signed 8 bit to "sign and magnitude" handler needed [PART ONE]

[Paul D. DeRocco]

> From: stephen barncard
> 
> Continuing on the "exporting aiff files from audioclips" 
> quest, I discovered that 'some' (about 10%) of the exported 
> files data would not play correctly. They sounded distorted 
> and loud at full scale. Digging in further, I discovered that 
> 8 bit AIFF files are not signed (due to their history), but 
> instead are ( I think ) "sign and magnitude" 8 bit.
> 
> So I need to convert the data for just the 8 bit files.
> 
> I know this is dumb ass simple, but I've been staring at 
> these numbers for too ling and need some fresh eyes. I am 
> *seeking a handler in Livecode* to convert numbers between 
> 0-255 (what I think is) two's compliment 8 bit to (what I 
> think is) "sign and magnitude" numbers between 0-255.
> 
> that is, I need to convert a number series that looks like 
> below (word wrap is inevitable.)
> 
> 00000000 80 7F 7F 7F 7E 7E 7E 7D 7D 7C 7C 7B 7B 7B 7A 7A
> 00000010 79 79 78 78 78 78 78 77 77 77 77 77 77 77 77 77
> 00000020 77 78 78 78 78 79 7A 7A 7B 7B 7C 7D 7D 7E 7F 80
> 00000030 81 81 83 83 84 85 86 87 88 89 8A 8B 8C 8C 8C 8E
> 00000040 8E 8F 90 90 90 91 91 91 92 91 91 91 92 91 91 90
> 00000050 90 90 8E 8E 8D 8C 8B 8A 89 88 87 85 84 83 81 80
> 00000060 7F 7D 7B 7A 79 77 76 75 73 72 71 6F 6F 6D 6C 6B
> 
> (this is the same sine wave going above and below center in 
> both lists)
> 
> to be converted to numbers that look like this:
> 
> 00000000 00 FF FF FF FE FE FE FD FD FC FC FB FB FB FA FA
> 00000010 F9 F9 F8 F8 F8 F8 F8 F7 F7 F7 F7 F7 F7 F7 F7 F7
> 00000020 F7 F8 F8 F8 F8 F9 FA FA FB FB FC FD FD FE FF 00
> 00000030 01 01 03 03 04 05 06 07 08 09 0A 0B 0C 0C 0C 0E
> 00000040 0E 0F 10 10 10 11 11 11 12 11 11 11 12 11 11 10
> 00000050 10 10 0E 0E 0D 0C 0B 0A 09 08 07 05 04 03 01 00
> 00000060 FF FD FB FA F9 F7 F6 F5 F3 F2 F1 EF EF ED EC EB
> 
> (this is a SSND chunk dump from an audio file that plays correctly.)
> 
> (part two will show the PHP code I tried to convert but failed)

Looks like offset binary to me, so x - 128 is the two's complement version.
If you need to represent the latter as unsigned bytes, the easiest way is
probably (x + 128) bitAnd 255.

-- 

Ciao,               Paul D. DeRocco
Paul                mailto:pderocco at ix.netcom.com