATA *, ushort, ushort)" (declared at line 50 of "..\libcelt\dsplib.h")
> libcelt80/libcelt plc.c line 97 1278460811593 9315"

I assume dsplib.h also defines a (different) fir() function. Maybe 
there's a way to change the CELT TI headers to not include dsplib.h in 
this particular context?

> all of which seem to be related to an incompatible type.

No, it's just two unrelated functions with the same name.

> ...also, I tried to #undef NEW_PLC in my config.h but it did not exclude
> these files.

NEW_PLC is now the only PLC, so the macro doesn't even exist in 0.8.0.

Cheers,

    Jean-Marc

>
> *
> *
>
> *
> *
>
>
>
> ------------------------------------------------------------------------
> *From:* Jean-Marc Valin <jean-marc.valin at usherbrooke.ca>
> *To:* Mike Hooper Hooper <mihooper at bellsouth.net>
> *Cc:* celt-dev at xiph.org
> *Sent:* Tue, July 6, 2010 7:40:20 PM
> *Subject:* Re: [CELT-dev] V0.8.0 Problems
>
> OK, I think I just fixed half of the issue, but is it possible that you
> were not using the latest git? I committed a fix yesterday that seemed
> like it should have fixed the other half of the issue you're having.
>
> Cheers,
>
> Jean-Marc
>
> On 10-07-06 07:16 PM, Mike Hooper Hooper wrote:
>  > Jean-Marc,
>  > I have a few more that could be similar:
>  > In celt.c, in celt_decode_float, in the return statement
>  > "*return* celt_decode_with_ec_float(st, data, len, pcm, frame_size,
>  > *NULL*);"
>  >
>  > ...my compiler says
>  >
>  > "argument of type "float *" is incompatible with parameter of type
>  > "celt_sig *restrict" libcelt80/libcelt celt.c line 1888 1278457880500
> 9106
>  >
>  > However, in my config.h file I have :
>  >
>  > * *
>  >
>  > *#define* DISABLE_FLOAT_API
>  >
>  > * *
>  >
>  > *#define* FIXED_POINT
>  >
>  > Did I miss a #define somewhere?
>  >
>  > thx
>  >
>  > Mike
>  >
>  >
>  >
>  >
>  > ------------------------------------------------------------------------
>  > *From:* Jean-Marc Valin <jean-marc.valin at usherbrooke.ca
> <mailto:jean-marc.valin at usherbrooke.ca>>
>  > *To:* Mike Hooper Hooper <mihooper at bellsouth.net
> <mailto:mihooper at bellsouth.net>>
>  > *Cc:* celt-dev at xiph.org <mailto:celt-dev at xiph.org>
>  > *Sent:* Tue, July 6, 2010 7:04:42 PM
>  > *Subject:* Re: [CELT-dev] V0.8.0 Problems
>  >
>  > Hi Mike,
>  >
>  > Thanks again on that other bug. It turns out that on most platforms
>  > "celt_int32" is the same as "int" so the compiler doesn't complain.
>  > Unfortunately, I haven't found any way to test for these kinds of issues
>  > that happen only on 16-bit platforms and I've had to rely on people like
>  > you compiling on such platforms. Let me know if you find anything else.
>  > This one is now fixed in git.
>  >
>  > Thanks,
>  >
>  > Jean-Marc
>  >
>  > On 10-07-06 06:22 PM, Mike Hooper Hooper wrote:
>  > > Jean-Marc,
>  > >
>  > > I wish I could take credit for finding the bug, but I'll have to give
>  > > credit to my compiler....:)
>  > >
>  > > Let's now take a look at this code in bands.c
>  > >
>  > > " quant_band(encode, m, i, X, Y, N, b, spread, tf_change, lowband,
>  > > resynth, ec, &remaining_bits, LM, norm+M*eBands[i], bandE, 0);"
>  > >
>  > > My compiler says:
>  > >
>  > > "argument of type "int *" is incompatible with parameter of type
>  > > "celt_int32 *" libcelt80/libcelt bands.c line 951 1278452093812 9084"
>  > >
>  > > I'm not exactly sure where the problem is. Can you see it?
>  > >
>  > > MikeH
>  > >
>  > > -----Original Message-----
>  > >
>  > > From: Jean-Marc Valin [mailto:jean-marc.valin at usherbrooke.ca
> <mailto:jean-marc.valin at usherbrooke.ca>
>  > <mailto:jean-marc.valin at usherbrooke.ca
> <mailto:jean-marc.valin at usherbrooke.ca>>]
>  > >
>  > > Sent: Tuesday, July 06, 2010 6:00 PM
>  > >
>  > > To: Mike Hooper Hooper
>  > >
>  > > Cc: celt-dev at xiph.org <mailto:celt-dev at xiph.org>
> <mailto:celt-dev at xiph.org <mailto:celt-dev at xiph.org>>
>  > >
>  > > Subject: Re: [CELT-dev] V0.8.0 Problems
>  > >
>  > > Hi Mike,
>  > >
>  > > Indeed, you found a bug! I missed it because the symptoms you saw only
>  > >
>  > > occurs in fixed-point with compilers that don't have var arrays or
>  > >
>  > > alloca(). It's fixed in git now. For now (commit b6a5f9), the git
> master
>  > >
>  > > branch is bitstream compatible with 0.8.0. I'll probably issue a 0.8.1
>  > >
>  > > bugfix release after we fix any remaining problems you have.
>  > >
>  > > Cheers,
>  > >
>  > > Jean-Marc
>  > >
>  > > On 10-07-06 04:52 PM, Mike Hooper Hooper wrote:
>  > >
>  > > > Tim, et al,
>  > >
>  > > > I have run into several problems with V0.8.0. I will address them
>  > >
>  > > > seperately.
>  > >
>  > > > 1. My compiler is complaining about the following code in celt.c
> which
>  > >
>  > > > seems to define metric first as celt_word32, then as celt_word16.
> Am I
>  > >
>  > > > mis-interpreting something?
>  > >
>  > > > VARDECL(celt_word32, metric);
>  > >
>  > > > ALLOC(metric, len, celt_word16);
>  > >
>  > > > Thx
>  > >
>  > > > MikeH
>  > >
>  > > >
>  > >
>  > > >
>  > >
>  > > >
>  > >
>  > > > _______________________________________________
>  > >
>  > > > celt-dev mailing list
>  > >
>  > > > celt-dev at xiph.org <mailto:celt-dev at xiph.org>
> <mailto:celt-dev at xiph.org <mailto:celt-dev at xiph.org>>
>  > >
>  > > > http://lists.xiph.org/mailman/listinfo/celt-dev
>  > >
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.xiph.org/pipermail/opus/attachments/20100707/10e65ad5/attachment-0002.htm 


More information about the celt-dev mailing list