32 bit, 64 bit, etc. Explain, please.

Question

would anyone be so kind to explain what the following from the Logic Express 8 official description means:
"Internal audio resolution: 32-bit floating point; 64-bit precision where required"
i mean i know of 16 and 24. but what are those above?

Answer 1

When audio is recorded into computer format, it is normally recorded as 16 or 24-bit resolution. Think of these figures as representing the level of detail that can be recorded. 24-bit recording has 256 times more detail than 16-bit (though many people can't hear the difference).
Once the material is in digital form, it's in the format that you've recorded it in. However, if the computer is going to do substantial amounts of calculation (adding effects, summing signals, applying gain etc), some of the detail is lost, even more so if your signals are too hot and you get clipping.
To avoid this, Logic converts all your 16 or 24-bit files into 32 bits and then any lost detail when doing processing isn't significant. You can tell that Logic works in 32-bit because, if you freeze a track, the Freeze Files folder in your Logic Project folder contains 32-bit audio files. Open one in Quicktime and get the info for it.
You might think that working with greater levels of accuracy might slow the computer down but that's not really the case. Modern chips work with either 32 or 64-bit packets of data and it's no harder or slower to do a sum on a 32-bit value than it is to do the same on a 24-bit value.
Finally, when you bounce down your music, you choose the target bit depth (16 or 24) and the computer has to convert its 32 bit numbers into your chosen bit depth. Incidentally, this is where dithering comes in. You are automatically losing some level of detail going from 32-bit to 16-bit and dithering is a way of making sure that no artificial waveforms (called aliasing) are created when you do so.
HTH,
Pete