If you know the structure of the URL you are scraping you should be able to break-off the useless part of the string.
For example, if you know that the URL follows the form:
http://example.com/1392/Music/1392/Shahrivar/21/{Artist}-{Album}/{Track Information}
Roughly, I think the following would allow you to extract the information you want from a single link:
void Main (string[] args)
{
var example = @"http://example.com/1392/Music/1392/Shahrivar/21/Avicii%20-%20True/01.%20Avicii%20Ft.%20Aloe%20Blacc%20-%20Wake%20Me%20Up%20(CDQ)%20%5b320%5d.mp3";
var parts = example.split('/');
var album = parts[7];
var trackInfo = parts[8];
var trackParts = trackInfo.split('-');
var artist = trackParts[0];
var trackTitle = trackParts[1];
Console.WriteLine(trackTitle);
}
Here I am splitting the URL by '/', which is a messy solution, but for a simple case, it works. Then I am finding the index within the tokenized string where the desired information can be found. once I have the track information, I know the convention is to separate the Artist from the Title by a '-', so I split again and then have both artist and title.
You can refactor this into a method which takes the URL, and returns an object containing the Artist and song title. After that, you might want to use a string.Replace on the '%20' with ' '.