About this blog
Background - Fine-Tuning Linux for Audio
Ever since I first created this web site roughly back in 2008 I have been researching ways to fine-tune Linux in various ways. This lead me to go to the best sources of information and knowledge developed by professionals.
After nearly a decade of reading manuals on how to tune Linux from RedHat, IBM, Oracle and others I came across many kernel level, network and audio related adjustments that could be made after the operating system has been compiled and installed.
After years of trying different values, settings and values recommended by the professionals and values I ended up liking through empirical listening tests I quickly found out that the old way of thinking "If some is good, more is better" does not apply to audio. For example using brute force methods of applying the highest priority didn't always sound best, in fact it most often sounds worse.
In addition, I learned that there is a delicate balancing act at play with tuning the kernel for optimal results. This meant I had to try a lot of variables to listen for what really affects sound quality and what kernel level tuning plays well with others.
Now after nearly a decade, I believe I have narrowed down a small set of kernel level variables to adjust to give optimal results for all types of music and in all listening situations. In fact, on Debian based systems these values work for all music players equally well.
Later as the SBC;s were introduced a new set of challenges presented itself. The old tuning used for desktop x86 processors and computers didn't sound the same on a Raspberry Pi. The meant I had to do a little more fine-tuning for this type of processor, with limited RAM and CPU power. Then there was the challenge of being able to apply these values to different types of Linux, such as Arch Linux and Tiny Core Linux for example. I found these two require slightly different settings. Too aggressive and the sound would be harsh and irritating, not enough and there was not change. What I was after, was a sound that offered an extension in the upper and lower octaves, a better-focused sound stage, which offered much better vocals and separation. It was this focus that would be the Hallmark sound which made the improvement immediately noticeable.
- Improving the priority of the Audio group
- Improving the audio thread priority
- Improving the latency of the Operating System with Kernel adjustments
- Improve network latency
See my Repository on GitHub for an automated process to apply these tuning variables.