is not bit depth
. Bitrate means how much data in a specified time therefor it's stated with bits per second or kilobits per second. Its a value not really used in this context.
A hint to understand bitdeph is to think about a ladder that represents volume. On top of the ladder all 16 bits are set to 1 and this represent 0dBFS (FS means full scale and show that we're talking about digital dBs). Goin down one step on the ladder you're standing on 15bit or in dB -6dB cause 1bit =6dB.
With each step down you pass -6dB so when you reach the bottom of the 16bit/step ladder you're at -96dB. You can't go further it's just mud there which is called noisefloor
What you can do is use a 24bit/step ladder. Then you can reach the -144dB mark.
What does it mean? The more bits an audio interface provide the quiter the audio can be. Recording a very quiet audio signal with a very low bitdepth and then raising the volume (normalizing) after recording/sampling will result in audible noise cause you don't just raise the recorded signal but also the 'mud'.
Also.....(I hope I'm not telling BS with this but thats how I understand it)....more bits mean finer 'inbetween' dB values per step.
On a 16bit ladder the values from -96dB to -90dB are represented by just 1bit which can have 2 states (0 and 1). The values -89.9998dB to -84dB are represented by 2bits (which are 4 states 00, 01, 10, 11) therefor 6dB are divided in 4 parts. Goin up the ladder the states which represent 6dB increase and at the top at 16bit the 6dB are divided in 65535 parts.
I hope this is understandable and my thoughts are correct. Maybe this chart help
Sound examples (this is created in Audition with original 16bit samples so its not true n-bit recording but calculated)
This is the list
12bit with dither
10bit with dither
8bit with dither
8bit with dither (noiseshaped)
4bit with dither
4bit with dither (noiseshaped)
Old game music (chiptunes) is also low bitdepth.
The other important value is the samplerate but I'm tired now.....