SCST86@LIVERPOOL.AC.UK (03/12/91)
Hi There, I'm experimenting with sample compaction and decompation. I've developed an algorithm of my own but the quality it gives is not very good (pretty aweful actually B-)) I was wondering if anyone in NetLand had any articles or advice that they could mail me. I'm dealing with 8 bit samples at any rate you care to mention. The idea is that I take data from a slow stream (like a floppy) and play it back in real time. Thanks in advance. Mail to..... Jason Timmins My Card.... +-------------------------+ Dept. Of Computer Science, | | University Of Liverpool. | Jason Timmins | JANET: SCST86@UK.AC.LIV.IBM | (I can waggle my ears) | OR: SCST86@UK.AC.LIV.UXA | | TELE: (UK) 0860 600 758 +-------------------------+
rjc@geech.ai.mit.edu (Ray Cromwell) (03/17/91)
In article <91071.103504SCST86@LIVERPOOL.AC.UK> SCST86@LIVERPOOL.AC.UK writes: >Hi There, >I'm experimenting with sample compaction and decompation. I've >developed an algorithm of my own but the quality it gives is not very >good (pretty aweful actually B-)) > >I was wondering if anyone in >NetLand had any articles or advice that they could mail me. > >I'm >dealing with 8 bit samples at any rate you care to mention. The idea is >that I take data from a slow stream (like a floppy) and play it back in >real time. > >Thanks in advance. > >Mail to..... >Jason Timmins My Card.... +-------------------------+ >Dept. Of Computer Science, | | >University Of Liverpool. | Jason Timmins | >JANET: SCST86@UK.AC.LIV.IBM | (I can waggle my ears) | > OR: SCST86@UK.AC.LIV.UXA | | > TELE: (UK) 0860 600 758 +-------------------------+ I don't know whether this would help or not, but there are several programs on the Amiga that play back samples in real time from floppy up to speeds of about 14khz. Any faster and it starts falling behind while the floppy catches up. On the Amiga, sound samples are encoded in the standard IFF FORM 8SVX using fibonacci delta compression. I think this method compresses samples by 50% by encoding 8 bit samples as 4 bit changes from the last sample. The way the player program works is by double or quadruple buffering samples read from the disk. The Amiga's Audio and Disk are driven by DMA, so you can load and play sounds with little processor intervention. The major overhead is the seek delays on the disk. Try ftping from abcfd20.larc.nasa.gov in directory /incoming/amiga. Look file a file called 'AGMSPlaysound'. It contains C source code for the player. You may find some help on encoding methods and buffering techniques in the source. Hope it helps... BTW, I know the techniques use in the source work. I've used it to play an entire disk long sample of the theme to 2001, and a long sample of Madonna's VOGUE. (about 45 seconds worth.)