Then What about Crosstalk distortion? Do't I need to consider it if just use a Tl074?
crosstalk is not distortion. It is a signal on one channel that leaks into another channel.
Simply look at its datasheet. Its crosstalk is at -120dB which is about 1/1000th as much as you can hear if only one channel has a signal.
But I guess the gain is 474?!
The first opamp has an inverting gain of 100k/4.7k= 21.3 times.
The second opamp has a non-inverting gain of (100k/4.7k) + 1= almost 22.3 times.
The third opamp has a gain of 100k/100k= 1.
The total gain is 474. I don't know why I said 505.
The difference in gain between 474 and 505 is so small that you won't hear any difference.
So those values should be good to be used for audio?
The cutoff frequencies of the three capacitor add together which reduces the bass response.
Are you telling that I will have almost -9dB at the lower thereshot (i.e 20Hz)? If so, it would not be so good, right?!
No deep bass.
What will happen if it takes long time to be charged?
Then the amplifier will not work for 5 seconds after it is turned on and will not work properly until about 23 seconds after it is turned on.
Why C1 should be non-polarized too?
I had the circuit mixed up since you have an offset voltage. Only C2 should be non-polarized.
There you just suggested not to use high values feedback resistor, but now you are telling that I CAN use 1M resistor for the feedback??!! If I do so then There would not be any problem with anything??!! Is that due the BJT input transistors for AD826?
Yes.
The AD826 has a fairly high input bias current for its good very high frequency response. Then its input impedance is fairly low and reduces the gain if a high value feedback resistance is used.
But the TL072 has Jfet inputs that have almost no input current so its input impedance is extremely high.