[flac-dev] [PATCH] doc: specify that quantized LPC shift must be non-negative

[lvqcl]

Tristan Matthews wrote:

> Refs http://sourceforge.net/p/flac/bugs/424/
> -				Quantized linear predictor coefficient shift needed in bits (NOTE: this number is signed two's-complement).
> +				Quantized linear predictor coefficient shift needed in bits (NOTE: this number is signed two's-complement and must be non-negative).


Maybe it's better to explain why?  At least something like this:

     "(NOTE: this number is signed two's-complement; but, due to implementation detail, it must be non-negative)"