This archive contains allophone generated by the SPO256-AL2. The files contains signed 8bit PCM samples. The PWM digital out pin has been captured, and PCM samples calculated from 4 properly aligned consecutives PCM pulses, DC offset has been removed.
The original allophones were converted to raw files to be played at rate 48k, 1 channel, 8 bits , signed integer
For example, to play ar allophone with sox : "sox.exe -r 48k -c 1 -b 8 -e signed-integer -t raw ar.wav -d"
The following table contains the length of each allophone as well as a "blocking" time.
The blocking time is the time that the allophone will take before a new allophone will be collected from the input buffer.
That is, when a allophone is written to the SPO, if a new allophone is written during the blocking time will cause the LRQn output to be high until the end of this blocking time.
Example 1: single allophone write
t=0 : write /OY/ to SPO
-> SBY goes low, LRQn goes high
t=0+tLRQn :
-> /OY/ get processed by SPO, input buffer is free, LRQn goes low
t=2945+tLRQn : End of allophone
-> SBY goes high
Example 2: two consecutive allophone
t=0 : write /OY/ to SPO input buffer
-> SBY goes low, LRQn goes high
t=0+tLRQn : /OY/ allophone is beeig transfered and processed, start of /OY/ blocking delay
-> /OY/ get processed by SPO, input buffer is free, LRQn goes low
t=50 : write PA1 to SPO
-> LRQn goes high, input buffer in full
t=2117+tLRQn : End of blocking delay, input buffer is transfered
-> LRQn goes low
t=2945+tLRQn : End of /OY/, PA1 is getting processed, start of PA1 blocking delay
t=2945+64+tLRQn : End of PA1
-> SBY goes high
Note : When a allophone is played, its last part will play in loop unless it is followed by a pause. By looking at the allophone waveform, it seems that this could be emulated by looping the last 92 samples.
