FreddyV wrote on 2022-02-07, 12:45:
The best would be to do a Disney sound Source like, With a bigger buffer and increased output frequency.
A nice improvement over the original DSS would indeed be the ability to adjust the FIFO buffer size and the output frequency. The original DSS had a fixed 16-byte buffer and a non-adjustable output frequency of around 7kHz. This resulted in sound effects sounding somewhat lower-pitched and slower in some games that officially supported the DSS, but didn't bother to down-sample the sound assets, which were often recorded at 11kHz. Being able to tweak the buffer size and playback frequency would help with these games, and would be a nice improvement over the original design overall.
It is also possible to do a Parallel port "Covox" with DMA Support, as the parallel port support DMA.
Aside from the required complexity for such a solution, it would limit compatibility to parallel ports that support ECP. Also, there is no existing software support for such a mode.
A compromise would be some kind of enhanced mode hat would use the IRQ line of the parallel port as a more efficient way to inform the software that the FIFO buffer is full (or empty). (As opposed to having to poll the state of one of the status pins, which Is the case with the original DSS.)
But before we blow up the scope too much: just a single device with software-switchable Covox speech Thing and Disney Sound Source compatibility would already be wonderful. With adjustable FIFO buffer size and playback frequency being low-hanging fruit w.r.t. improvements over the original standard.