problem with understanding audio wave

Hey!

I have problem with understanding result of recorded audio.
I connected mic with simple non inverting op-amp, dc bias and arduino. So my signal is at 2,5V when nothing happens and changes a little bit when I generate sound. I record audio using TMRpcm and save it on sd card. Then I play it using Audacity.
Here's my problem. When I record sound I also measure voltage changes and I can see that these changes are very small, something up to 0.5V. I can hear sounds on Audacity but they are quite weak. So I thought that I can change amplification a little to get better, stronger sound. I thought that I have 0-5V range so changes of voltage should be really few times bigger to use that range. But when I change amplification even not much, so that voltages changes are something between 0.3-1V I get something like this:
result

So it's going from the bottom to the top and back. I could image that happening if voltage was changing from 0V to 5V but that was not the case.
Please help me understand why it is behaving that way and how else I can strengthen signal to get sound more distinct from noise.

So I thought that I can change amplification a little to get better, stronger sound. I thought that I have 0-5V range so changes of voltage should be really few times bigger to use that range

Yes that will make the sound better. But you seem to be not doing it. So we need to see a schematic of your setup. With suitable amplification you should get a full scale sound recording. It would seem you are not.

I am not sure that libiary records sound that Audacity can play.

So I thought that I can change amplification a little to get better, stronger sound. I thought that I have 0-5V range so changes of voltage should be really few times bigger to use that range. But when I change amplification even not much, so that voltages changes are something between 0.3-1V I get something like this:
result...

The Arduino doesn't have an analog output. TMRpcm isn't really PCM or analog. It's [u]PWM[/u] switching between zero and +5V. If you low-pass filter it (and high-pass filter it to take-out the bias) you can get analog. But, I assume it takes a really good active filter to get "good" analog.