Nabble - Mad - Dev

playback/decoding problem

2008-12-02T00:05:21Z

Hello,

i have a problem with one stream that is not correct decoded by libmad. I use libmad-0.15.1b-r5 and mplayer with the following command line:

mplayer -ac mad http://gffstream.ic.llnwd.net/stream/gffstream_w19a

This is a mp3 radio stream and i get ugly noise and spikes. I can try another stream and it works very well. Can anyone tell me why this stream is not correct decoded. I have tried some other decoders
but its always the same problem.

Thanks,

Mathias

---

OUTPUT:

mplayer -ac mad http://gffstream.ic.llnwd.net/stream/gffstream_w19a

Connecting to server gffstream.ic.llnwd.net[87.248.219.36]: 80...
Cache size set to 320 KBytes
Cache fill: 15.00% (49152 bytes)
Audio only file format detected.
==========================================================================
Forced audio codec: mad
Opening audio decoder: [libmad] libmad mpeg audio decoder
AUDIO: 44100 Hz, 2 ch, s16le, 128.0 kbit/9.07% (ratio: 16000->176400)
Selected audio codec: [mad] afm: libmad (libMAD MPEG layer 1-2-3)
==========================================================================
AO: [alsa] 48000Hz 2ch s16le (2 bytes per sample)
Video: no video
Starting playback...

---

Playback issues

2008-10-26T13:00:03Z

Hi, I am trying to make a sound library for the Wii that uses libmad for
the mp3 decoding. However, I seem to have a problem, my code reports a
lot of lost sync errors when decoding the frames. Could someone please
take a look at my code and see if I am doing something obviously wrong?

You can find my code at svn://henke37.cjb.net/sound/ , the main file is
newMP3Player.cpp .

Madplay improved patch for alsa audio

2008-10-19T05:25:46Z

Hi,

Attached is an improved patch for alsa backend of madplay.
* fix allocation of buffer, buffer_size is number of frames
* use default 16 bit depth (some hardware has small buffers, 16/24 bit
difference is very small)

Regards,

Micha

--- madplay-0.15.2b/audio_alsa.c 2008-10-18 15:10:16.000000000 +0200
+++ madplay-0.15.2b/audio_alsa.c.new 2008-10-18 15:03:27.000000000 +0200
@@ -28,31 +28,30 @@

#include <errno.h>

-#define ALSA_PCM_OLD_HW_PARAMS_API
-#define ALSA_PCM_OLD_SW_PARAMS_API
#include <alsa/asoundlib.h>

#include <mad.h>

#include "audio.h"

-char *buf = NULL;
-int paused = 0;
+#define BUFFER_TIME_MAX 500000

-int rate = -1;
-int channels = -1;
-int bitdepth = -1;
-int sample_size = -1;
-
-int buffer_time = 500000;
-int period_time = 100000;
-char *defaultdev = "plughw:0,0";
+unsigned char *buf = NULL;
+int paused = 0;
+
+unsigned int rate = 0;
+unsigned int channels = -1;
+unsigned int bitdepth = -1;
+unsigned int sample_size = -1;
+
+unsigned int buffer_time;
+unsigned int period_time;
+char *defaultdev = "plughw:0,0";

snd_pcm_hw_params_t *alsa_hwparams;
snd_pcm_sw_params_t *alsa_swparams;

-snd_pcm_sframes_t buffer_size;
-snd_pcm_sframes_t period_size;
+snd_pcm_uframes_t buffer_size;

snd_pcm_format_t alsa_format = -1;
snd_pcm_access_t alsa_access = SND_PCM_ACCESS_MMAP_INTERLEAVED;
@@ -66,14 +65,20 @@
snd_pcm_hw_params_t *params,
snd_pcm_access_t access)
{
- int err, dir;
-
+ int err;
+
/* choose all parameters */
err = snd_pcm_hw_params_any(handle,params);
if (err < 0) {
printf("Access type not available for playback: %s\n", snd_strerror(err));
return err;
}
+ /* set the access type */
+ err = snd_pcm_hw_params_set_access(handle, params, alsa_access);
+ if (err < 0) {
+ printf("Sample format not available for playback: %s\n", snd_strerror(err));
+ return err;
+ }
/* set the sample format */
err = snd_pcm_hw_params_set_format(handle, params, alsa_format);
if (err < 0) {
@@ -87,29 +92,38 @@
return err;
}
/* set the stream rate */
- err = snd_pcm_hw_params_set_rate_near(handle, params, rate, 0);
+ err = snd_pcm_hw_params_set_rate(handle, params, rate, 0);
if (err < 0) {
printf("Rate %iHz not available for playback: %s\n", rate, snd_strerror(err));
return err;
}
- if (err != rate) {
- printf("Rate doesn't match (requested %iHz, get %iHz)\n", rate, err);
- return -EINVAL;
- }
+ err = snd_pcm_hw_params_get_buffer_time_max(params, &buffer_time, NULL);
+ if (err < 0) {
+ printf("Unable to retrieve buffer time: %s\n", snd_strerror(err));
+ return err;
+ }
+ if (buffer_time > BUFFER_TIME_MAX)
+ buffer_time = BUFFER_TIME_MAX;
/* set buffer time */
- err = snd_pcm_hw_params_set_buffer_time_near(handle, params, buffer_time, &dir);
+ err = snd_pcm_hw_params_set_buffer_time_near(handle, params, &buffer_time, 0);
if (err < 0) {
printf("Unable to set buffer time %i for playback: %s\n", buffer_time, snd_strerror(err));
return err;
}
- buffer_size = snd_pcm_hw_params_get_buffer_size(params);
+ if (period_time * 4 > buffer_time)
+ period_time = buffer_time / 4;
/* set period time */
- err = snd_pcm_hw_params_set_period_time_near(handle, params, period_time, &dir);
+ err = snd_pcm_hw_params_set_period_time_near(handle, params, &period_time, NULL);
if (err < 0) {
printf("Unable to set period time %i for playback: %s\n", period_time, snd_strerror(err));
return err;
}
- period_size = snd_pcm_hw_params_get_period_size(params, &dir);
+ /* retrieve buffer size */
+ err = snd_pcm_hw_params_get_buffer_size(params, &buffer_size);
+ if (err < 0) {
+ printf("Unable to retrieve buffer size: %s\n", snd_strerror(err));
+ return err;
+ }
/* write the parameters to device */
err = snd_pcm_hw_params(handle, params);
if (err < 0) {
@@ -123,6 +137,7 @@
int set_swparams(snd_pcm_t *handle,
snd_pcm_sw_params_t *params)
{
+ unsigned int start_threshold;
int err;

/* get current swparams */
@@ -136,13 +151,7 @@
if (err < 0) {
printf("Unable to set start threshold mode for playback: %s\n", snd_strerror(err));
return err;
- }
- /* allow transfer when at least period_size samples can be processed */
- err = snd_pcm_sw_params_set_avail_min(handle, params, period_size);
- if (err < 0) {
- printf("Unable to set avail min for playback: %s\n", snd_strerror(err));
- return err;
- }
+ }
/* align all transfers to 1 samples */
err = snd_pcm_sw_params_set_xfer_align(handle, params, 1);
if (err < 0) {
@@ -190,7 +199,7 @@
rate = config->speed;

if ( bitdepth == 0 )
- config->precision = bitdepth = 32;
+ config->precision = bitdepth = 16;

switch (bitdepth)
{
@@ -241,7 +250,7 @@
return -1;
}

- buf = malloc(buffer_size);
+ buf = malloc(buffer_size * sample_size);
if (buf == NULL) {
audio_error="unable to allocate output buffer table";
return -1;
@@ -279,7 +288,7 @@
int play(struct audio_play *play)
{
int err, len;
- char *ptr;
+ unsigned char *ptr;

ptr = buf;
len = play->nsamples;

libmad-0.15.1b does not compile on Fedora 9

2008-10-11T08:27:11Z

Hi,

I've just installed a clean Fedora 9 distro, and I tried to compile
libmad-0.15.1b (which I had done fine previously, under Fedora 5),
I got the following error :

gcc -DHAVE_CONFIG_H -I. -I. -I. -DFPM_INTEL -DASO_ZEROCHECK -Wall
-march=i486 -g -O -fforce-mem -fforce-addr -fthread-jumps
-fcse-follow-jumps -fcse-skip-blocks -fexpensive-optimizations -fregmove
-fschedule-insns2 -fstrength-reduce -MT version.lo -MD -MP -MF
.deps/version.Tpo -c version.c -fPIC -DPIC -o .libs/version.o
cc1: error: unrecognized command line option "-fforce-mem"
make[2]: *** [version.lo] Error 1

gcc version 4.3.0 does not have the "-fforce-mem" option... what can I do ?
is there a way (a configure option) to disable this ?

Thx,
Joao

--
Joao Moreira de Sa Coutinho
joao.moreira@...
Tel.: 08.70.23.48.29
Mob.: 06.28.90.83.67

Re: madplay: How to select a specific ALSA card

2008-09-19T01:20:27Z

joker1999 wrote:

Thomas Reitmayr-2 wrote:

Hi,
I just started using madplay a short while ago and I was wondering
whether it is possible to select an ALSA card different from the default
one for audio output.
>From looking at the code it seems like the PATH part of --output is
used, however
madplay --output=alsa:plughw:1 myfile.mp3
chokes as the TYPE alsa does not exist (in audio.c).

Did I miss something in the command line or is this a bug / missing
feature (for which I could provide a patch)?

Thanks,
-Thomas

You can try:
madplay -o wave:- something.mp3 | aplay

Regards

You may also need
mount /dev/sda1 -o remount,async,noatime
on nslu2 to get better usb transfer, so you won't experience hangs during playback.

Regards

Re: madplay: How to select a specific ALSA card

2008-09-14T13:35:47Z

Thomas Reitmayr-2 wrote:

Hi,
I just started using madplay a short while ago and I was wondering
whether it is possible to select an ALSA card different from the default
one for audio output.
>From looking at the code it seems like the PATH part of --output is
used, however
madplay --output=alsa:plughw:1 myfile.mp3
chokes as the TYPE alsa does not exist (in audio.c).

Did I miss something in the command line or is this a bug / missing
feature (for which I could provide a patch)?

Thanks,
-Thomas

You can try:
madplay -o wave:- something.mp3 | aplay

Regards

Re: Converting fixed point numbers

2008-07-12T13:07:04Z

On Jul 12, 2008, at 10:56 AM, Reiner wrote:

> madplay uses the following in audio_linear_round
>
> /* quantize and scale */
> return sample >> (MAD_F_FRACBITS + 1 - bits);
>
> if bits is 16 (unsigned short e,g) we will shift 13 bits, whats
> about the 3 MSB bits ???
>
> With this I think only the first digit of the whole part is taken,
> what is with the sign and the other 2 digits from the whole part,
> are they never used?

The whole part is used to represent samples that are above full scale
[-1.0, 1.0) but madplay clips these prior to scaling, so at this point
the whole part just contains copies of the sign bit.

--
Rob Leslie
rob@...

Converting fixed point numbers

2008-07-12T10:56:50Z

Hello,

madplay uses the following in audio_linear_round

/* quantize and scale */
return sample >> (MAD_F_FRACBITS + 1 - bits);

if bits is 16 (unsigned short e,g) we will shift 13 bits, whats about
the 3 MSB bits ???

With this I think only the first digit of the whole part is taken, what
is with the sign and the other 2 digits from the whole part, are they
never used?

Thanks

Reiner

libmad documentation project

2008-06-09T07:46:21Z

Hi,

I recently started working on a small audio processing project. I chose
to use mad library for dealing with mp3 files.

Since I needed to study it for getting enough understanding of it, and
the package TODO file calls for contributors to documentation project, I
tried to make my notes useful for the programmers community.

One thing I like in java programming is the documentation facility,
which generates browseable API documentation directly from comments
inserted in source files.

The same functionality is available for C/C++ and other languages, with
Doxygen (open-source) program, so I decided to use it.

You'll find all details and files here:
http://m.baert.free.fr/contrib/docs/libmad/

Your comments and suggestions are welcome.
TIA for them.

--
Best regards,
Michelle Baert

Re: madplayer cross-compile on ARM11 compilation error "impossible constraint in 'asm'"

2008-05-21T06:05:11Z

Thank you so much for your help, Rob!

Regards,
Jeff

> From: rob@...

> To: mad-dev@...
> Date: Sun, 18 May 2008 19:40:12 -0700
> Subject: Re: [mad-dev] madplayer cross-compile on ARM11 compilation error "impossible constraint in 'asm'"
>
> On May 18, 2008, at 12:33 AM, Jianfeng Li wrote:
> > I am now trying to cross-compile the madplayer on an arm11 based
> > embedded system. The compiler I am using is from montavista
> > mobilinux 5.0.
> >
> > When I tried to build the madplay, for the file resample.c, I got 2
> > errors on line 89 and 104. The error information is "impossible
> > constraint in 'asm'. When I tried to build the libmad, I also met
> > the same error. But it seems that during compilation if "FPM-ARM"
> > option is enabled then the error is solved. But here I checked the
> > Makefile, there is no such option of "FPM" as in the Makefile of
> > libmad.
> ! >
> > So can anybody help to tell me why this error happens, it is due to
> > the compiler's constraint/requirement? How should I try to solve
> > this error? Is "trying to enable the FPM" a good option to try?
>
> Running the libmad `configure' script is the usual way to cause an
> appropriate FPM macro to be defined. You can force the selection using
> the --enable-fpm option (e.g. --enable-fpm=arm) in case the
> automatically chosen default is inappropriate for some reason.
>
> Note that the FPM selection is carried into the mad.h header file
> which is used by madplay et al. When cross-compiling, the resulting
> mad.h could be different from the build system's mad.h so it's
> important to check your include paths.
>
> --
> Rob Leslie
> rob@...
>

Invite your mail contacts to join your friends list with Windows Live S! paces. It's easy! Try it!

Stream Decode

2008-05-20T10:29:20Z

Hi!

I have the following question - is it possible to decode a stream using MAD?
For example: I have shoutcast stream and I read it in small portions
(say, 40960 bytes), decode it and encode back to mp3 using LAME
(chaning bitrate, samplerate, etc...).

The problem is that I have to call mad_stream_buffer with each piece of
stream data, decode/encode it and then if I listen the resulting mp3
file - I notice "noise" between theese small decoded pieces.
Seems like it happen not only with VBR files.

Pseudo-code:

while ( bytes = read(file_id, buff, SHOUT_BUFFER_LENGTH ) ) {

mad_stream_buffer(&stream, buff, bytes );

while(1){
mad_frame_decode(&frame,&stream)
....
int ret = lame_encode_buffer( lameGlobalFlags,
left,
right,
nsamples,
mp3Buf,
mp3Size );
// Play resulting mp3 data

}

}

Is it possible to decode a stream without theese "edge effects" ? May be
I should not re-initialize MAD buffer after each read in some way?
Please help.

Re: madplayer cross-compile on ARM11 compilation error "impossible constraint in 'asm'"

2008-05-18T19:40:12Z

On May 18, 2008, at 12:33 AM, Jianfeng Li wrote:

> I am now trying to cross-compile the madplayer on an arm11 based
> embedded system. The compiler I am using is from montavista
> mobilinux 5.0.
>
> When I tried to build the madplay, for the file resample.c, I got 2
> errors on line 89 and 104. The error information is "impossible
> constraint in 'asm'. When I tried to build the libmad, I also met
> the same error. But it seems that during compilation if "FPM-ARM"
> option is enabled then the error is solved. But here I checked the
> Makefile, there is no such option of "FPM" as in the Makefile of
> libmad.
>
> So can anybody help to tell me why this error happens, it is due to
> the compiler's constraint/requirement? How should I try to solve
> this error? Is "trying to enable the FPM" a good option to try?

Running the libmad `configure' script is the usual way to cause an
appropriate FPM macro to be defined. You can force the selection using
the --enable-fpm option (e.g. --enable-fpm=arm) in case the
automatically chosen default is inappropriate for some reason.

Note that the FPM selection is carried into the mad.h header file
which is used by madplay et al. When cross-compiling, the resulting
mad.h could be different from the build system's mad.h so it's
important to check your include paths.

--
Rob Leslie
rob@...

madplayer cross-compile on ARM11 compilation error "impossible constraint in 'asm'"

2008-05-18T00:33:50Z

Hi, All,

I am now trying to cross-compile the madplayer on an arm11 based embedded system. The compiler I am using is from montavista mobilinux 5.0.

When I tried to build the madplay, for the file resample.c, I got 2 errors on line 89 and 104. The error information is "impossible constraint in 'asm'. When I tried to build the libmad, I also met the same error. But it seems that during compilation if "FPM-ARM" option is enabled then the error is solved. But here I checked the Makefile, there is no such option of "FPM" as in the Makefile of libmad.

So can anybody help to tell me why this error happens, it is due to the compiler's constraint/requirement? How should I try to solve this error? Is "trying to enable the FPM" a good option to try?

Many thanks in advance!

Regards,
Jeffen

Invite your mail contacts to join your friends list with Windows Live Spaces. It's easy! Try it!

Re: Initite loop bug in libid3tag-0.15.0b

2008-01-30T04:18:22Z

Rahul Banerjee wrote:
> Hi,
>
> I am new to the usage of this mad library. I have a question regarding
> this. Is the fast forward playback of mp3 is supported with this library?

Well yes, sort of. libmad will simply decode a stream.
The burden is placed on the user however to direct the
decoder to whatever bits you'd like to decode. So to
implement fast fwd, trick play, etc.. you'll need to
point the decoder at the bits you'd like to render which
say constitute snippets of audio to be rendered from a
stream in fast forward.

This requires repositioning the stream pointer ahead in
the stream and rendering frames until you have an
error free decode and are synchronized. Render as much
of decoded output to the audio device for a single hop,
reposition the decoder in the stream and repeat.

-john

--
john.cooper@...

Re: Initite loop bug in libid3tag-0.15.0b

2008-01-29T18:49:36Z

Hi,

I am new to the usage of this mad library. I have a question regarding this. Is the fast forward playback of mp3 is supported with this library?

Thanks and Regards,

Rahul Banerjee

On 1/13/08, Kentaro Oda <odaken@...> wrote:

Hi, for the maintainer libid3tag.

I found an infinite loop bug in libid3tag-0.15.0b library, which
causes memory overflow.

The problem occurs when parsing an ID3_FIELD_TYPE_STRINGLIST field,
specifically when data to be parsed is ended with '\0'.
In this case, **ptr == 0, but the condition end - *ptr is 1 so loop
continues infinitely.

*** field.c     2003-04-19 09:14:33.000000000 +0900
--- field-patched.c     2008-01-13 16:08:22.000000000 +0900
***************
*** 291,297 ****

       end = *ptr + length;

!       while (end - *ptr > 0) {
       ucs4 = id3_parse_string(ptr, end - *ptr, *encoding, 0);
       if (ucs4 == 0)
         goto fail;
--- 291,297 ----

       end = *ptr + length;

!       while (end - *ptr > 0 && **ptr != '\0') {
       ucs4 = id3_parse_string(ptr, end - *ptr, *encoding, 0);
       if (ucs4 == 0)
         goto fail;

--
Kentaro Oda

Initite loop bug in libid3tag-0.15.0b

2008-01-12T23:17:36Z

Hi, for the maintainer libid3tag.

I found an infinite loop bug in libid3tag-0.15.0b library, which
causes memory overflow.

The problem occurs when parsing an ID3_FIELD_TYPE_STRINGLIST field,
specifically when data to be parsed is ended with '\0'.
In this case, **ptr == 0, but the condition end - *ptr is 1 so loop
continues infinitely.

*** field.c 2003-04-19 09:14:33.000000000 +0900
--- field-patched.c 2008-01-13 16:08:22.000000000 +0900
***************
*** 291,297 ****

end = *ptr + length;

! while (end - *ptr > 0) {
ucs4 = id3_parse_string(ptr, end - *ptr, *encoding, 0);
if (ucs4 == 0)
goto fail;
--- 291,297 ----

end = *ptr + length;

! while (end - *ptr > 0 && **ptr != '\0') {
ucs4 = id3_parse_string(ptr, end - *ptr, *encoding, 0);
if (ucs4 == 0)
goto fail;

--
Kentaro Oda

madplay: How to select a specific ALSA card

2008-01-08T10:13:45Z

Hi,
I just started using madplay a short while ago and I was wondering
whether it is possible to select an ALSA card different from the default
one for audio output.
>From looking at the code it seems like the PATH part of --output is
used, however
madplay --output=alsa:plughw:1 myfile.mp3
chokes as the TYPE alsa does not exist (in audio.c).

Did I miss something in the command line or is this a bug / missing
feature (for which I could provide a patch)?

Thanks,
-Thomas

Re: memory leaking in mad_layer_III?

2007-09-27T03:50:40Z

Rob Leslie wrote:
> On Sep 26, 2007, at 2:55 PM, Niels Roosen wrote:
>> I will post the relevant code tomorrow, should I do that as attachment
>> or inline?
>
> Either is fine unless it's long, in which case attachment is probably
> preferred.

You're not going to believe this, but I seem to have solved the problem.
While scrutinizing the relevant code for posting I discovered a rather
odd construct in my code (don't know why I didn't see that before) and
after some more investigation I discovered that I was lying earlier. In
the initialization function (which is also responsible for calling the
mad_*_finish() functions if necessary) the condition for actually
calling the cleanup functions was not working at all! This was because I
mixed up something during development and clearly never tested this
properly. So I fixed that and didn't get -any- references to lost memory
in libmad anymore \o/.

Anyway, to summarize: yet another case of tunnel vision, so you try to
explain your problem to someone else and during that process find out
what is going wrong. So thanks nevertheless, I guess I needed this :-)

So I humbly apologize for being such a retard. Nevertheless, mad rocks
my socks, it is fast and reliable unlike the mpg123 decoder which often
fails to synchronize frames, miserably failing to decode an entire
stream... :-)

Cheers,

Niels

--
"One of the most frightening things about your true nerd, for many
people, is not that he's socially inept - because everybody's been there
- but rather his complete lack of embarrassment about it." - Randy
Waterhouse in Cryptonomicon

n: Arnie / Arnicus | e: niels@...
w: http://okkernoot.net/ |
irc: irc.okkernoot.net #geek | icq: 31817478
aim: arniea a arniea | msn: arnie@...

http://jungletrain.net/ - 24/7 Drum and Bass Internet Radio

Re: memory leaking in mad_layer_III?

2007-09-26T16:53:11Z

On Sep 26, 2007, at 2:55 PM, Niels Roosen wrote:
> I will post the relevant code tomorrow, should I do that as
> attachment or inline?

Either is fine unless it's long, in which case attachment is probably
preferred.

--
Rob Leslie
rob@...

Re: memory leaking in mad_layer_III?

2007-09-26T15:55:04Z

Rob Leslie wrote:

> On Sep 26, 2007, at 3:40 AM, Niels Roosen wrote:
>> Am I missing something or could it be that using mad this way hasn't
>> been tested excessively enough and is it true that the memory
>> alllocated in mad_layer_III() on lines 2530 and 2538 (libmad version
>> 0.15.1b) is not freed properly? I would be delighted to either get
>> confirmation on this or some advice on how to solve this issue.
>
> The memory allocated in mad_layer_III() is freed in mad_frame_finish()
> and mad_stream_finish(); you should make sure these are definitely being
> called on the same structures you pass to mad_frame_decode().
>
> Since you say you are making these calls, I'm at a loss to offer any
> specific advice without seeing your code.

What I do is initialize the decoder, decode incoming data per frame
until the client disconnects, and then, when another client connects,
run the mad_stream_finish() and mad_frame_finish() on the pointers to
the relevant structures before running the initialization functions
again, in order to clean up any remaining data in the decoder. I took
this from some example code but unfortunately don't know what any more,
but I suppose that is how it should be done.

I will post the relevant code tomorrow, should I do that as attachment
or inline?

Regards,

Niels

--
"One of the most frightening things about your true nerd, for many
people, is not that he's socially inept - because everybody's been there
- but rather his complete lack of embarrassment about it." - Randy
Waterhouse in Cryptonomicon

n: Arnie / Arnicus | e: niels@...
w: http://okkernoot.net/ |
irc: irc.okkernoot.net #geek | icq: 31817478
aim: arniea a arniea | msn: arnie@...

http://jungletrain.net/ - 24/7 Drum and Bass Internet Radio

Re: memory leaking in mad_layer_III?

2007-09-26T10:03:51Z

On Sep 26, 2007, at 3:40 AM, Niels Roosen wrote:
> Am I missing something or could it be that using mad this way
> hasn't been tested excessively enough and is it true that the
> memory alllocated in mad_layer_III() on lines 2530 and 2538 (libmad
> version 0.15.1b) is not freed properly? I would be delighted to
> either get confirmation on this or some advice on how to solve this
> issue.

The memory allocated in mad_layer_III() is freed in mad_frame_finish
() and mad_stream_finish(); you should make sure these are definitely
being called on the same structures you pass to mad_frame_decode().

Since you say you are making these calls, I'm at a loss to offer any
specific advice without seeing your code.

--
Rob Leslie
rob@...

memory leaking in mad_layer_III?

2007-09-26T04:40:27Z

Hello,

In the past while I have been rounding up development for a streaming
server for the radio station I am part of. However, during the last
stress testing phase I noticed that memory usage slowly increases as
streaming clients disconnect and connect. At first I was convinced it
was some leak in my own code but valgrind didn't confirm that this was
happening. However, I noticed that there was quite some lost memory
originating in the mad_layer_III() function in layer3.c:

==5957== 7,701 bytes in 3 blocks are definitely lost in loss record 68 of 85
==5957== at 0x481B72C: malloc (in
/usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
==5957== by 0x485F1C1: mad_layer_III (in /usr/lib/libmad.so.0.2.1)
==5957== by 0x485429A: mad_frame_decode (in /usr/lib/libmad.so.0.2.1)
==5957== by 0x7775: decode_data (server.c:318)
==5957== by 0x9387: server (server.c:980)
==5957== by 0x4874164: pthread_start_thread (in /lib/libpthread-0.10.so)
==5957== by 0x4A176E9: clone (in /lib/libc-2.3.6.so)

==5957== 13,824 bytes in 3 blocks are definitely lost in loss record 71
of 85
==5957== at 0x481CE9E: calloc (in
/usr/lib/valgrind/x86-linux/vgpreload_memcheck.so)
==5957== by 0x485F215: mad_layer_III (in /usr/lib/libmad.so.0.2.1)
==5957== by 0x485429A: mad_frame_decode (in /usr/lib/libmad.so.0.2.1)
==5957== by 0x7775: decode_data (server.c:318)
==5957== by 0x9387: server (server.c:980)
==5957== by 0x4874164: pthread_start_thread (in /lib/libpthread-0.10.so)
==5957== by 0x4A176E9: clone (in /lib/libc-2.3.6.so)

One would assume this is just caused by memory not being free()'d up
when the program exits, but the longer the server has run, the more
memory was definitely lost, indicating that the memory loss originates here.

I use the mad library in such a way that I manually line up data and
decode this data per frame, which is required to ensure certain
functionality of the streaming server. I made sure that when I
reinitialize the decoder every time I run the mad_*_finish() operations
on all the structs I use before running mad_*_init() again. However, I
suspect that not all allocated memory is freed as it should.

Am I missing something or could it be that using mad this way hasn't
been tested excessively enough and is it true that the memory alllocated
in mad_layer_III() on lines 2530 and 2538 (libmad version 0.15.1b) is
not freed properly? I would be delighted to either get confirmation on
this or some advice on how to solve this issue. If more information is
required I will do my best to provide this.

Regards,

Niels Roosen

--
"One of the most frightening things about your true nerd, for many
people, is not that he's socially inept - because everybody's been there
- but rather his complete lack of embarrassment about it." - Randy
Waterhouse in Cryptonomicon

n: Arnie / Arnicus | e: niels@...
w: http://okkernoot.net/ |
irc: irc.okkernoot.net #geek | icq: 31817478
aim: arniea a arniea | msn: arnie@...

http://jungletrain.net/ - 24/7 Drum and Bass Internet Radio

Re: Calculation error

2007-09-07T03:33:33Z

On Fri, Aug 24, 2007 at 12:16:12PM +0200, Tim Dylla wrote:
> I did some testing on my own implementation of mp3, and used MAD as
> reference.
> Therefore I generated a sine-sweep from 0Hz to 22kHz ( as far as this is
> possible ;) ) at 0dBFS, and 48kHz Samplerate. Then I encoded it with
> Lame for a single Channel MPEG-1 Layer III Bitstream, 48kHz Samplerate,
> 320kbps Bitrate. (I don't know the exact Lame-version and optionstring
> anymore, but I didn't use anything weird)

> I'm not a friend of posting large files on a mailing list without any
> agreement with the moderator, so if someone is interested in the
> bitstream and the resulting raw file, pleas feel free to send me a message.

Could the generation be done with common software? (sox? clm? nyqvist?)
That could make a sort of full automatic test system if scripted.

Also, a file like that might find a home on Wikimedia or Internet Archive.

--
To know what you prefer, instead of humbly saying Amen
to what the world tells you you ought to prefer,
is to have kept your soul alive.
-- Robert Louis Stevenson

libMad Vs ffmpeg

2007-08-24T16:05:29Z

I am porting MP3 at ARM9 platform.

If I use only fixed-point(Integer), I wonder which one is faster between MAD mp3 decoding and ffmpeg mp3 decoding.

Re: Calculation error

2007-08-24T12:53:37Z

Hi list again,

there is a bug in madplay 0.15.2b, which skips some frames in certain
conditions. Rob Leslie seems to have a patch already in the pipeline.
libmad is not affected at all.

Best,
Tim Dylla

Tim Dylla schrieb:

>
>
> Andre schrieb:
>> Tim Dylla <tim.dylla@...> wrote:
>>> MAD in contrast introduces a perceivable error right on the first
>>> half wave of the sine sweep. I think, it's an overflow in the
>>> requantization or the IMDCT (when I remember right, the huffman
>>> decoder produces very high output at this moment).
>>
>> Just a hunch, but if you are able to build MAD from source, maybe you
>> could try the follow patch posted to the list a few years back...
>> perhaps it helps ??
>>
>> diff -ruN mad-0.14.2b_orig/libmad/layer3.c
>> mad-0.14.2b/libmad/layer3.c
>> --- mad-0.14.2b_orig/libmad/layer3.c Thu Nov 8 15:28:02 2001
>> +++ mad-0.14.2b/libmad/layer3.c Tue Mar 25 15:21:12 2003
>> @@ -910,7 +910,7 @@
>> fprintf(stderr, "requantize overflow (%f * 2^%d)\n",
>> mad_f_todouble(requantized), exp);
>> # endif
>> - requantized = MAD_F_MAX;
>> + requantized = MAD_F_MAX / 2;
>> }
>> else
>> requantized <<= exp;
>
> Nope, same Result. I'm sorry, I'm still sticking with my diploma, so I
> don't have the time for extensive tests right now. Maybe someone can
> confirm the problem and have a look at? Feel free to contact me for the
> Bitstream in Question and the correct decoder output.
>
> Best,
>
> Tim Dylla
>
>
>>
>>
>> Andre
>> --
>>
>>
>
>

Re: Calculation error

2007-08-24T07:35:08Z

Andre schrieb:

> Tim Dylla <tim.dylla@...> wrote:
>> MAD in contrast introduces a perceivable error right on the first
>> half wave of the sine sweep. I think, it's an overflow in the
>> requantization or the IMDCT (when I remember right, the huffman
>> decoder produces very high output at this moment).
>
> Just a hunch, but if you are able to build MAD from source, maybe you
> could try the follow patch posted to the list a few years back...
> perhaps it helps ??
>
> diff -ruN mad-0.14.2b_orig/libmad/layer3.c
> mad-0.14.2b/libmad/layer3.c
> --- mad-0.14.2b_orig/libmad/layer3.c Thu Nov 8 15:28:02 2001
> +++ mad-0.14.2b/libmad/layer3.c Tue Mar 25 15:21:12 2003
> @@ -910,7 +910,7 @@
> fprintf(stderr, "requantize overflow (%f * 2^%d)\n",
> mad_f_todouble(requantized), exp);
> # endif
> - requantized = MAD_F_MAX;
> + requantized = MAD_F_MAX / 2;
> }
> else
> requantized <<= exp;

Nope, same Result. I'm sorry, I'm still sticking with my diploma, so I
don't have the time for extensive tests right now. Maybe someone can
confirm the problem and have a look at? Feel free to contact me for the
Bitstream in Question and the correct decoder output.

Best,

Tim Dylla

>
>
> Andre
> --
>
>

Calculation error

2007-08-24T04:16:12Z

Hi List,

I did some testing on my own implementation of mp3, and used MAD as
reference.
Therefore I generated a sine-sweep from 0Hz to 22kHz ( as far as this is
possible ;) ) at 0dBFS, and 48kHz Samplerate. Then I encoded it with
Lame for a single Channel MPEG-1 Layer III Bitstream, 48kHz Samplerate,
320kbps Bitrate. (I don't know the exact Lame-version and optionstring
anymore, but I didn't use anything weird)
As expected, the resulting bitstream can be decoded without any obvious
error with my Decoder and also with the dist10 decoder.
MAD in contrast introduces a perceivable error right on the first half
wave of the sine sweep. I think, it's an overflow in the requantization
or the IMDCT (when I remember right, the huffman decoder produces very
high output at this moment).
I used MAD 0.15.2b in conjunction with MADFrontend on a
32bit-Windoze-machine. The option string was -b24 -d

I'm not a friend of posting large files on a mailing list without any
agreement with the moderator, so if someone is interested in the
bitstream and the resulting raw file, pleas feel free to send me a message.

With regards,

Tim Dylla

[Offtopic] Stories about Process Change

2007-08-23T09:26:59Z

Hi MAD-Hackers, Hi Rob,
I am researching what happens if a developer attempts to change the way
a project works (in contrast to changing the software being developed) and
have picked MAD as a small OSS project in long-term existence that seems
to use a closed development strategy. I would be very interested in
hearing if there was ever the need to modify the development process and
if yes how successful have those changes been.

I put up a survey of 13 questions (should take between 10 and 20 minutes)
at

http://survey.mi.fu-berlin.de/public/survey.php?name=changeEpisodeSurvey

but you can also email me directly if you prefer.

Thank you very much,

Christopher Oezbek

P.S.: If you are thinking about making such a change or are currently on
it,
feel free to drop me a note so I can share what I have learned so far.

P.P.S.: Results will be made available at

http://www.inf.fu-berlin.de/w/SE/SurveyChange

over the next month, where you can also find further information.

--
Christopher Oezbek
Freie Universität Berlin
http://www.inf.fu-berlin.de/~oezbek/

problem with the seeking over

2007-08-17T02:13:28Z

Hi,

Ok, I managed to seek in a frame

Thanks for your help

Cab

Re: mad-dev Digest, seeking problem with libmad

2007-08-16T03:52:42Z

Hi,

You said I had to map the current file at a seek point

but I just can't do itn there's a problem with the function mmap itself

here's my source: (from minimad)

int main(int argc, char *argv[])
{
struct stat stat;
void *fdm;

if (argc != 1)
return 1;

int fd;
FILE*fichier=fopen("essai.mp3","r+");
fd=fileno(fichier);
lseek(fd,10,SEEK_SET);
if (fstat(fd, &stat) == -1 ||
stat.st_size == 0)
return 2;
fdm = mmap(0,stat.st_size/2-1 , PROT_READ, MAP_SHARED,
fd,stat.st_size/2);
if (fdm == MAP_FAILED)
return 3;

decode(fdm, stat.st_size/2-1);

if (munmap(fdm, stat.st_size/2-1) == -1)
return 4;

return 0;
}

I've got an error return 22

I know it's because the offset is not set to 0

thanks for your help

Re: test-bitstreams

2007-07-29T22:53:37Z

Hi Tim,

We, at Patni, have obtained the test bitstreams for strict
conformance of the MP3-Decoder from the following link:
http://standards.iso.org/ittf/PubliclyAvailableStandards/

Here you might want to click on the 11172-4:1995
"Compliance testing bitstreams" link for the
conformance test vectors. These vectors do test most of the
features of the MP3-Decoder.

Regards,
Abeezar Vasi.

-----Original Message-----
From: mad-dev-bounces@...
[mailto:mad-dev-bounces@...]On Behalf Of Tim Dylla
Sent: Saturday, July 28, 2007 7:37 PM
To: MAD-Development List
Subject: [mad-dev] test-bitstreams

Hi list,

since I'm doing final tests on my mp3-decoder implementation for a
Blackfin DSP, I'm searching for teststreams coming with a decoded
reference.

The original ISO/IEC/Fraunhofer-teststream package contains only the
decoded reference of the compliance bitstream. But (as you certainly
know) it doesn't test the whole functionality of MP3 (not even short
blocks), and the decoder in this package is crap (limited compliance,
far away from being a reference).

I'm searching for more bitstreams in company with their decoded
reference to avoid adding up deviance of the used Encoder to my test
results. Has anyone some resource regarding this? I'm interested in
MPEG1 and 2 - Layer III bitstreams.

Thanks in advance,

Tim

http://www.patni.com
World-Wide Partnerships. World-Class Solutions.
_____________________________________________________________________

This e-mail message may contain proprietary, confidential or legally
privileged information for the sole use of the person or entity to
whom this message was originally addressed. Any review, e-transmission
dissemination or other use of or taking of any action in reliance upon
this information by persons or entities other than the intended
recipient is prohibited. If you have received this e-mail in error
kindly delete this e-mail from your records. If it appears that this
mail has been forwarded to you without proper authority, please notify
us immediately at netadmin@... and delete this mail.
_____________________________________________________________________

test-bitstreams

2007-07-28T08:06:49Z

Hi list,

since I'm doing final tests on my mp3-decoder implementation for a
Blackfin DSP, I'm searching for teststreams coming with a decoded
reference.

The original ISO/IEC/Fraunhofer-teststream package contains only the
decoded reference of the compliance bitstream. But (as you certainly
know) it doesn't test the whole functionality of MP3 (not even short
blocks), and the decoder in this package is crap (limited compliance,
far away from being a reference).

I'm searching for more bitstreams in company with their decoded
reference to avoid adding up deviance of the used Encoder to my test
results. Has anyone some resource regarding this? I'm interested in
MPEG1 and 2 - Layer III bitstreams.

Thanks in advance,

Tim

Re: Computation Speed

2007-07-23T12:53:04Z

john cooper schrieb:
> Tim Dylla wrote:
>
>> Now I wonder: how much MIPS needs MAD? Say, for the x86-Version
>> optimized for Accuracy? Has anyone measured this? Or is there a Webpage
>> with all that info, and I was just dumb? :)
>
> A rough approximation is about 50MIPS for a 128bps stereo
> decode at 44.1khz.

Thanks a lot,

tim

>
> -john
>

Re: Computation Speed

2007-07-23T11:51:19Z

Tim Dylla wrote:

> Now I wonder: how much MIPS needs MAD? Say, for the x86-Version
> optimized for Accuracy? Has anyone measured this? Or is there a Webpage
> with all that info, and I was just dumb? :)

A rough approximation is about 50MIPS for a 128bps stereo
decode at 44.1khz.

-john

--
john.cooper@...

Computation Speed

2007-07-23T11:12:44Z

Hello List,

I'm new to this list, let me introduce myself.

I'm working on my Diploma for Applied Informatics, I have to implement
the MP3-Decoder for the Analog Devices BF-533 DSP. It's the the final
testphase regarding accuracy, where MAD really convinces. Since I want
to leave the needed MIPS of my Algorithm low, I had to make some
compromise between Accuracy and Speed.

Now I wonder: how much MIPS needs MAD? Say, for the x86-Version
optimized for Accuracy? Has anyone measured this? Or is there a Webpage
with all that info, and I was just dumb? :)

Best Regards,

Tim Dylla