[opensuse-kde] Sound on SuSE/Linux
I am going to bring this question to this mail list as I think it is more apropos to discuss it here... I have just been through a proverbial nightmare trying to configure sound on my system and getting some of the various tools to work. Am running under SuSE11.2 x86_64, KDE 4.4.2. I will admit up front I am NOT a Linux guru but I am enough of a software engineer to recognize an overly complex architecture when I see one. In all of my wanderings around the internet and maillists, I have discovered there as so many different layers, tools and configuration requirements that I have got to wonder if there is any sort of overall sound system-architecture vision in place and a GUI guide for users, being developed, to follow. We have got LAYERS and LAYERS of ALSA, ESOUND, JACK, OSS, ESD, aRTs, PULSEAUDIO, GSTREAMER, PHONON and tools like XINE, KAFFEINE, MYTHTV and MPLAYER which go off in their own directions... codecs to locate... I mean seriously folks, for a user to set up and configure a sound system this is getting out of control! How to choose? What to choose? What and how to configure? What to install? How to hook up and link together? And NO this is far from an automated process of any sort, and NO it is NOT transparent to a user! I KNOW I just went through hell just trying to get Amarok to play (and XINE and KAFFEINE as well) along with the projectM-pulseaudio visualization tool.. Just to cite my latest headache as an example, (and this was just one of many I had!!!) I found that running projectM-pulseaudio, which is suppose to connect to the pulseaudio sound server, randomly chooses which sound device on my system as its "connectHelper". (I had two, one of which was actually DISABLED in YaST but apparently that has no effect on projectM-pulseaudio!!) And projectM often chose the disabled sound device which left it unresponsive to sound output from any of the other applications. Here is another - I discovered that XINE and KMIX were each acting as if they had control over the master volume level setting! I had a horrible time tracking down why KAFFIENE was not playing anything until I happened on and installed the XINE-GUI application and within it I discovered that it too had a sense of volume control. When I turned it up, KAFFIENE started working fine! It was as if XINE and KMIX were in a series connection too because once I turned up the volume level in XINE-GUI I could also control the volume level with KMIX! IMHO this is a horrid architectural flaw and can cause users endless amounts of grief trying to figure out why something like KAFFIENE is not working for them. One and ONLY one master volume level control should be allowed/active at a time per sound device, but this sort of error arises when no one is overseeing the overall sound architecture. Same goes for Mute controls, another really nasty problem I had tracking down in KMIX... I have no idea how to track down where such a problem really lies and what I must do to configure things correctly. YaST? PulseAudio? ProjectM? Is this a bug or is something misconfigured? Especially given this extremely complex sound architecture, how do I differentiate? Very few of the application GUI setup tools come with much in the way of guiding me to a solution and none are presenting me with a clear picture of just how everything works together with confidence building test procedures. The KDE Systems Settings for Multimedia does not present very many options either , with little guidance, and I even noted the loss of the old ALSA configuration tool. I suspect some architect(s) is trying to simplify what a user has to configure, but hiding configuration options does not allow a GUI to guide a user to a solution. It only makes it harder to find one when the GUI fails to achieve a satisfactory result. The designer(s) of the KDE Systems Settings GUI are correct in believing that it is not a good idea to simply offer all configuration options in a flat view either, but a well design GUI will guide a user to properly apply and set configurations options as needed until a satisfactory result can be achieved for that user's purposes. Websites I visited give me a rather complicated picture (probably outdated too in most cases) of their own perspective of the sound architecture but nowhere am I ever presented with a clear picture of how all these different components actually work and/or connect/integrate together and what applications use them, how and how to configure.... Is there any hope in the near future that this will get simpler and easier for us poor audiophiles? Will appreciate hearing thoughts and NO I am not intending to throw rocks... Just that if I had such a hard time, I got to wonder how others are fairing and I think providing feedback and starting discussions is useful to engineers... Marc...
Am Freitag 02 April 2010 04:09:00 schrieb Marc Chamberlin:
running under SuSE11.2 x86_64, KDE 4.4.2 I guess you mean openSUSE 11.2 with KDE SC 4.4.2. "SuSE" and "KDE 4" are long dead.
ALSA The Linux sound architecture.
ESOUND, Not installed by default. Not needed. Dead.
JACK, Not installed by default. Not needed.
OSS, Not installed by default. Not needed.
ESD, Not installed by default. Not needed. Dead.
aRTs, Not installed by default. Not needed. Dead.
PULSEAUDIO, Not installed by default. Not needed. Sucks.
GSTREAMER, Codec architecture which just uses ALSA. No need to configure it.
PHONON KDE's abstraction layer. Only of interest for developers. No need to configure it.
and tools like XINE, KAFFEINE, MYTHTV and MPLAYER which go off in their own directions... codecs to locate... I mean seriously folks, for a user to set up and configure a sound system this is getting out of control! How to choose? What to choose? What and how to configure? What to install? Just live with the default install and don't try to mess with it. If you don't mess with it, it works fine. -- To unsubscribe, e-mail: opensuse-kde+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kde+help@opensuse.org
On 4/2/2010 3:35 AM, Markus wrote:
Am Freitag 02 April 2010 04:09:00 schrieb Marc Chamberlin:
running under SuSE11.2 x86_64, KDE 4.4.2
I guess you mean openSUSE 11.2 with KDE SC 4.4.2. "SuSE" and "KDE 4" are long dead.
ALSA
The Linux sound architecture.
ESOUND,
Not installed by default. Not needed. Dead.
JACK,
Not installed by default. Not needed.
OSS,
Not installed by default. Not needed.
ESD,
Not installed by default. Not needed. Dead.
aRTs,
Not installed by default. Not needed. Dead.
PULSEAUDIO,
Not installed by default. Not needed. Sucks.
GSTREAMER,
Codec architecture which just uses ALSA. No need to configure it.
PHONON
KDE's abstraction layer. Only of interest for developers. No need to configure it.
and tools like XINE, KAFFEINE, MYTHTV and MPLAYER which go off in their own directions... codecs to locate... I mean seriously folks, for a user to set up and configure a sound system this is getting out of control! How to choose? What to choose? What and how to configure? What to install?
Just live with the default install and don't try to mess with it. If you don't mess with it, it works fine.
Thanks Markus for your input, you have an "interesting" view on things.... But No, the default install did NOT work for me, and applying obvious/intuitive solutions did NOT work for me either. When I brought up openSuSE for the first time, I did NOT have any sound period. Yet I knew the computer hardware was working fine as I can dual boot to Windoz where sound would work. This left me with the daunting task of trying to figure out what had gone wrong. So off I went to try and explore and understand how the sound architecture works to see what and where something could go wrong. And THAT is when I, or any other user will come face to face with, and be appalled by, all these different layers, trying to read, comprehend and follow what each is all about. Bringing up my sound system, and some of the various tools/applications did not involve a single point of failure either. Instead I had to hunt and experiment by trial and error, ask lots and lots of questions, got lots of various qualities of answers, and encountered lots of confusion by others as well. It was a frustrating experience, sometimes I would get one piece working, like ALSA after I finally tracked down the problem to a misconfigured KMIX with hidden controls, only to have it fail on me again later, for mysterious reasons as I tried to get yet another tool (Amarok, Kaffeine yada yada) up and running... I am not arguing about the need to fix any particular bug, nor am I arguing for the need to remove any of these layers, I am sure each brings to the user/application developer unique and valuable features. What I am arguing for is the need for a much simpler model of the sound architecture that works as an integrated whole and can be presented to the user in such a way that is easy for him/her to understand and modify/control as needed. To accomplish that will require oversight by a small elite team of system architects. Most large software projects usually come with such a core team that can set guidelines and standards, and make sure subprojects and components comply with the overall architectural vision. IMHO this seems to be missing, over all of the releases of openSuSE that I have installed, (which goes back for quite a few years now) I am not seeing the sound system getting simpler, instead it is getting harder to understand... KDE is one of the major presentation layers to users, so to me this group seems like a good place to start such a discussion. It is from this desktop model (and others as well) that the user will be presented with the model of the sound architecture, and with the major tools on how to set it up and get various applications working. And I cannot say that I have ever felt that the tools such as YaST, or the Systems Managers of KDE have ever presented me with the right set of guides and models to successfully get these various sound applications to work in a straight-forward manner... Marc...
On Friday, April 2, 2010, Marc Chamberlin wrote:
I am going to bring this question to this mail list as I think it is more apropos to discuss it here...
I have just been through a proverbial nightmare trying to configure sound on my system and getting some of the various tools to work. Am running under SuSE11.2 x86_64, KDE 4.4.2. I will admit up front I am NOT a Linux guru but I am enough of a software engineer to recognize an overly complex architecture when I see one.
In all of my wanderings around the internet and maillists, I have discovered there as so many different layers, tools and configuration requirements that I have got to wonder if there is any sort of overall sound system-architecture vision in place and a GUI guide for users, being developed, to follow. We have got LAYERS and LAYERS of ALSA, ESOUND, JACK, OSS, ESD, aRTs, PULSEAUDIO, GSTREAMER, PHONON and tools
There are similar layers in other operating systems. In Windows, you may find several types of operating system low level drivers: WDM, ASIO, GSIF and WaveRT, frameworks and programming APIS: MCI, MME, DirectX (DirectSound), WKS, OpenAL... In Linux: ALSA is the OS driver model, replacing the old deprecated oss. The sound servers: esd(=esound), jack, pulseaudio, arts are system wide or session daemon/services, each one providing a set of features and APIs. Gstreamer, Mplayer, Xine, VLC are frameworks/applications on top of the lower layers, providing integrated multimedia, video and audio codecs, and not only sound. Phonon is a much higher level abstraction/simplification API, for Qt4/KDE. But I agree that the overall picture is not trivial at all.
like XINE, KAFFEINE, MYTHTV and MPLAYER which go off in their own directions... codecs to locate... I mean seriously folks, for a user to set up and configure a sound system this is getting out of control! How to choose? What to choose? What and how to configure? What to install? How to hook up and link together? And NO this is far from an automated process of any sort, and NO it is NOT transparent to a user! I KNOW I just went through hell just trying to get Amarok to play (and XINE and KAFFEINE as well) along with the projectM-pulseaudio visualization tool..
Integration is something that usually doesn't come out of the box when you mix different software components. When the components are in a heavy development beta stage, that explains the complexity you are experiencing. Integration of complex software components is the mission, and should be solved or alleviated, by the Linux distributions.
Just to cite my latest headache as an example, (and this was just one of many I had!!!) I found that running projectM-pulseaudio, which is suppose to connect to the pulseaudio sound server, randomly chooses which sound device on my system as its "connectHelper". (I had two, one of which was actually DISABLED in YaST but apparently that has no effect on projectM-pulseaudio!!) And projectM often chose the disabled sound device which left it unresponsive to sound output from any of the other applications.
This problem is known as the "PulseAudio routing to hell", well explained by Colin Guthrie here: http://colin.guthr.ie/2010/02/this-is-the-route-to-hell/ It is not the only problem regarding pulseaudio. The crackling noises, low audio quality for some sound drivers (mainly the integrated Intel HDA soundcards) is another one. I hope that some day these problems can be fixed in pulse or somewhere. There may be bugs at every layer of the sound infrastructure. [...]
Is there any hope in the near future that this will get simpler and easier for us poor audiophiles? Will appreciate hearing thoughts and NO I am not intending to throw rocks... Just that if I had such a hard time, I got to wonder how others are fairing and I think providing feedback and starting discussions is useful to engineers...
Marc...
I don't know if the problems are going to be solved in the near future. People that can help to solve the problems should listen and recognize that the problems exist, in the first place. The integration of all the pieces requires collaboration, and good faith. A factor that may also help is some funding by the companies leading the Linux distros. Integration is not going to happen simply throwing the pieces together and hoping a lucky miracle. Regards, Pedro -- To unsubscribe, e-mail: opensuse-kde+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-kde+help@opensuse.org
participants (3)
-
Marc Chamberlin -
Markus -
Pedro Lopez-Cabanillas