Maybe this is what you already have:
You can always add a positive bias to your audio signal and use a pot with 1024 taps.
jim-p First- I appreciate your interest, and trying to help. Let me comment on the circuit diagram you presented. Both stages in your circuit are non inverting op amps. In my situation if the gain is increased, it increases the output stage in mili-volts. It also increases the point where the circuit goes off. Meaning you can still hear the audio when it should be no sound. The final stage of my circuit is a instrumentation amplifier. There is no problem increasing the output with one resistor, but when I do that, it creates more audio in the off position. I am aware of Pots with 1024 taps, which would make transition between steps much less noticeable. I am feeding an audio signal to a GMR sensor. There is a non inverting OP amp feeding the audio signal to the GMR sensor, then out of the GMR to a instrumentation amplifier. Really simple circuit. I don’t see any place in the circuit where using a Digital Pot with Arduino would solve the problem. Furthermore: One reason I was using the GMR circuit is because of the low current draw, which makes it ideal for battery operation. If I add a digital Pot and a Arduino I am adding more current. Meaning it would run a battery down much quicker. In that case a small wall wart power supply would be needed.
When you said " I can get my GMR audio circuit to go to zero exactly what did you mean? If it's zero and you add a gain of 100 it should still be zero.
jim-p Unfortunately when I get my circuit to go to zero, if I add gain then the circuit will not go to zero with the added gain. If I drop the gain so it goes to zero, then the wide open does not go all the way wide open.
I don't what your circuit is but obviously something is wrong if adding gain causes the attenuation to change. My circuit does require +/- supplies but you can achieve very large signal attenuation and the x2 gain stage does not affect the first stage.
jim-p I am probably not explaining things enough, and talking about a circuit that is not displayed. To understand, think of it this way. The GMR sensor reacts to a magnet moving toward it or away from it. The output of my circuit not only depends on op-amps, or in my case a instrumentation amplifier, but what the sensor is sending to the input. For example: If I get a wide open output of 100 mV in the off position I still have 5 to 8 mV. If I lower the reading of the off position to zero. Then the wide open output comes down from 100 mV an equal amount. If I do what you suggest and add gain to the output, say make it wide open 120 mV. then in the off position I get 25 to 28 mV. It is about putting the output in a range of zero to a wide open value of 100 mV. Maybe setting that rage would be possible with a digital pot on the output of the circuit.
Audio volume in instruments (organs) has been done for decades with an LDR and light bulb.
That simple opto-isolated principle seems a lot easier than some digipot.
Leo..
Well in simulations with my circuit I can reduce the input signal to a few tens of microvolts. You will never get it to absolute zero.
jim-p That is my same results. Since it is audio and those microvolts get amplified, I can hear sound when the sound is supposed to be totally off. Thanks.
Hi, @KeithHilton
Can you please tell us what your overall project is?
What is the source of your 100mV peak signal?
Why 100mV and not 1V or 5V?
Thanks... Tom.... 
I guess I don't understand what you are doing. Are you taking the output of my circuit and amplifing it even more?
Are you trying to drive a big loudspeaker?
jim-p and TomGeorge the circuit I just posted should answer your questions—”What is my circuit?” This circuit is copied from the NVE website, and the ALTOxx sensor is a GMR that NVE corporation makes. I have “NO” project–this is simply a hobby learning how to use Arduino. I am simply feeding in a simple audio signal, instead of the battery input, and trying to control it with a magnet. The OP amp is arranged as non inverting. You can change the gain by changing the resistors. If your audio signal is not strong enough you can put a OP amp pre-amp to feed the audio signal to the GMR. If you don’t like the OP amp arrangement in the diagram, try using a instrumentation amplifier. If you have any more questions about the GMR or the circuit, you can contact NVE corporation and they have a person that answers Emails. Hope this helps.
We term any circuit and controller and component that is being discussed a project.
You are finding that the lower the source signal the greater to tendency to amplify noise as well.
Tom....
I don't because the attenuation will depend on how well mached the resistors are. For a inamp the attenuation will depend on the Common mode rejection ratio of the amp.
You would also need to trim out the offset voltage.
It's an interesting application.
TomGeorge, can’t my project be: To learn more about Arduino and how to code? What is wrong with learning how to use Arduino. Please explain that to me TomGeorge? I really want to know why you are trying to be the police? All I am wanting to do is learn more about Arduino.
TomGeorge, the source of my audio signal is a small sine wave generator. I chose 0 to 100mV to keep the signal small and more manageable. Hope this information helps.
Yes it can.
Can you post a schematic of what you are experimenting with at the moment/
Tom....