The intention of this page is to give a very basic overview that explains what can be done with mp3 and why. This page can by no means explain the finer concepts of mp3 or sound engineering.

There are two major variables that determine mp3 sound quality.

• number of bits per sample
• sample rate

The more bits you have to describe the sound amplitude, the better is your ability to find a value that is very close to the actual sound amplitude. The closer the bits recreate the real amplitude value, the closer is the generated sound to the original. The amazing thing is that this has a large influence on high frequencies and a moderate impact on lower frequencies (the algorithms and the following electronics will create a sine wave almost automatically).

The sample rate basically determines the highest frequency that can be recorded. The highest frequency is determined by the "sample theorem". Slower sample rates reduce high frequencies.

O.K. so here is the message: use high sample rates and high bit resolutions to create fantastic sounds … (just kidding) actually don't. There is one more thing that you should always consider when you think sound and sound quality. How and where do I reproduce the sound. In our case the answer is:

• on a crappy PDA/mobile phone speaker
• in the woods or on the street

Now here comes the secret. No matter how many bits you invest and no matter what sampling rate you apply. Your music can not be better than the replay device and situation (if you think about this, this knowledge might actually save you a lot of money when you consider your next HiFi car audio system).

Back to the issue at hand. Play around with the sampling rate and the bit depth to find a mix that is still good enough for the cartridge. I am typically using 16 bits and 16 kHz mono (because I can easily remember the numbers ). This setting reduces a standard 30 minute mp3 file (the kind that you would use on your mp3 player) from 54 MByte to 4 MByte. In other words it gives you roughly 4 minutes worth of mp3 play time out of the 1/2 MByte that we can currently use in Earwigo. This should be good enough for any cartridge that I can think of (exception: constant background music, which might be an interesting concept ).

One last word with regards to sound quality. The above settings can absolutely not create a HiFi sound signal (neither can your PDA speaker). The above setting is good enough to play voice or voice type signals. If you need music, increase the sample rate to 32 kHz. Remember we are not discussing HiFi, this is about a cartridge with sound injection.

• Voice: 16 Bit, 16 kHz, mono
• Music: 16 Bit, 32 kHz, mono

Free tools to convert mp3 settings: Audacity or GoldWave