Re: Any more graceful way to saturate adds?

On Apr 13, 10:27 am, "Rod Pemberton"  <spamt...@[EMAIL PROTECTED]
> wrote:
> If Jim needs to distinguish 255 from overflow as Terence noted, perhaps
this
> is sufficient:

I do not need to distinguish 255 from overflow.

The quick background for this is that it is the clamping function for
adding a signed 8-bit integer to an unsigned 8-bit integer.  The
function is an adapted ADPCM routine; whereas most ADPCM works on
signed 16-bit samples, I have a need for it to work on unsigned 8-bit
samples.
The inner loop of the routine has a previous unsigned sample, and a
new signed sample, and needs to add them together to get the new
sample.  The result needs to clamp at both 0 and 255.

With this in mind, what is the best solution, assuming the previous
unsigned sample is in DL and the new signed sample is in AL?