mad-dev

mad-dev@lists.mars.org

682 discussions

Resampling 44.1k to 48k
by Ira Weiny 09 May '01

09 May '01

I have an application which needs to decode mp3's to 48kHz. As most mp3's are 44.1 I am going to have to resample. I am using libmad to decode the mp3 just fine by starting with the minimad source. I then integrated the resampling algorithim in resample.c. I noticed from the comments that this was not the best resampling algorithim available. Has anyone else done something better with fixed point math? Ira

1 0

MAD timer API
by Rob Leslie 08 May '01

08 May '01

The basic timer datatype is mad_timer_t; it represents a unit of time. You can initialize a timer by setting it either to a zero constant or to some other explicit value: mad_timer_t timer; timer = mad_timer_zero; /* or */ mad_timer_set(&timer, seconds, frac_numer, frac_denom); Timers have a precision of 1/352800000 second, so you can use the frac_numer and frac_denom arguments to specify a fractional value as a rational number. The resulting timer value, in seconds, is: seconds + (frac_numer / frac_denom) You can add one timer value to another with: mad_timer_add(&timer, timer2); You can also multiply a timer value by an integer: mad_timer_multiply(&timer, scalar); Timer values can be either positive or negative; to change the sign of a timer: mad_timer_negate(&timer); To get the absolute value: timer2 = mad_timer_abs(timer); Or, to determine the sign: sign = mad_timer_sign(timer); /* < 0 negative; = 0 zero; > 0 positive */ In general you can compare any two timer values: comp = mad_timer_compare(timer1, timer2); /* comp < 0 : timer1 < timer2 */ /* comp = 0 : timer1 = timer2 */ /* comp > 0 : timer1 > timer2 */ Finally, there are methods for extracting the value of a timer. To convert a timer value to selected UNITS (see libmad/timer.h for unit constants): count = mad_timer_count(timer, UNITS); /* rounded down */ If you assume a basic unit of seconds, you can get the remaining fractional part of a timer in arbitrary terms with: frac_numer = mad_timer_fraction(timer, frac_denom); The precision of the timers is such that frac_denom can be any valid MPEG audio sampling frequency to get sample-granular results. Finally, you can get a string representation by supplying a sprintf-style format and selecting various units: mad_timer_string(timer, dest, format, UNITS, frac_UNITS, subparts); For example, to get HH:MM:SS.DDD you could use: mad_timer_t timer; char dest[13]; mad_timer_string(timer, dest, "%02lu:%02u:%02.%03u", MAD_UNITS_HOURS, MAD_UNITS_MILLISECONDS, 0); (MAD_UNITS_HOURS implies the next two format specifiers are for minutes and seconds.) There are also selectors for video frame units, and even drop-frame time coding is supported: mad_timer_string(timer, dest, "%02lu:%02u:%02u;%02u", MAD_UNITS_HOURS, MAD_UNITS_29_97_FPS, 0); After each call to mad_frame_decode() (or mad_header_decode()), frame.header.duration is filled with a timer value representing the playing time of the audio in the frame. You can accumulate these times with mad_timer_add() to keep an accurate time position as you play the audio. -- Rob Leslie rob(a)mars.org

1 0

autoconf macros for mad
by Sam Clegg 24 Apr '01

24 Apr '01

Has anyone thought of adding some autoconf macros (e.g. mad.m4) to libmad so that other packages using it can use aclocal to simply verify the version installed on a given build system? I work on a couple of projects that would find this useful. sam -- sam clegg email: samc(a)superduper.net www : http://www.superduper.net pgp key : http://www.superduper.net/~samc/key.gpg

1 0

mad fails to build when gettext is not installed and a make distclean patch
by Sean 'Shaleh' Perry 23 Apr '01

23 Apr '01

So it seems there is some kind of typo happening in the autoconf magic. If gettext is not installed when the make process reaches the po/ directory the build dies because make tries to run 'no'. This appears to be some kind of oops from when configure registered that gettext did not exist. Also, distclean fails to remove the po/*.gmo files as well as the created intl/libintl.h link. This causes the debian build process to fail when it tries to run diff against the original source and the current debianized tree. Enclosed is a patch for this. --- mad-0.13.0b.orig/intl/Makefile.in +++ mad-0.13.0b/intl/Makefile.in @@ -171,7 +171,7 @@ clean: mostlyclean distclean: clean - rm -f Makefile ID TAGS po2msg.sed po2tbl.sed + rm -f Makefile ID TAGS po2msg.sed po2tbl.sed libintl.h maintainer-clean: distclean @echo "This command is intended for maintainers to use;" --- mad-0.13.0b.orig/po/Makefile.in.in +++ mad-0.13.0b/po/Makefile.in.in @@ -192,7 +192,7 @@ clean: mostlyclean distclean: clean - rm -f Makefile Makefile.in POTFILES *.mo *.msg *.cat *.cat.m + rm -f Makefile Makefile.in POTFILES *.gmo *.mo *.msg *.cat *.cat.m maintainer-clean: distclean @echo "This command is intended for maintainers to use;" Oh, and while I am thinking about it (and this is purely cosmetic) why are so many files marked +x in the tree?

1 0

TTY controls patch
by Rob Leslie 16 Apr '01

16 Apr '01

Hi folks, Soon after releasing 0.13.0b I decided to hack on TTY controls for `madplay'. The result is here: ftp://ftp.mars.org/pub/mpeg/testing/tty-0.13.0b.patch Many thanks go to Matthew Green for his initial implementation of this, although I've taken a slightly different approach. I've not tested this thoroughly which is why I'm making it available as a patch rather than making a new release. I'd appreciate feedback, especially from anyone using the audio_sun module (Solaris and NetBSD, probably.) TTY controls are enabled by default unless stdin is not a terminal or the -o (--output) option is used. You can also enable/disable the controls explicitly with --tty-control or --no-tty-control. The TTY control keys are: p pause s stop (return to beginning of current file) f forward (skip to next file) b backward (*) q quit i information (non-verbose modes only) (*) if pressed within the first 4 seconds of a playing file, skip to the previous file; otherwise replay the current file from the beginning Cheers, -rob

2 1

Signals patch
by Rob Leslie 13 Apr '01

13 Apr '01

There is another patch for `madplay' I forgot to mention, contributed by Diomidis Spinellis, which (according to Diomidis) allows you: 1. to specify a list of filenames as an @ argument to madplay (as in mpg123) and 2. control the player with two signals: SIGUSR1 to play the previous song SIGUSR2 to play the next song This patch may conflict with the TTY patch, but here it is nonetheless, modified only to allow clean patching against 0.13.0b: ftp://ftp.mars.org/pub/mpeg/contrib/signal-0.13.0b.patch I may incorporate this functionality into a future release. Cheers, -rob

1 0

MAD 0.13.0b available
by Rob Leslie 11 Apr '01

11 Apr '01

MAD version 0.13.0b is now available. Highlights of this release include: - complete file decode, including last frame - support for reading Xing VBR headers in the plug-in for Winamp - fixed visuals display in the plug-in for Winamp for 24- and 32-bit output - support for Layer III ancillary data extraction - fixed MPEG-2 Layer III intensity stereo decoding - new `--downsample' option for madplay (reduced decoding overhead) - fixed PIC-related crashing bug under ARM - libmad should now be thread-safe More details can be found in the `CHANGES' file in the release. The source code is here: ftp://ftp.mars.org/pub/mpeg/ The plug-in for Winamp is here: http://www.mars.org/home/rob/proj/mpeg/mad-plugin/ Cheers, -- Rob Leslie rob(a)mars.org

1 0

New mp3/madplay frontend (fwd)
by Nicolas Pitre 01 Apr '01

01 Apr '01

Might be of some interest here as well... ---------- Forwarded message ---------- Date: Sat, 31 Mar 2001 23:11:54 -0500 From: Greg Berenfield <gberenfield(a)berenfield.com> To: ipaq(a)handhelds.org Subject: [iPAQ] New mp3/madplay frontend Greetings, This is to announce the release of pyPAQ for the ipaq. pyPAQ is a python/gtk frontend for madplay that has many features with a form factor suited to the ipaq. pyPAQ requires a linux distribution with python, specific python libraries, and also the latest version of madplay + madtime (.12.5b), all of which are available as pre-built binaries on the pyPAQ website. Features include: Playlist support with intuitive file-selection FF/RW slider control ID3 Tag support Built-in mixer (volume/treble/bass) Everything is available at http://ipaq.sytes.net pyPAQ is based on the work of Edward Muller (original pyPAQ, intuitive interface design), and Ulf Betlehem (cplay - ncurses python mp3-frontend) but all bugs are my own. All problem reports, comments, and suggestions are greatly appreciated. Regards, Greg B.

1 0

Re: PR 12312 - madplay dumps core
by Andre 25 Mar '01

25 Mar '01

Rob Leslie <rob(a)mars.org> wrote: > > BTW, I just stumbled across something you might want to look at: > http://groups.google.com/groups?hl=en&lr=&safe=off&th=b1aed43880d58f4d&ic=1… Oops, looks like my fault :-) I had vaguely considered whether the code was PIC safe or not, and assumed that because the table of constants is declared inline, and because the ARM has easy access to the program counter (unlike MIPS for example) the assembler would generate the address of the array from the current PC, but I didn't check the final code. Anyway, the good news is that the fix is fairly easy - the attached patch should do the job. It is however untested (*) so please try it out and let me know. [ (*) It is untested mainly because I still haven't managed to get Linux working on my ARM 'Integrator' board. I have compiled a kernel, but the 'milo' bootloader for this platform seems to be ARM proprietary and therefore unavailable. If anyone can help me with this I would be most grateful. ] Andre -- ____________________________________________________________ Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie

2 1

(no subject)
by chf＠sunplus.com.tw 24 Mar '01

24 Mar '01

Dear Rob, and All: I download MAD plusg-in 0.12.5b.exe from www.mars.org And Decoding Result of in_mad.dll is different from in_mp3.dll Using in_mad.dll: vol of right channel is larger than left. but if using in_mp3.dll, left and right have same vol. If Possible, please tell me which one is correct ? Best Regards, Judy 23.March, 2001 (See attached file: 8.mp3) (See attached file: 8.mp3)

1 0

problem about MAD plug-in for winamp
by chf＠sunplus.com.tw 23 Mar '01

23 Mar '01

1 0

changing the tempo in realtime
by Gad Hayisraeli 05 Mar '01

05 Mar '01

could it be possible to change the tempo of an mp3 song in realtime, using the mad ? (i mean tempo = like in the MIDI format - not changing the pitch but just the speed that the musical notes being playing.) can this change within the mad code cause heavy latency ? Gad

2 1

Compile libmad on other architecture
by Sunny Chan 28 Feb '01

28 Feb '01

Hello, I am trying to integrate libmad into my MP3 player project - I am trying to presuade make to compile using arm-elf-gcc & co instead of normal gcc - however if I do ./configure --target=arm-elf --enable-fpm=arm The autoconf script didn't change the compiler used by make therefore I can't compile it... How can I change the libmad so that it will use arm-elf-gcc, etc instead of normal gcc? Thanks!

2 3

wrapping MAD (mad rocks)
by Rob Leslie 23 Feb '01

23 Feb '01

FYI -rob Read and respond to this message at: http://sourceforge.net/forum/message.php?msg_id=114204 By: coplan Just thought that everyone might like to know that our project, the SoundBox Project ( http://sourceforge.net/projects/soundbox ), has chosen MAD to be our key command-line MP3 player. For those of you who don't know, SoundBox is a project to turn your old computer parts into a digital audio jukebox for your stereo. It's actually a database front end, and music selection system that will make things easier. Besides, it gives us something to do with our old processors. =) We're only in planning stages now, but we don't mind help when it comes to certain parts of the project. One of the big things is that we're going to be wrapping madplay with a perl script -- and we're going to be throwing songs from the database at madplay on the fly. It'll be an interesting project. MAD rocks -- we're really impressed with the sound quality.

1 0

MAD FreeBSD "Port"
by Jason R Mastaler 23 Feb '01

23 Feb '01

FreeBSD users; MAD is now part of the "Ports" collection which offers simplified software installation. See the `audio/mad' directory within your ports tree or <http://www.freebsd.org/ports/audio.html>. Enjoy.

1 0

Mad with VC++ 6.0
by Guillaume.Delile＠infineon.com 20 Feb '01

20 Feb '01

Hi everybody, I'm a new user of this wonderful mp3 decoder. As anyone port mad under windows (VC++) ? I'm just looking for a simple decoder ( mp3 -> pcm) not a player. Thank you, and again good job ! Guillaume.

2 1

mad performance question
by Gad Hayisraeli 19 Feb '01

19 Feb '01

i'm trying to get the best from madplay (0.2.12b), and my goal is to jump to certain milisecond in the song, when i get an interrupt (keyboard). i use mad_stream_skip within mad_header_dcode() to acheve this. while monitoring the latency between getting the interrupt and having the jump done, i noticed that both in assabet and intel PIII i acheve latency of between 2-50 miliseconds. by saying latency, i mean that while i wanted to jump to X position, i jumped de-facto to X+20ms, for example. can i make it better ? (i want to acheve <10ms or fixed latency up to 40ms) maybe my measure isnt right ? ( i take stream->current_frame time after the skip, when i get into mad_header_decode(), and the latency calculated by: latency=desired time-current time. maybe current time isnt calculated exctly ? mabe i need to use real-time linux kernel ? Gad

4 9

Re: resample problem..
by Rob Leslie 15 Feb '01

15 Feb '01

J.Saravanan wrote: > Thanks for the reply.. > >>> output: sample frequency 44100 Hz not availlable;closest 41666 Hz >>> output: cannot resample 44100 Hz to 41666 Hz >>> >>> And the program quits. >>> Can sombody tell me how to fix this problem. >> >> Currently `madplay' only accepts hardware sampling frequencies within 3% of >> the requested frequency. >> >> You could increase this tolerance to 6% by modifying player.c around line >> 503. > > Is that the line start with 3L * ... > I've changed it with 6L * ... Yes. > The result is little different.. My board completely hangs. After some > time the linux kernel crashes with panic. > > Any other Idea.. Seems to me it is a problem with your sound driver in the kernel. -- Rob Leslie rob(a)mars.org

1 0

MAD 0.12.5b available
by Rob Leslie 13 Feb '01

13 Feb '01

Hot on the heels of 0.12.4b, MAD 0.12.5b is now available with even more performance improvements. This version incorporates the long-awaited Huffman decoding optimization that performs requantization in the same step. Many thanks to Andre McCurdy for his help that led to this release. There are still a few more improvements to be made, but this probably represents the bulk of what can be optimized for significant gains. The source code release: ftp://ftp.mars.org/pub/mpeg/ The MAD plug-in for Winamp: http://www.mars.org/home/rob/proj/mpeg/mad-plugin/ The MAD home page: http://www.mars.org/home/rob/proj/mpeg/ Cheers, -- Rob Leslie rob(a)mars.org

2 1

Re: [mad-user] resample problem..
by Rob Leslie 12 Feb '01

12 Feb '01

"J.Saravanan" wrote: > I've compiled madplay for my arm target ( Its a CLPS7500 evaluation > board). When I try to run the application I get the following error > message... > > output: sample frequency 44100 Hz not availlable;closest 41666 Hz > output: cannot resample 44100 Hz to 41666 Hz > > And the program quits. > Can sombody tell me how to fix this problem. Currently `madplay' only accepts hardware sampling frequencies within 3% of the requested frequency. You could increase this tolerance to 6% by modifying player.c around line 503. (Maybe this is a better tolerance? Follow-ups to mad-dev please...) -- Rob Leslie rob(a)mars.org

1 0

MAD 0.12.4b available
by Rob Leslie 11 Feb '01

11 Feb '01

MAD 0.12.4b is now available. This version incorporates a number of performance improvements for all platforms. It also includes (untested) native fixed-point math support for the PowerPC platform, contributed by David Blythe. Since I've made some significant changes to the way the fixed-point math routines are maintained, I'd appreciate feedback from anyone using MAD on a PPC to make sure I haven't broken anything. Finally, this version changes the way mixed short blocks are handled in Layer III with respect to alias reduction, correcting another apparent ambiguity in ISO/IEC 11172-3. Since mixed blocks are rare (many encoders don't support them) this has until now gone unnoticed. Thanks to H.O. for reporting this. The source code release is here: ftp://ftp.mars.org/pub/mpeg/ Winamp users should notice improved performance in the MAD plug-in: http://www.mars.org/home/rob/proj/mpeg/mad-plugin/ The MAD home page is, of course: http://www.mars.org/home/rob/proj/mpeg/ Cheers, -- Rob Leslie rob(a)mars.org

5 11

alias reduction for mixed short blocks
by Rob Leslie 09 Feb '01

09 Feb '01

A quick note on the alias reduction change for mixed short blocks: Haruhiko Ogasawara informed me that in spite of ISO/IEC 11172-3's claim that "alias reduction is not applied for granules with block_type == 2 (short block)", alias reduction most likely should be applied to the lower two subbands of mixed short blocks, as these are long blocks. Of the decoders that support mixed blocks at all, most indeed apply alias reduction on the lower subbands, including recent FhG implementations. Apparently version 3.0 of FhG's reference decoder l3dec does NOT apply alias reduction here, however, so there's a bit of ambiguity. Consequently, I've added a new option --enable-strict-iso to `configure' to revert to the previous behavior. This is not enabled by default. A few examples of files encoded with alias vs. no alias reduction are here: http://www.geocities.com/nazone2001/ You'll find that the mixed_alias file has aliasing artifacts when decoded with MAD 0.12.3b but not with MAD 0.12.4b, and for the mixed_no_alias file vice versa. Presumably the mixed_alias file is the one which is correctly encoded. I know there are some LAME developers on this list: will the mixed block support coming in LAME perform alias reduction on the lower two subbands of mixed short blocks? -rob

1 0

(no subject)
by Ivan Mi 07 Feb '01

07 Feb '01

Hello, I started to perform some application under Win CE, which uses libmad algorithm . But there are several function (write, read, fcntl), which are not supported under Win CE. As far as I understand instead of these function can be used - fread, fwrite- with FILE* , instead of handle. But function, which associate handle with FILE* - _fdopen is not supported under WIN CE, too. I'll be very glad if you share some ideas how to solve this problem. Thank you in advance. Ivan ____________________________________________________________________ Get free email and a permanent address at http://www.netaddress.com/?N=1

2 1

xmms
by Sam Clegg 02 Feb '01

02 Feb '01

I was wondering if anyone had written an XMMS input plugin that uses libmad? If nobody is doing this I will start it right away. Also I have written an mp3 player using libmad's lowlevel API that can read and play shoutcast streams. I realize this might be beyond the scope of a pure decoder but it would be nice if you play shoutcast streams from the commandline with madplay. Should this kind of thing go into the source tree for madplay? sam -- sam clegg email: samc(a)superduper.net www : http://www.superduper.net pgp key : http://www.superduper.net/~samc/key.gpg

4 3

Winamp 3 Alpha release...Time to start thinking of a WA3 MAD Plugin?
by kaiwei 02 Feb '01

02 Feb '01

As was read all over the net, WA3 alpha has been released to let the general public be their free alpha testers. Perhaps it's time to start working on a WA3 MAD plugin so that when WA3 is finally stable, we can have a working version of MAD ready? I'm no developer but I've read that it's pretty easy to port 2.x plugins to the 3.x architecture. Something I read on a decoder test website post recently on the list. "In conclusion, the MAD decoder has no major decoding bugs, and the only problem compared to other decoders is that it clips the ends off many files. It doesn't quite live up to its accuracy claims - it is better than other 16-bit decoders, but doesn't yet represent the best that can be achieved." Room for improvements? kaiwei

2 1

← Newer
1
...
21
22
23
24
25
26
27
28
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

mad-dev