Quote:
Originally Posted by omasciarotte /img/forum/go_quote.gif
It's designed to minimally process the data read from disk. That said, one generally cannot or should not simply read the data from disk and send it to the I/O. There is usually transcoding, like floats to fixed conversion and redithering, before the data can be passed to the output interface.
|
If you are going to be manipulating the data in the original audio file you will certainly need to do things like transcoding, fixed to float (and back) conversion, re-dithering, etc.
The problem is...a bit-perfect computer music playing program should not be doing any of that (not one that I want to use, at least). It should not attempt to change one single bit of the original data (16-bits in...16-bits out, untouched). That means that transcoding, float conversions, and re-dithering are unnecessary.
Again, if you want manipulated audio data (something that the people who created the original CD you purchased did not feel was appropriate, or else they would have done it themselves), you can achieve that with just about any music playing program, but it might require some experimentation to get the combination that sounds right for one's own ears. Perhaps Amarra provides a better user interface for someone that is experimenting the digital signal processing/manipulation than the existing programs (Foobar2000, iTunes, WinAmp, etc.).
And, yes, if you aren't messing with the original data before sending it onto the driver, you have some pretty simple buffering requirements (async reads into a buffer, when it reaches a reasonable level begin outputting, read more data into the buffer when there is sufficient room but before starvation occurs...pretty basic stuff). Now, if you need to modify that data and put the output elsewhere, and that modification is CPU intensive, you may need to get a bit more creative.
