Linux-audio-dev

linux-audio-dev@lists.linuxaudio.org

3 participants
7535 discussions

Re: [LAD] Audio problem for numerical optimization

by Fons Adriaensen

On Mon, Mar 03, 2014 at 11:28:49PM +0100, "Jeremia Bär" wrote: > I'm not sure I get the difference here. As I understand, optimization includes > (1) choice of suitable algorithm for the problem, (2) smart implementation, i.e. > writing code such that the compiler can apply smart optimization (e.g. aliasing) > and (3) optimizing for a particular microarchitecure (leveraging instruction > level paralellism for a fixed CPU model). The difference is that my (2) only requires programming skills, while (1) would require familiarity with the application domain. A trivial example: say you have a vector of 1024 floating point values and you need to compute log10 of all of them. In a general purpose routine you have no choice but to test each value x for x > 0.0, and that is something we like to avoid inside a loop or in vector code. But if you know the application domain, you may know that all values will be > -0.001, and that adding 0.002 to them won't affect the result in any way that matters in practice. So you can just compute log10 (x + 0.002) and remove the test. In real-world cases such changes may be much more invasive. For example, when designing a demodulator/decoder for a telecom system, you will have a 'degradation budget', say 0.1 dB. This means that your algorithm is allowed to perform as if the S/N ratio of the input signal was 0.1 dB less than it really is. You're free to spend that 0.1 dB wherever you want. But you can do this only if you understand the consequences of e.g. using a less accurate computation at some point, and are able to demonstrate (by analysis) that you remain within the budget by doing so. This requires understanding the algorithm at a much deeper level than would be required to code it given a detailed description. > We will have to submit code to a university-internal repository and it will run > through some software plagiarism system. Would that be a problem? No, but I can't reveal the actual way it will be used. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow)

11 years, 11 months

[LAD] Audio problem for numerical optimization

by user web210p1

Hi! I'm currently taking a course in University called "How to write fast numerical code". We are to do a group project of 3 persons where we choose a numerical algorithm and optimize it over the course for a particular processor (e.g. apply vectorization, cache optimization etc...). I was wondering if there is an interesting algorithm in audio processing that we could apply. In particular, if there is any library developer out here that would be interested in an optimization of one of the algorithms used, we would be happy to contribute. Ideally, the problem is compute intensive (i.e. bottleneck of algorithm is not the memory hierarchy). FFT is used as an example in the lecture so it cannot be used, unless there is a significant additional computation involved. Thanks very much for suggestions, Jeremia

11 years, 11 months

[LAD] JACK on Gigaport HD+ at 4800kHz (reduced channel count)

by Jörn Nettingsmeier

Hi *! I'm trying to get my Gigaport HD+ to run at 48000kHz. The specs say it is capable of 8ch @ 44k1/16, 6ch at 44k1/24 and 48k/24. When I try to start it at 48k, it comes up ok but ends up running at 44k1. It shows 8 channels, so that's expected. Now, how do I tell it to use only 6, so that I can get to 48k? I've tried setting -o6, which gives the usual "cannot set playback channel count". Next I looked at amixer -D hw:3 controls, and there is a playback switch map that is set to "on,on,on,on,on,on,on,on". I set it to "on,on,on,on,on,on,off,off" and retry. No luck. Still fails to set playback channel count, JACK still comes up with 8 outs and falls back to 44k1. What am I missing? Is there some .asoundrc magic that might come to the rescue? Maybe define a device that has only 6 channels to begin with? But how do I do that? Any hints much appreciated. Best, Jörn -- Jörn Nettingsmeier Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487 Meister für Veranstaltungstechnik (Bühne/Studio) Tonmeister VDT http://stackingdwarves.net

11 years, 11 months

[LAD] Spam on LAA

by Andres Cabrera

Hi, I saw a message with this title in LAA: lm3jo en deb version 1.1.1-5 It has a link to a deb file, which looks very suspicios. Is it malware for Debian? It would be the proof that this is the year of the Linux desktop. Cheers, Andrés

11 years, 11 months

Re: [LAD] NetJack question

by Mark Hills

On Mon, 24 Feb 2014, David J Myers wrote: > I need to take the audio input from my sound hardware on one Linux box, > send it over the network to a second Linux box and play it back on the > second box sound hardware. > Can this be done using jackd and jack_netsource without modification? If > not, can it be done with a simple Jack applet of my own? If not, how can > I implement this quickly without writing my own streaming server and > client? Without wanting to hijack the thread (your subject is NetJack but your question more general), I had a similar use case and this could be useful to you: http://www.pogo.org.uk/~mark/trx/ Thanks -- Mark

11 years, 11 months

Re: [LAD] NetJack question

by Paul Davis

On Mon, Feb 24, 2014 at 10:38 AM, David J Myers < david.myers(a)amg-panogenics.com> wrote: > > user@ubuntu:~$ jackd -d net > jackd 0.121.2 > JACK 1 is now at 0.124.1 ... upgrading is recommended. You are using a version that is heading towards 4 years old.

11 years, 11 months

Re: [LAD] NetJack question

by Adrian Knoth

On Mon, Feb 24, 2014 at 11:58:21AM -0000, David J Myers wrote: > I need to take the audio input from my sound hardware on one Linux box, send > it over the network to a second Linux box and play it back on the second box > sound hardware. Yes. Using the latest jackd1 release will make this easier, however, you can also go for whatever your distribution has available. Basically, you run $ jackd -d alsa on the first box and $ jackd -d net on the second. If you have jackd2, it would be "netone", not "net". You then use Fons' zita-ajbridge on the second machine to connect to your physical I/O. (they replace the older alsa_in/out tools) The latest jackd1 can do the last two steps in one step. Maybe netjack2 is also an option for you. Go for whatever works. Cheers -- mail: adi(a)thur.de http://adi.thur.de PGP/GPG: key via keyserver

11 years, 11 months

[LAD] NetJack question

by David J Myers

Hi there, I'm trying to find out whether I can use NetJack1 to do the following:- I need to take the audio input from my sound hardware on one Linux box, send it over the network to a second Linux box and play it back on the second box sound hardware. Can this be done using jackd and jack_netsource without modification? If not, can it be done with a simple Jack applet of my own? If not, how can I implement this quickly without writing my own streaming server and client? Many thanks for your time. DavidJ

11 years, 11 months

Re: [LAD] [LAU] JACK latency API clarifications

by Lieven Moors

On Fri, Feb 21, 2014 at 03:29:40PM -0500, Paul Davis wrote: > On Fri, Feb 21, 2014 at 3:27 PM, Lieven Moors <lievenmoors(a)gmail.com> wrote: > > > Aren't most clients checking for > > the sample rate in the process callback anyway? > > > if they are, then they are doing it wrong. That would be me ;-) I probably did this because I thought it could change at period boundaries. Could you explain why it is wrong?

11 years, 11 months

[LAD] FVWM-Crystal-3.3.2 is out

by Dominique Michel

This is an exceptional announcement. You may ask why announce a FVWM-Crystal release on the LAA, fvwm-crystal being a desktop. That's simple, fvwm-crystal was improved in 3.3.0 with a key modifier editor which let you change its key modifiers. This will affect only the extensive set of fvwm-crystal own key-bindings, and the key bindings of other applications will remain fully unaffected. That's a very easy and fast way to have the extensive key-bindings set of applications like ardour or emacs to not collide with the desktop. During the last 2 years, a lot of bug fixes was done, the existing features was improved and unified between the recipes, and new features was added. The list of changes is so huge that it is better to try it for yourself, if you don't already have done it. During the last few months, a particular effort was made to fix some nasty bashisms. That should make fvwm-crystal to work much better with distributions using dash as default shell, and fix one hopefully unnoticed security issue. A short and very partial story of the last releases: 3.3.2 bring 2 new recipes with ACPI support (battery, cpu speed and temperature). 3.3.1 bring a new and complete Dutch translation and a Preference Editor for the preferences that don't have menu options. It bring too a pot file for easy translations in new languages. 3.2.6 see an updated Russian translation and x-terminal-emulator support (Debian). 3.2.4 add mount/umount/pmount/pmount-gui support with the built-in desktop icons. 3.2.1 finally succeed to remove the restart with all preference changes but one. 3.2 see a complete rewrite of the Thunar desktop icons manager. Not only the buttons was redone from the scratch into a single button, which fix the flickering, but also a new implementation of the associated actions was done. Now it's possible to associate any file manager you can think about with these icons, 2 are supported at the same time, as well than custom commands. ... For a complete list, see the ChangeLog file. Download: https://sourceforge.net/projects/fvwm-crystal/files/?source=navbar Website: http://fvwm-crystal.sourceforge.net/ Project site: https://sourceforge.net/projects/fvwm-crystal/?source=navbar Enjoy, Dominique

11 years, 11 months

Jump to page:

2026

2025

2024

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