If you liked Pop Sweeet Pop, and found Dystopian Waltz too noisy, then this is for you: Perfect Song. Hopefully this time you'll get the message about commercial music deprived of creative and artistic substance. If not, then maybe you'll like the picture, which is probably a failed attempt at increasing the sales of this song since it's free, just like everything else on verdigris.mu. But you get the idea.

I used various bits of things to create this one, borrowing recordings from freesound.org: Crowd Laughters, Small croud yelling YEAH. However the final recorded "found sound" at the end of the song was made by me one morning.

The music was entirely generated with Splat, the source code and sound samples are available for download with this article. Only the voice was recorded acoustically, with some parts going through a small amplifier with a home-made valve distorsion. The solo in the middle of the song was generated randomly (so it's a bit different each time the song is re-generated) by choosing some notes within a fixed set and random variations of timing and levels as shown by this piece of Python code copied from qwirks.py with some extra comments to explain what is going on:

            # pick a random note from the given chord
            note = random.choice(chord)
            # alter the start time of the note (T is the duration of a quarter)
            t0 += random.choice([-T/4, 0, 0, 0, 0, 0, T/2])
            # alter the duration of the note as a multiple of T
            nlen = 1.0 + random.uniform(-0.25, 0.5)
            # alter the level of the note in dB, between -4 and -2 dB
            lvl = -3.0 + random.uniform(-1.0, 1.0)
            # alter the stereo pan of the note within +/- 3 dB
            pan = random.uniform(-3.0, 3.0)
            # intermediate variable with the resulting levels for each channel
            levels = (dB(lvl+pan), dB(lvl-pan))
            # call the method to generate the note with the resulting values
            self._note(t0, nlen, note, levels)
The drum sound samples were borrowed from the free software Hydrogen "Ultra Acoustic Kit", originally made with real drum recordings, and sequenced with Splat as already mentioned earlier. Disclaimer: while it's extremely accurate and consistent, there's no way you could possibly compare it with a real human drummer playing real drums (as previously discussed at the pub with the respectable Paul Richards).

Finally, the picture features a classic pin-up enjoying listening to the radio, now in public domain, and was edited with the free software Inkscape using a free font to add the song title. The final audio mix was made with the free software Audacity running on a Debian GNU/Linux operating system. While I'm at it, this website is powered by the free software Django framework running on a free software Apache web server (just like about 50% of the web servers on the planet) on a Debian GNU/Linux system. And by free, I mean free as in freedom since you get all the source code and permissions to modify it and find out what it actually does, not just as in "free beer" (which is also a nice thing).