It's not worth $39, certainly not $49. I haven't even used it, just looked at your web page: but I have a strong professional background in this area, so I know what it does.<p>There are many shareware programs which do this, and others are free: this goes double for plugins, and there's a better than-even chance that a musician who likes computers already has some kind of multitrack software which accomodates plugins. They likely also have a hardware audio interface with a special low-impedance input for plugging in a guitar and got equivalent software for free. So that's black mark #1 - you're not offering much new functionality.<p>Black mark #2 is your interface. It is beautifully clean, I applaud that. However, while an audio waveform is wonderful for engineers, it's not so much for musicians. It might be better to generate a low-resolution spectral display, by applying (say) a 256-band FFT which won't cause any computer to break a sweat but would give a more musically meaningful presentation, and set you apart from the competition. As you have already implemented pitch-independent time stretching, you probably know how to do this.<p>Personally, I would not waste so much space on the album art either, but that's me.<p>I'd like to see you add some more features, which wouldn't break your interface. While decomposing a a piece of music into polyphonic pitch information is hard (see <a href="http://www.celemony.com/cms/index.php?id=635" rel="nofollow">http://www.celemony.com/cms/index.php?id=635</a> for the technology leader), deriving the 8key* of the music from a relatively coarse FFT analysis is not so difficult. Having the key signature, and indeed the bpm, appear on the left side or over blocks in the audio display should be easily achievable and would add a lot of value.<p>Another thing you could do is offer a button to extract or remove the top (usually vocal) melody line. This is easy: take a stereo file, invert one channel, and add it to the other. this will give you a (mono) karaoke track, since the vocals are almost always panned to the center. You'll lose the kick drums and bass too, but you could limit to the bandwidth of the human voice. Invert and add the karaoke channel to both sides of the stereo waveform, and you'll get <i>only</i> the center, allowing the musician to either copy or accompany the lead vocal and not much else. There's nothing more complex than a multiply-add operation going on here ((waveform.position.leftchannel.samplevalue * -1) + waveform.position.rightchannel.samplevalue ...etc.) so you should be able to do this in realtime with virtually no performance hit.<p>Finally, consider taking the live stream of input from the microphone input (eg the guitar), FFting it, and doing a loose correlation with the FFT of the playing track to derive a little 'accuracy' meter. Of course there will be much more audio information in the track unless it's also an acoustic guitar solo, but the frequency response of the live input should be a subset of this as long as the musician is playing in time. When they start to drag or lead the backing track, the energy of individual frequency bins will exceed those of the backing track.<p>Sorry for being so harsh, but I do think you're asking too much for something that most people will perceive as little more than a 'hello world' audio program. If I were you I would protoype the functionality in reaktor or some other visual-programming tool, add some more useful and unusual modes of display or operation as above, and then bring that back to your clean interface.