asbak wrote:Adhering to poorly worded documentation like this is what got us into this mess in the first place. Yeah, audio will work with a standard kernel and no tuning and you can even record, but so what?
If bad, outdated, documentation is what you are complaining about, than I would suggest to double check few of your statements:
asbak wrote:- Standard kernels suck for low-latency audio
- A PREEMPT or low latency kernel is highly recommended and WILL outperform a standard kernel and works great for general purpose use and is easy to install
A low latency kernel is
a standard kernel compiled by selecting appropriate parameters at compilation time. A RT kernel is also patched, so it is not a standard kernel. At some point the RT patch will be included in the standard tree, so one will be able to select the RT preemption model at compile time, without patching. RT will become a feature of the standard kernel. What is reported by khz is in essence correct information, as whatever parameters you choose to compile a standard kernel with, you get a standard kernel out unless you apply patches too. Beside, I do not understand the confrontational tone, since it proves your point that in most cases RT is not necessary. On this note:
Whilst I agree that documentation is old in many places, and a lot of outdated popular wisdom circulates from years back when things worked differently, I don't understand the arrogance by which you expose your points. You can be right also by not being so confrontational. This is a comment about the language you chosen, not about you as a person. I believe it will help you having a constructive exchange in a public forum if you choose to use a language that doesn't sound that abrasive. Just my humble tip. Take it or leave it.
asbak wrote:The documentation needs to be re-written to make it specific, accurate and relevant and to put an end to the masses of disinformation out there which isn't doing the Linux audio cause any favours.
I am myself guilty of not experimenting enough, study enough, and contribute to the documentation but please, unless you are already doing so, if you have adequate reproducible evidence about what kind of configuration is effective, and about what bits of information are outdated, then contribute to the documentation.