DIY Digital synthesis, anyone?

Tony Clark clark at
Wed Oct 30 15:20:44 CET 1996

   Actually, the 'C30 TI DSP chip is a full 32-bit processor.  While the 
'C50 is only 16-bit, I don't know about the 'C40.  But in any case, if 
bit resolution is what you want, the 'C30 clearly has the advantage.  
What is neat about this is that (since most people will only use 16-bit 
D/A converters) you can push two channels of audio through the processor 
bus at the same time.
   It's funny, but what I've noticed is what I said originally, you'll 
get a couple of camps of people who like one kind of DSP or another.
   I will credit Motorola for having the best architecture of any chip 
manufacturer.  I first learned microprocessor programming on a 6809.  
Their microprocessors rock.
   On the issue of programming, I have no idea how the Motorola ones are 
set up, but the 'C30 does take an effort to optimize code on an assembler 
level.  That's mainly due to its pipelined architecture.  With 4 levels 
of instruction execution, it is possible to have the 'C30 perform a 
Fetch, Decode, Read, and Execute on data all in the same clock cycle.  In 
this scenario, you can crunch data at the clock speed of the DSP chip.  
   Of course you don't have to program the DSP to this level of 
performance.  So that can save a person a bit of assembler nightmare.  
Everyone's individual preference and performance will vary.  :)


I can't drive (my Moog) 55!
Tony Clark -- clark at

> In my opinion Texas DSPs are awful to program (with assembler) compared 
> to Motorola's DSPs. Also most of Texas DSPs are only 16bit while Motorolas
> are 24bits. This is why Motorolas are more often used in audio applications.
> I would choose Motorola, but I agree with you that for filtering Texas DSPs
> are great because they are designed for that.
> Pekka Martikainen
> cubase at

More information about the Synth-diy mailing list