# Dealing with quantization level limiting dsp

There is a analog signal in the form as below:

$$f(t)=A \sin(\omega(t) t+\phi)+B \sin(\omega(t-t_1) t+\phi_1)$$ Where $$B<<A$$

We are going to calculate $\phi, \phi_1, t_1$,using dsp. But since $B<<A$ and due to quantization level limitation, we can not extract data precisely. Is there any suggetion how to deal with that?

Thank you for your kind replies

• I'd suggest trying non-uniform quantization, increasing the number of quantization levels in the range $(-B,B)$. – MBaz Nov 27 '15 at 16:05
• For non-uniform quantization, the quantization noise decreases for low values, but its sabotaged for higher values. – Kami A Nov 27 '15 at 18:38
• That's correct, but the higher values have more power, so they can tolerate more noise power too. Just something to consider -- this may not be the solution you need. – MBaz Nov 27 '15 at 19:20
• How w(t) is different from w(t-t1)? Do you have a combination of CW wave with chirp? – Moti Nov 27 '15 at 23:38
• f(t) is a combination of a frequency-moduled signal with a time domain shifted version of itself – Kami A Nov 28 '15 at 9:55