Pulse Audio Dynamic Volumes (Take-2)

So today I tried using PulseAudio Dynamic volumes again, and the quick conclusion is this: They are bad, so very very bad.

Previously I had been cruelly subjected to Dynamic Volumes on with my Desktop through it's several speakers and sub-woofers. Now I tried it on my laptop using headphones / earbuds. The experience was actually worse.

On the surface Dynamic volumes seems nice, because in theory it should raise the total volume of the system as necessary. There is however one drawback. It raises the MASTER VOLUME for the entire system, which means all other applications and any subsequent applications will inherit that volume level.

So say you raise the volume on a piece of music you are playing (in Amarok for example), which is nice. They you load a video in your VLC. The video's audio is louder to begin with AND the VLC volume is set to the previous amount, which was 140%. Suddenly your ears are blasted with maximum audio volume. Painful and wrong.

Seriously. I'd fucking sue a company for earing damage if this wasn't free software with absolutely no warranty. (Okay, I wouldn't sue an open-source free company, even then, because I love open-source, but you get the point.)

PulseAudio's Dynamic Volume Levels are Really Bad

Problem

So today I got to enjoy some seriously awful pain from the developers of PulseAudio. Now I kind of like PulseAudio, when it works. Which hasn't always been consistent.

When using Fedora, I used my scroll wheel over the Amarok icon to increase the volume. This not only raised the volume of Amarok, but it also raised the master volume for my ENTIRE SYSTEM.

If one application requests it's volume level be increased, PulseAudio raises that programs volume and the Master volume... and leaves them at that level.

Now every new sound made by any other applications, which all fall under the Master Volume, are at potentially high levels.

There is also the situation where you manually cannot LOWER the volume if some other application is requesting it be higher. I found myself in that position today where I madly scrolled down over the Kmix icon, only to watch the volume flicker between 93-94% maximum.

Apparently this is not bug, it's working as designed. But I can personally attest that it is designed WRONG! https://bugs.kde.org/show_bug.cgi?id=288594

Solution

The solution is simple. Edit (or create) the following file.

~/.pulse/daemon.conf

Make sure the following setting is enabled:

flat-volumes = no

Save that file and reboot. (I don't know if you have to reboot or if there is an easier way.) If you need help knowing how to edit a file under Linux, here is my simple guide.

Ranting on the Side

Yes, I'm angry. This is awful  and this is exactly the type of bullshit that gives people real reasons to legitimately criticize Linux. I shouldn't have my computer system suddenly explode with overwhelmingly high sound and STAY that way without my control.

My speakers are physically set to a very high volume because they are mounted around my room in stereo and not easy to access physically. I've controlled them digitally for over a decade. I will not have the freedom and power to control them with the flick of a keyboard button stolen from me because of this ineptly implemented feature.

I'm not saying you can't have a different paradigm for how audio should work, but god-damn-it you should not break (let alone shatter) the existing world to experiment. And that's exactly what this is: an experiment gone wrong. I honestly cannot see a situation where this could even possible work (as it is currently) for a user like me and thus the case for 90% of the rest of the world.

PulseAudio - A new system to solve many problems

Overview
What PulseAudio is, and why it is great for Linux.

Forward
Someone once asked me why PulseAudio is being pushed and I only had some vague ideas why. This article briefly mentions some new reasons I did know. I thought you might be interested:

http://www.phoronix.com/scan.php?page=news_item&px=NzIwMg

I have been reading up on PulseAudio and like I said before, if it becomes the standard (and I think it will) it will be great and solve a lot of the problems in Linux. Sufficient to say, if PulseAudio can do all that I have ever seen listed on websites for features, it will be the end all of sound systems for Linux. :)

Just to name a few features PulseAudio has that I like:

• Support for a far wider range of audio devices ( Bluetooth/Apple Airport)
• Flat volume support (similar to Vista's audio controls)
• On-the-fly reconfiguration of audio devices
• Native support for 24-bit samples.
• It will run on Windows, both natively if desired or through Cygwi.
• Support for allow networked sound (between Windows and Linux)
  

These last two two features are things I have personally desired from a sound system. Previously I had never successfully enabled transporting of sound across a network from a Windows Machine to a Linux machine. (The occasion for wanting to do this was that I had a Windows machine in my room, running games which I was controlling from my Linux desktop via Synergy. I did not want to setup a second pair of speakers to hear the games. I wanted to have the audio from that computer piped into my main desktop speakers.)

About two years ago, PulseAudio was rather unsupported (maybe even broken), at least in Ubuntu. Maybe it was the programs... or maybe it was the kernel. I don't know. It could have been both. But now most of my programs run PulseAudio just fine. I remember the annoying problems I would have with OSS audio in the past. This transition to PulseAudio is much less of a problem for me.

I suppose a user must need to be somewhat advanced to experience the occasional difficulties with software audio protocol conflicts and difficulties. I can safely tell you that I definitely experienced more problems with audio 6 years ago when I began to seriously use Linux. I also experienced more when I was using a wider variety of software.


In the past 6 years I have been watching, there has been larger move towards supported ALSA in most applications, and now support for PulseAudio seems to be coming along nicely. The best part of PulseAudio is that, to my understanding, as long the application uses ALSA, ARTS or ESD, all of those are wrapped under the PulseAudio umbrella.

So if PulseAudio is actually working on your system, the system is backwards and forwards compatible, and all sound play nicely together. Well done, huh?

Of course, there is the issue that PulseAudio will have more overhead than systems like ALSA, which of course has more overhead than OSS. Personally though, I seriously doubt this is going to even be noticed on desktops with Dual-cores and even more so as Quad-cores become commonplace. Also, we are not talking about a lot of overhead here, especially when the system is not doing much.