Hello guys,
I've set up an iOS application for OTA fingerprinting using the same audio recording methods/settings as the EchoPrint sample application. I have not yet tried the sample app, but my app is working in terms of communicating with the API, creating and sending fingerprints, retrieving track information, and so on.
The problem is that it very rarely works. I'm testing both in the iOS Simulator and on an iOS 5 iPod, using songs I've made sure exist in the public database. Re-playing the recorded snippet tells me it records fine -- it has background noise on the iPod but Apple's voice memo app has the same noise, the recordings are seemingly the same quality. I'm testing at a variety of volumes (and holding device from different distances from speakers), making sure there's no sound in the background, and I've tried recording snippets of 20, 30 and 40 seconds as well as changing the recording settings (sample rate, stereo/mono etc) -- nothing seems to help.
I've been able to identify a few songs successfully, and only one song has been consistently identified (yet it still fails sometimes and you need pretty decent volume). As it is, this would not be useful at all for OTA identification of songs -- certainly not close to something like Shazam or SoundHound.
Is this how it works for you as well, or should it work better?
If it should work better, what might I be doing wrong (since it obviously is working, but very poorly so)?
Otherwise, can you recommend another service that offers OTA media recognition / acoustic fingerprinting?
If you'd like me to provide any code, please let me know. Would greatly appreciate advise. Thanks!