[FFmpeg-devel] [PATCH] ALS: Solve Issue 1657

[Michael Niedermayer]

On Tue, Jan 05, 2010 at 03:59:41AM +0100, Thilo Borgmann wrote:
> Am 05.01.10 02:43, schrieb Michael Niedermayer:
> > On Tue, Jan 05, 2010 at 02:23:09AM +0100, Thilo Borgmann wrote:
> >> Am 05.01.10 01:43, schrieb Michael Niedermayer:
> >>> On Tue, Jan 05, 2010 at 12:34:53AM +0100, Thilo Borgmann wrote:
> >>>> Am 05.01.10 00:30, schrieb Thilo Borgmann:
> >>>>> Hi,
> >>>>>
> >>>>> issue 1657 seems to be caused by negative indices used in [].
> >>>>> See: http://roundup.ffmpeg.org/roundup/ffmpeg/issue1657
> >>>>>
> >>>>> Using *() resolves this issue.
> >>>>>
> >>>>> Tested with gcc 4.0 on MacOS 10.6. There were other versions/compilers
> >>>>> mentioned in roundup, maybe these could be tested by someone (you)?
> >>>>>
> >>>>> I'm sorry, my svn still seems to be broken and produces unusable patches
> >>>>> (%ld...). Nevertheless I can apply them if the workaround is ok.
> >>>>>
> >>>>
> >>>> Some artifacts left in als_data.h. Ignore the old patch, updated patch
> >>>> attached.
> >>>>
> >>>> -Thilo
> >>>
> >>>>  alsdec.c |    4 ++--
> >>>>  1 file changed, 2 insertions(+), 2 deletions(-)
> >>>> 987821d84540420efa6f2e67be17094074e638f8  als_issue1657.rev1.patch
> >>>> Index: libavcodec/alsdec.c
> >>>> ===================================================================
> >>>> --- libavcodec/alsdec.c	(Revision 21025)
> >>>> +++ libavcodec/alsdec.c	(Arbeitskopie)
> >>>> @@ -%ld,%ld +%ld,%ld @@
> >>>>              y = 1 << 19;
> >>>>  
> >>>>              for (sb = 0; sb < smp; sb++)
> >>>> -                y += MUL64(lpc_cof[sb],raw_samples[smp - (sb + 1)]);
> >>>> +                y += MUL64(lpc_cof[sb], *(raw_samples + smp - (sb + 1)));
> >>>
> >>> patch ok
> >>
> >> Applied.
> >>
> >>>
> >>> independant of this, it could be that if lpc_cof was reversed
> >>>
> >>> for (sb = 0; sb < smp; sb++)
> >>>     y += MUL64(lpc_cof[sb], raw_samples[sb]);
> >>
> >> In the second case the index has to be negative which is not possible
> >> with this approach.
> > 
> > i dont see why
> > also it should be 
> > raw_samples[sb]
> > and
> > raw_samples++
> 
> Hmm. If you could please elaborate a bit more, I will pick this up
> tomorrow and will see if I can make it better.

one way:

-    for (; smp < bd->block_length; smp++) {
-        y = 1 << 19;
-
-        for (sb = 0; sb < opt_order; sb++)
-            y += MUL64(bd->lpc_cof[sb],raw_samples[smp - (sb + 1)]);
-
-        raw_samples[smp] -= y >> 20;
+    for (raw_samples+=smp; raw_samples < raw_samples_end; raw_samples++) {
+        y = 1 << 19;
+
+        for (sb= -opt_order; sb<0; sb++)
+            y += MUL64(lpc_cof[sb],raw_samples[sb]);
+
+        raw_samples[0] -= y >> 20;
    }

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Old school: Use the lowest level language in which you can solve the problem
            conveniently.
New school: Use the highest level language in which the latest supercomputer
            can solve the problem without the user falling asleep waiting.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20100105/1c03909d/attachment.pgp>