bit depth refers to the resolution of amplitude. the higher the bit depth the more dynamic range between loud and soft parts. this is important in orchestral music and movie soundtracks. it allows for quieter sounds to sound less distorted.
one common effect is Bit Bistortion where the bit depth is lowered to 32, 16, 8, or even lower. the result is that it shapes the sound into square waves which is a type of distortion.
you can think of a speaker pushing the air in and out. the bit depth is the number of positions between the maximum in and out position. a square wave sort of makes the speaker jump from all in to all out and back very very fast in a rough way. this is what can blow a speaker. a sine wave is smooth and more gentle. lower bit depth makes your waves more squarish.
imagine your speaker can push out 12 inches and pull in 12 inches. the pattern of in and out vibration is determined by the co inaction of frequencies being played. if that pattern is play at have the speakers capacity of volume, it only moves between 6 inches in and 6 inches out. if you double the volume then it can play the full range of the speaker. the loudest parts will move it the full 12″-12″ spread me song that quieter parts will only travel the 6″-6″ spread or even 1″-1″ of the spread. if your recording is at a low bit forth then these quiet parts in the 1″-1″ spread will be sort of jumpy and squarish and sound like that Bit Distrotion effect.
however, if your music is very loud— compressed to make the softer parts more loud — like most dance music,bthen there is not much dynamic range to where you need to hear quiet parts in great detail, and the difference between 24 and 16 will probably not be noticed. yes, the very quiet parts will be more squarish, but they probably won’t be noticed on the dance floor as even they will be very loud compared to natural sounds.
now in recording, you need more bits because you’re going to turn quieter recordings up on the mix and every time you increase the volume of a recording you are effectively lowering bit depth. 24 bits allows you to record lower to avoid overs or distortion from a signal being too loud (and squaring off the waves as the loudest part because ‘the speaker can’t go any further out or in’ squaring off the tops and bottoms of the waves) and still turn it up or normalize it and have good quality of dynamic range or bit depth.
in other words, 24 bit allows you to record a sound that is very quiet and DOUBLE the volume 8 times and still have the amplitude resolution of a sound recorded at full 16bit.
so if your tracks are recorded correctly they can be mixed down to 16bit and sound perfectly suitable for Electronic Dance Music.
if you’re doing ambient binaural soundscapes on a high end PA like me, it’s probably not suitable. 🙂
—Caleb
ps. Sample Rate is not a reference to amplitude resolution but frequency resolution. it’s how offten a sample of an analog sound is made when converting it to digital. in order to capture a particular frequency you need to capture it at both the high and low part of the wave so you need the sample rate to be twice the frequency of the wave you want to digitize. so 5k or 5,000 Hz (oscillations per second) needs a sample rate of 10k samples per second to capture it. a 44.1k sample rate captures up to 22.05 kHz — way beyond most human hearing.
96k sample rate captures up to 48kHz frequencies, but tha means that 24kHz gets sampled 4 times and so it produces a more accurate sample. 12k gets sampled 8 whole times. 6k gets made up of 16 different samples. this is starting to be the range where human hearing is most sensitive (300-3k Hz).
it also means the samples line up in time more accurately for a more phase aligned sampling or cleaner sound. ‘transparent’. ‘open’.
but again this is most important when recording and if recording and mixing (producing) is done correctly on high sample rate and high bit rate equipment then playback need not be at such a high level.
I do however think sample rate is more important than bit depth as phase distortion from a low sample rate causes more listening fatigue than amplitude distortion from a low bit depth (unless you take very very quiet material and turn it up 4 to 8 times).
eventually when storage costs become irrelevant and processing power costs do as well, we’ll see 192kHz sample rate and 32bit depth become the norm. you have to realize how much power/speed is required to process a sample with 256 times the size (from 16bit to 24 bit) when adding effects or doing real time analysis for software controlled syncing and mixing multiple streams(decks and samples) each added dimension of bit depth (size of sample) and frequency depth (number of samples) exponentially adds demand to your system and we’re just not there yet in affordable processor capability.
DJs really have the best deal as their requirements are so much less in terms of latency, bit depth, sample rate, and even file size than recording engineers or those playing live instruments through software (round trip latency).
