Linux-audio-dev September 2007

linux-audio-dev@lists.linuxaudio.org

55 participants
40 discussions

by Predrag Viceic

Hi, For some unknown reason, I seem to be unable to set SND_PCM_FORMAT_FLOAT format flag when initialising alsa driver. This happens with my both cards (on is the laptops AC'97 Audio Controller and the other a similar onboard cart) This line used to work with alsa 0.9 but return an "Invalid argument" error with alsa 1.0.14a : snd_pcm_hw_params_set_format(alsa_pcm_handle, alsa_hwparams, SND_PCM_FORMAT_FLOAT_LE) If I remove this line, the driver gets initialised but as integer sample format, which, of course, generates noise as I'm sending 32bit floats to the output... Example code at lines 1031-1041 : http://cvs.savannah.nongnu.org/viewvc/freecycle/src/soundplayer.cpp?annotat… Please, please help ! Predrag http://freecycle.redsteamrecords.com

15 years, 5 months

Re: [LAD] Playing Audio Over CAT5

by Tosif Ahmed

>>Well, you have a problem in that the wrt doesn't do floating point very well. MAD says it supports 100% fixed-point (integer) computation, and i am using libmad as of now. to do my decoding >> Otherwise I'd suggest netjack. I checked out netjack, but i dont think i want that, as i said i want the audio from the router directly to the speaker through CAT5 and balun assembly. Wat i am confused about is after the decoding how, wud i send it over the ethernet and how wud analog signals travel over ethernet.

15 years, 5 months

Playing Audio Over CAT5

by Tosif Ahmed

Hi, I am an undergrad student working on a project related to playing audio on a Linksys WRT54GL Router through cat5 cable. I have the following setup. A Linksys WRT54GL router with OpenWRT firmware on it. What i want to do is to play some audio file (prefrably mp3 format), and stream the output over Ethernet through CAT5 cables and convert it through a balun. I have broken the process in following steps 1. Get the audio file and decode it (through libmad or a similar library) 2. Convert this to analog form 3. Send it over Ethernet I understand the first part, and have already started the implementation, but the other two steps are where i have problems. To play it directly on a speaker i would need to convert it to Analog, but how do make those analog signals travel over ethernet? As you can i dont have much knowledge about audio programming, so i would appreciate any help i could get on this mailing list. Thanks.

15 years, 5 months

denormal handling

by Georg Holzmann

Hallo list! I am just thinking about the right strategy for denormal handling in a floating point (single or double prec) audio application (and yes I already read the docs of the different methods at musicdsp and so on ...) Basically my question is, if it is enough to simply turn on the Flush-to-zero and Denormals-are-zero mode and then compile everything with -msse -mfpmath=sse ? I know it won't run on older Pentium3,2 etc. - but for the machines which support this feature, is this enough ? Thanks for any hint, LG Georg

15 years, 5 months

/dev/uinput

by Jens M Andreasen

Hi! I am trying to create a userspace driver with module uinput and I am getting into trouble when I terminate the program that creates the device. After a while the machine will crash ... The object is to create a joystick device, and I might be missing some important bits of documentation. So far I am not writing any data to the interface, but js_demo will find it and open it, showing the expected number of axis. The name of the interface looks like random though, which worries me and gets me to suspect that something is wrong Any idea why the machine is crashing or has a pointer to some example I could study? As long as I just let the program run, there is no problem ... Only on termination. This is what I have gotten so far in the uinput department: --8<------------------------------------------------- #include <fcntl.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <signal.h> #include <linux/input.h> #include <linux/uinput.h> int create_uinput() { int fd = -1; fd = open ("/dev/uinput", O_RDWR); if (fd <= 0) { fprintf (stderr,"could not open uinput device\n"); exit(EXIT_FAILURE); } memset (&uinp, 0, sizeof (struct uinput_user_dev)); strncpy (uinp.name, "NoJoy!", 7); // FIXME! uinp.id.version = 4; uinp.id.bustype = BUS_USB; ioctl (fd, UI_SET_EVBIT, EV_ABS); ioctl (fd, UI_SET_EVBIT, EV_KEY); ioctl (fd, UI_SET_ABSBIT, ABS_X); ioctl (fd, UI_SET_ABSBIT, ABS_Y); ioctl (fd, UI_SET_ABSBIT, ABS_Z); ioctl (fd, UI_SET_ABSBIT, ABS_THROTTLE); ioctl (fd, UI_SET_KEYBIT, BTN_TOP); ioctl (fd, UI_SET_KEYBIT, BTN_TOP2); ioctl (fd, UI_SET_KEYBIT, BTN_BASE); ioctl (fd, UI_SET_KEYBIT, BTN_BASE2); ioctl (fd, UI_SET_KEYBIT, BTN_BASE3); ioctl (fd, UI_SET_KEYBIT, BTN_BASE4); // Create device write (fd, &uinp, sizeof (uinp)); if (ioctl (fd, UI_DEV_CREATE)) { fprintf (stderr,"could not create uinput device.\n"); exit(EXIT_FAILURE); } sigset(SIGINT,destroy_uinput); sigset(SIGTERM,destroy_uinput); return fd; } void destroy_uinput() { fprintf(stderr,"NoJoy says: BYE!\n"); if(uinp_fd > 0) { ioctl (uinp_fd, UI_DEV_DESTROY); close (uinp_fd); } exit(0); } --

15 years, 5 months

Re: Direct Stream Digital / Pulse Density

by Maitland Vaughan-Turner

Fons Adriaensen <fons(a)kokkinizita.net> sez: > > On Thu, Sep 27, 2007 at 09:03:32PM -0700, Maitland Vaughan-Turner wrote: > > > There was something like a 50% success rate in choosing the audio > > source correctly. > > Which means it was just a random selection... > Obviously... (not that it proves anything either way) I just thought it was an interesting coincidence that this article appeared right on the heels of our discussion of the subject. I *did* say maybe you're right and maybe it *is* all in my head. hehe, no need to rub salt in it. :) ~Maitland

15 years, 5 months

question on dynamic linking

by nescivi

Hiho, I am having a discussion on the supercollider front about what is the proper way for dynamic linking. as far as I know, you use ldconfig and have the library location that programs dynamically link to defined in /etc/ld.so.conf but what is supposed to happen if the user just installs the program to a directory in his home directory? how should the dynamic linking be defined? esp. if the user does not have the root rights to change anything in /etc/ld.so.conf or to run ldconfig. I did not find anything quick on the net about this, so maybe one of you can enlighten me what is the "proper" way of dealing with this. sincerely, Marije

15 years, 5 months

Re: [LAD] denormal handling

by Georg Holzmann

Hallo! > our experience with ardour has been that DC bias is measurably more > effective at reducing CPU load than DAZ, FTZ or both combined. DAZ and > FTZ do both help significantly, however. One more question: is it not necessary to deactivate DAZ, FTZ again after the application (or operation) ? Or is this done automatically ? (because I cannot see it e.g. in your ardour code) Because e.g. in this document: http://developer.apple.com/documentation/Performance/Conceptual/Accelerate_… they set it back afterwards. Thanks, LG Georg

15 years, 5 months

Re: [LAD] Re: Direct Stream Digital / Pulse Density

by Maitland Vaughan-Turner

Dominique Michel <dominique.michel(a)citycable.ch> sez: > > Personally, I don't like it. I prefer very much a good stereo sound in the > original language (with some kind of text if it is a language that I don't > understand) like on the Swedish TV. > > For that PCM-DSD stuff. I prefer PCM because we can archive a good sound > quality with a much lower bandwitch. DSD was fine at the beginning of digital > recording because it was nothing else (for what I know), but for today's > professional audio, DSD is a waste of resources because of the huge needed > bandwitch. > > Dominique > This hits near something I was wondering: Wouldn't it be pretty easy to do lossless compression on a DSD stream. Since it's only one bit, It seems like you could use simple run-length encoding to achieve pretty good results. ~Maitland

15 years, 5 months

Re: [LAD] Re: Direct Stream Digital / Pulse Density

by Maitland Vaughan-Turner

Gordon JC Pearce <gordonjcp(a)gjcp.net> sez: > > On Tue, 2007-09-25 at 19:31 +0200, Fons Adriaensen wrote: > > > A nice variation on this theme occured years ago at an AES conference. > > The speaker wanted to demonstrate that 'digital' sound was crap, by > > using the familiar 'push down the extended arm' test. Test persons > > listening to analog sound could easily resist, while they lost all > > force when listening to a digital recording. > > > > What the speaker didn't know was that the PA system used to play the > > tracks was fully digital... > > I once helped prepare the equipment for a double-blind test of speaker > cables. All the golden-eared audophiles picked out one cable as being > far superior to the others, with better clarity and definition in the > upper harmonics and tighter more defined bass or some such bollocks. > > I did have to buy my Mum a new extension lead for her lawnmower, though. > Sixty feet of Black and Decker's finest, with Speakon plugs soldered to > it. > > Gordon That's pretty freaking funny. Reminds me of the Penn and Teller where they sell the diners in a trendy restaurant water from a garden hose on the patio. haha http://www.youtube.com/watch?v=XfPAjUvvnIc it's really funny stuff; worth watching if you have a couple minutes. Incidentally, I get my water from a mountain spring up the road (can't do much better than that, eh?), although I *have* bought a bottle of two of Evian in my day :) As for the double blind audio test, I reckon you guys have all seen the new AES journal by now (if not, go to your mailbox). There is a double-blind test where people were played DVD-A and SACD's, but some of them were passed through an extra A/D/A stage at 16bit-44.1khz. There was something like a 50% success rate in choosing the audio source correctly. Maybe it *is* all in my head...? Although, I wonder if ear training has anything to do with it. I'm super dorky sometimes, and I used to mix multitrack projects to different bit-depths/sample-rates and then try to train myself to hear the differences between them. hahaha, I figure most people don't do that sort of thing... ~Maitland

15 years, 5 months

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

Linux-audio-dev September 2007