Re: wierd behaviour of lm741
No.
You need to bias the inputs using resistors, for example. Use a divider for the + input: say 100kΩ / 100kΩ. The - input should get the feedback: a resistor from the output, let's call it R1 and a resistor in series with a capacitor to ground; let's call this resistor R2 and the cap C.
The + and - inputs will now be biased at about half the supply voltage, that is 2.5V. The + input is obviously biased by the divider. The - input is biased from the output through R1. C ensures R1 and R2 do not form a divider in DC, just AC. Hence, the - input will be at 2.5V, and so will the output.
Now you apply the signal to the + input, through another capacitor, which will prevent your signal source from shorting to ground the DC voltage established by the divider.
The gain of the circuit is, of course, G=1+R1/R2. So if you need a gain of 10, choose for example R1=100kΩ and R2=10kΩ.
The two capacitors should be chosen large enough for the lowest audio frequency to pass through them: 1 to 10µF should do. Their + side is towards the inputs of your opamp.
The output should most likely be AC coupled, so use another capacitor, say 10uF, connected with its + side to the output of your opamp.
So, to answer your question, when you have no audio signal, the output will be at 2.5V DC. The AC signal will be zero with no input, that is, you get no audio out.
Good luck!
Added after 23 minutes:
After a little thinking I decided a picture is worth a thousand words.