How to encoding linear signal to logarithmic signal with less bits?

[tompak]

Hi,

I am new to encoding signals ... However, I have a problem that I am not sure how to solve:

- I have a 12-bit linear signal, which I want to encode/compress into the 6-bit logarithmic signal representing 0:-0.5:-31,5 dB from max amplitude. How can I do that?

- When received, how do I decode/map it back?

Best regards

 

[FvM]

You are asking about AC or DC signal?

 

[andre_luis]

I have a 12-bit linear signal, which I want to encode/compress

The nature of a signal tells a lot about the suited technique to encode it. As you are referring to a linear signal, depending on the slope maybe even a compression would be not necessary.

how do I decode/map it back?

Honestly, I do not believe that a logarithmic approach could feature the lossless compression.

 

[FvM]

I don't hear the OP requiring lossless compression. Logarithmic compression is used for some applications, e.g. good old telecom µ-law. If performed on a digital input signal, it's simple number mapping.