View Single Post
Old 2012-07-04, 09:09 PM
paddington's Avatar
paddington paddington is offline
TTD Staff
Join Date: Jan 2005
Location: UK
Re: 16 Bit vs 24 Bit

The answer is that 24 bit depth gives the digital sample more steps for *dynamic reproduction* - bit depth has nothing to do with frequency response.

24 bit audio sounds much more natural, because there are more steps of volume than with 16 bit.

All digital stuff is either on or off. There is no analog glide for volume swells, which are intrinsic to music. Dynamics play a huge part of the music experience (different levels of playing - soft to loud) and 24 bits in a sample increases the number of volume levels which can be sampled - so the volume swell (or decay) you hear with your ears can be much better replicated.

To give you an idea of what sample *rate* is, verses sample *depth*:

Sample rate:
A 192k rate has more to do with the quality of the top end, not so much getting the top end.

Example: 44.1 kHz stereo sampling yields 22050 per channel, according to mr Nyquist, which is enough for human hearing range + some headroom to handle the aliasing of the roll-off above the range where it can be heard.

Since it is yielding a full (human) range spectrum, 44.1 kHz is adequate for reproducing the range of human hearing - 20 Hz - 20 kHz. Higher sample rates simply increase the quality of the higher frequncies.

Bit depth: governs the perceived smoothness of the transitions between the volume levels by increasing the number of steps in a volume swell or decay. The smoothness afforded by 24 bit depth is just as important as the increase in available levels. Not only are there MANY more volume levels to better approximate the analog experience, the steps between them are MUCH less noticeable to your subconscious hearing and it just sounds more natural.

The Dynamic Range is not increased from 16 bit to 24 bit... the # steps inside that range are increased and the resultant sound is more natural - "less digital sounding" - as the sound swells or falls.

- "Dynamic Range" = technical term, measured in decibels - the range from the lowest to the loudest sound in the recording.

- "Dynamics", related to musicality = musical term meaning volume levels at which the music is played during a performance. Soft parts, loud parts, crescendos, decrescendos, etc.

The two aren't really related.

-> 16 bit audio yields 65k levels for amplitude (volume levels).

-> 24 bit yields nearly 17 million levels for amplitude (volume levels).

Which would yield a more perfect analog reproduction of a trumpet crescendo? 65,585 steps per second or 16,777,777 (whatever the number is)?

For the 33% more storage space, the ~255x improvement in amplitude resolution is quite a bargain, to me.

I remember reading a study done that showed the average human ear to be able to detect around the number of levels yielded by 20-22 bit digital audio.

So, no, maybe we can't all hear the full 24 bit resolution - but isn't that the point? If we can hear WAY better than 16 bit can deliver, 24 bit, in theory, should be yielding enough levels for the steps to be completely indiscernible to the human ear, if you go by that study's results.

Which is the whole point of digital audio... to perfectly reproduce the original analog sound - to our ears. 16 bits isn't enough to do that.

24 bits get the job done.
"There are some of these recordings where it is just a whirring, and you cannot hear the music. " - Jimmy Page, 2007 / JUL / 26
Reply With Quote Reply with Nested Quotes