Asio documentation

[Keldon Alleyne]

Did you use any additional documentation for asio apart from the basic information from steinberg?

[Michael Tippach]

Keldon said:

Did you use any additional documentation for asio apart from the basic information from steinberg?

As far as ASIO: TTBOMK, there is none! Even the documentation in the ASIO SDK is rather unclear on this and that...

Everything else was to be learned the hard way, trying to get _my_ reading of the holy ASIO scriptures in line with the visions, revelations and readings of the ASIO spec that occurred to developers of audio software applications.

Quite an entertaining ride, given that it has already happened and I won't have to do it ever again ;D

[Keldon Alleyne]

Well I might do a better job of using DirectKS instead of ASIO; plus I have much more freedom!

Bottom line is that I want the lowest latency solution possible, and that might be the way to do it.

[Keldon Alleyne]

I have decided not to bother with ASIO for two reasons:
- 1: ASIO is proprietary and therefore means I cannot do any GNU work using it (just look at Audacity)
- 2: I think it is unfair that you have to go through so much trouble to use ASIO in a GNU project so I want to implement DirectKS with the intention to aid the open source community, although having said that the DirectKS examples should do that but it has not been done yet
- 3: Yes, the ASIO documentation is appalling, doesn't work well with GCC, and I would enjoy using the direct route and could even create my own adapter

The Windows DDK is still installing; but I should be able to find the appropriate dll name and function call from the help file so that I could [for example] begin to call it in RosAsm. I really don't want to have to install Visual Studio because it caused terrible problems with the debugger - although I'm not sure if I just didn't have some HDD corruption, I can't be sure. I can risk it again though, only takes an hour to install.

I'll codename the project TLTL (too little too late) since Microsoft claim to be providing low latency access in Vista's default sound stream, therefore eliminating the need for such applications.

Another thing that strikes me is that the documentation claims to be able to execute the DirectKS examples user mode; I will have to see this when the DDK installs - but I very much doubt it ^_^ If so then that is GREAT!

[keldonalleyne]

Odd, for some reason the board never did give me activation for keldon85!

EDIT: so it does execute in user mode, and although ASIO4All has all ASIO outputs go to the same sound output the DirectKS sample had to be modified to get the right output. Probably an attribute that the sample is not checking, but it's hard to believe that it executes in user mode!!!

And I don't know why but I had this down as a bigger task than it was, doesn't seem to be anything too difficult in contrast to the bigger picture.

[Michael Tippach]

If you choose WDM/KS because ASIO is somewhat "underdocumented", you're heading the wrong way, methinks.

All the DirectKS sample will let you do is play a 16bit/48Khz stereo PCM wave file on cards that support it and it does not work with WaveRT drivers under Vista either.

There is virtually _nothing_ regarding audio input (otherwise you will start to wonder e.g. why some audio devices would insist on a nonzero dwChannelMask in the WAVEFORMAT_EXTENSIBLE, meaning why would an input request a speaker configuration?

Thinking about adding WaveRT? The MS web site has a document about WaveRT, which is outdated and the only way to get the missing information "legally" (i.e. from public sources) is to pry it from the WDK headers and apply some guesswork plus experimentation.

Not to discourage you, but ease of use, compatibility and documentation are certainly not the right arguments in favor of WDM/KS over ASIO.

[keldonalleyne]

Hmmm, WaveRT looks great and is definitely a consideration for the future. But even NI are waiting for more hardware support of its specification. By then Microsoft should have improved their documentation.

As for WDM/KS vs ASIO there is just so little information available that a decision is difficult. Learning as much on both may be the best option! I would want to implement recording after around 3-5 years (maybe sooner if progress permits) but it is not a priority feature, I just want low latency output.

I know of other problems such as hardware reporting compatibility for modes it does not really support too, but maybe open sourcing could provide fixes and patches for that