• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Can we hear the bottom bits of 24?

Tangband

Addicted to Fun and Learning
Joined
Sep 3, 2019
Messages
999
Likes
978
Ever record anything? Theres always noise. Room noise, instrument noise (especially elect. guits), mic noise, syths etc have electronics noise, preamp noise. If you get a signal thats 80db above the noise your doing a good job, youll never get 100db S/N.
Agree, but shurely 24 bit on the digital side is technically better than 14 bits? , You also have the recording marginal , about -10 dB to avoid clipping in A/D when recording real instrument in a concert hall.
Its fantastic that we now have digital systems that have much better resolution than all analog electronics, something one could only dream of 25 years ago.:)
 
OP
E

earlevel

Senior Member
Joined
Nov 18, 2020
Messages
314
Likes
391
Perhaps a more persuasive way to make the demo is brief -5db to adjust to say normal listening volume about 80db. Then drop in 6dB decrements to the 16 bit limit. Hold there ad have user turn up volume to make that clear, then start with the -6dB drops again. My guess is by the time one gets to -20 bits won't hear squat. Then drop 21 wont hear squat 22 wont hear squat, 23 wont.... And then ask again if they heard it. I believe that demonstrating that it's not just the last one, but perhaps the last 4 that are inaudible with their system maxed which may help to drive the point home.
OK, I did something similar to what you suggest. My signal, as before, is digital (bit twiddling), so no need to dither. Peak to peak amplitude is two steps (one plus lsb, one minus) or whatever bit level is announced, 5-bit to 24-bit. I won't link to it directly, due to the confusion earlier, but it's at the top of this article, which gives a more detailed explanation:

A listening test
 
Last edited:
  • Like
Reactions: JRS

JRS

Addicted to Fun and Learning
Joined
Sep 22, 2021
Messages
821
Likes
710
Location
Albuquerque, NM USA
OK, I did something similar to what you suggest. My signal, as before, is digital (bit twiddling), so no need to dither. Peak to peak amplitude is two steps (one plus lsb, one minus) or whatever bit level is announced, 5-bit to 24-bit. I won't link to it directly, due to the confusion earlier, but it's at the top of this article, which gives a more detailed explanation:

A listening test
Very well done! I get to -15 bit and it was over--with a decent amp (using laptop + Sundara's) , might have pulled a couple of more bits, but it was clear as day with that signal--I could hear it, then barely, and nothing--no doubt. Be nice to sticky that with whatever listening tests that have been compiled. If not, ASR should start one--sort of an invitation to new browsers, yea lets see what ears you got. Settle that at the get go. Again, nicely done.
 

danadam

Addicted to Fun and Learning
Joined
Jan 20, 2017
Messages
582
Likes
828
Perhaps a more persuasive way to make the demo is brief -5db to adjust to say normal listening volume about 80db. Then drop in 6dB decrements to the 16 bit limit.
Something like that, although it starts at 0 dBFS:

 
OP
E

earlevel

Senior Member
Joined
Nov 18, 2020
Messages
314
Likes
391
Very well done! I get to -15 bit and it was over--with a decent amp (using laptop + Sundara's) , might have pulled a couple of more bits, but it was clear as day with that signal--I could hear it, then barely, and nothing--no doubt. Be nice to sticky that with whatever listening tests that have been compiled. If not, ASR should start one--sort of an invitation to new browsers, yea lets see what ears you got. Settle that at the get go. Again, nicely done.
I listened via two interfaces:

USB -> Topping DX7 Pro -> balanced +4 -> ILoud MTM: If I listened loud 24-26" from speakers, I'd get to 16-bit and then the abyss. But I went louder to the point the dialog was grudging a bit—I set The Bluest Blues to the loudest i could listen, louder than I would normally, still good sounding by at the limits of the MTMs. A louder song would have completely distorted, my aim was to be overly generous. This time I could hear 17-bit, and "detect" 18-bit, then silence. Headphones had the same results, but not for sure I heard 18-bit, I don't want to repeat at that volume.

Firewire -> TASCAM DM-3200 digital mixer -> analog mains out to Equator Q10 studio monitors. These are mid-fields, I was about six feet. I got to 16-bit, then had to work at it. Cranked The Bluest Blues just short of the limiters in the monitors kicking in—loud by sweet. That revealed some low level noise coming in on a channel that I haven't noticed at normal volume, so killed that (it wasn't an analog input, I think that was a digital effects channel). I got to 17-bit, but not 18.

But at normal listening and mixing volumes, 16 would be the limit. Unless it's loud, I'd have to build a room. I'm lucky to live in a quiet neighborhood, with a detached office/studio with no plumbing, but I'd have to do even better. I think if someone could get to 18-bit without getting impractical with volume, they are in an extremely good situation.

edited: The Blues Blues, not Bluest Blue, big difference ;)
 
Last edited:
  • Like
Reactions: JRS

JRS

Addicted to Fun and Learning
Joined
Sep 22, 2021
Messages
821
Likes
710
Location
Albuquerque, NM USA
I listened via two interfaces:

USB -> Topping DX7 Pro -> balanced +4 -> ILoud MTM: If I listened loud 24-26" from speakers, I'd get to 16-bit and then the abyss. But I went louder to the point the dialog was grudging a bit—I set Bluest Blue to the loudest i could listen, louder than I would normally, still good sounding by at the limits of the MTMs. A louder song would have completely distorted, my aim was to be overly generous. This time I could hear 17-bit, and "detect" 18-bit, then silence. Headphones had the same results, but not for sure I heard 18-bit, I don't want to repeat at that volume.

Firewire -> TASCAM DM-3200 digital mixer -> analog mains out to Equator Q10 studio monitors. These are mid-fields, I was about six feet. I got to 16-bit, then had to work at it. Cranked Bluest Blue just short of the limiters in the monitors kicking in—loud by sweet. That revealed some low level noise coming in on a channel that I haven't noticed at normal volume, so killed that (it wasn't an analog input, I think that was a digital effects channel). I got to 17-bit, but not 18.

But at normal listening and mixing volumes, 16 would be the limit. Unless it's loud, I'd have to build a room. I'm lucky to live in a quiet neighborhood, with a detached office/studio with no plumbing, but I'd have to do even better. I think if someone could get to 18-bit without getting impractical with volume, they are in an extremely good situation.
I concur--I prize my hearing--that is what I have left after 60+ years of city living, and a habit of listening loud. So I cheated, I don't have enough gain to begin with, but still started at moderate level til faint and then maxed the output and picked up a few more. My new Topping D3 Pro+ should be here this w/e and I'll have the ability to drive hard (love that 1.5W--even with the big bass EQ should have enough gas to get to 112+). But I don't expect more than a couple of bits, and I still plan to cheat by riding the gain. All my gear is in storage for now so that will have to wait a bit. That your rig can pull 18 is pretty cool. My tweeters might--ribbons with 100dB efficiency and can take 100 steady state. By that time my expensive Scandi 7" midbass drivers are screaming in agony, Woofers ca do 120dB (94 and kW peaks before bumping). I may just have to drop a Faital 6PR140 in there for good measure. Haven't heard it, but that's the beauty of DSP/active. May not get it perfect, but close enough. I really wish they were round--gonna be a bitch getting those flush. Edit: that's a go--just checked Troel's website and he is keen on those mids: just picked up 4dB sensitivity!

Anyhooo, I think that last bit nonsense can be put to rest. Hell if someone can pull 20 I will be most impressed. If you grew up in he country and not on a farm and are 10 years of age, yea then maybe. Sadly, it's a bit like light pollution--a lot of folks have never really seen the Milky Way and haven't the foggiest notion of what they can't see. Likewise with noise pollution, only the damage is permanent.
 
Last edited:

JRS

Addicted to Fun and Learning
Joined
Sep 22, 2021
Messages
821
Likes
710
Location
Albuquerque, NM USA
Something like that, although it starts at 0 dBFS:

Nice, like the tone--again no missing it at near inaudibility. Only got to -60 but will retest this w/e with real amplification.
 
OP
E

earlevel

Senior Member
Joined
Nov 18, 2020
Messages
314
Likes
391
Nice, like the tone--again no missing it at near inaudibility. Only got to -60 but will retest this w/e with real amplification.
Just be aware of the fact this is really a 16-bit test, as it's titled—dithered. The noise floor is at -90 (actually peaks to 84 dB, apparently when the dither rolls an exact 1.0). Everything beyond that is progressively further under the noise floor (which is not say it's not a valid test, its point is to show how good 16-bit can do. Just different from a 24-bit test, especially mine which has a noise floor of -inf dB, which shows how good 24 can do).
 

JRS

Addicted to Fun and Learning
Joined
Sep 22, 2021
Messages
821
Likes
710
Location
Albuquerque, NM USA
I'm glad you pointed that out to me. But as you say, if anything it makes an even more compelling argument re the absurd claim that one can hear down to the last bit. One may as well say they can hear butterflies across the garden.
 

Robin L

Major Contributor
Joined
Sep 2, 2019
Messages
4,337
Likes
6,073
Location
1 mile east of Sleater Kinney Rd
I'm glad you pointed that out to me. But as you say, if anything it makes an even more compelling argument re the absurd claim that one can hear down to the last bit. One may as well say they can hear butterflies across the garden.
Through a double pane window in a rainstorm.
 
OP
E

earlevel

Senior Member
Joined
Nov 18, 2020
Messages
314
Likes
391
@JRS, or anyone who knows ears or has a good reference for the workings in the sense of setting limitations on the quietest we can hear:

In discussion elsewhere, I found that a not-unreasonable guy differs from my conclusions about limits to what we can detect because he believes everything is continuous, therefore there is no hard limit the the least we can hear—even if we test ourselves, there may be some other person who can hear much lower in volume. (It doesn't help to talk about noise floor in our body, room, or electronics, because since it's know we can hear signal below noise, people also tend to think there is no limit there either.) Obviously, the world around us is a lot more discrete than that, even fro things we normally consider analog and therefore continuous—we get down to small numbers of electrons flowing, we get shot noise.

It seems obvious there will be some point where pressure waves at the ear can't deflect the mass and tension of the ear drum, much less passing through to affect the hair cells. (And even pressure waves spreading and weakening to the point they can't deflect air molecules.) Any material on physical limits that impose a minimum SPL (obviously frequency dependent) any human can hear?
 

Robin L

Major Contributor
Joined
Sep 2, 2019
Messages
4,337
Likes
6,073
Location
1 mile east of Sleater Kinney Rd
@JRSIt seems obvious there will be some point where pressure waves at the ear can't deflect the mass and tension of the ear drum, much less passing through to affect the hair cells. (And even pressure waves spreading and weakening to the point they can't deflect air molecules.) Any material on physical limits that impose a minimum SPL (obviously frequency dependent) any human can hear?
. . . or similarly, the plastic membranes of microphones. Bigger, thicker diaphragms [say, a Shure 58] can take higher levels of volume at the expense of such things as treble. Smaller, thinner microphones [small diaphragm condensers] have more extension on top but along with that more noise.
 

valerianf

Senior Member
Joined
Dec 15, 2019
Messages
402
Likes
234
Location
Los Angeles
Audio recording is not technically different of professional video recording.
The camera record the shot at a resolution between 6k to 8k.
Then the post production is making its magic tweaks.
At the end they deliver what ever resolution is needed for the customer: 720p for broadcasting, 1080p for BR, 4K for UHD BR or theater.
Everybody is happy because the image quality is up.

Same think for audio: if multichannel 24bits 96kHz becomes the audio recording standard, then every lower standard can be delivered to the customer including Dolby Atmos.
 

JRS

Addicted to Fun and Learning
Joined
Sep 22, 2021
Messages
821
Likes
710
Location
Albuquerque, NM USA
@JRS, or anyone who knows ears or has a good reference for the workings in the sense of setting limitations on the quietest we can hear:

In discussion elsewhere, I found that a not-unreasonable guy differs from my conclusions about limits to what we can detect because he believes everything is continuous, therefore there is no hard limit the the least we can hear—even if we test ourselves, there may be some other person who can hear much lower in volume. (It doesn't help to talk about noise floor in our body, room, or electronics, because since it's know we can hear signal below noise, people also tend to think there is no limit there either.) Obviously, the world around us is a lot more discrete than that, even fro things we normally consider analog and therefore continuous—we get down to small numbers of electrons flowing, we get shot noise.

It seems obvious there will be some point where pressure waves at the ear can't deflect the mass and tension of the ear drum, much less passing through to affect the hair cells. (And even pressure waves spreading and weakening to the point they can't deflect air molecules.) Any material on physical limits that impose a minimum SPL (obviously frequency dependent) any human can hear?
I tried to some posts ago, to argue that yes there are hard limits to hearing. People seem to get that one can only detect a single photon and it can't get any better. While we can conceptualize phonons, and model behavior of a collection of bodies in response to being struck by one, biology is anything but an isotropic array with well known electrical and kinetic coupling between neighbors. It's mushy, bony, gooey, hairy, and fuzzy, very fuzzy in the way calculations are made and information is encoded. By the time we get to an actual nerve fiber transmitting information from the cochlea to the brain, we have volleys of nice all-or-nothing events more or less corresponding to a 0 or +5V state. But even that gets mushy when you start to look at how these volleys of nerve firings are processed and reprocessed as one ascends higher orders if processing in the brain.

But winding back to where transduction occurs within the inner hair cells, researchers have made some remarkable inroads. Where the rubber meets the road happens below:

hair cell.PNG
1636247422387.png

What is meant by a channel such as the MET is a hollow protein pipe that allows for the movement of small monovalent and divalent cations such as potassium and calcium across the cellular membrane, which is accompanied by an electrical change or is coupled to a metabolic process. In this case, it's both. Inwardly directed calcium and potassium currents both depolarize the cell membrane, and activate the contractile proteins involved in dragging vesicles of the chemical glutamate to the edge of the cell where it is released (see quantal release below).

So under shear caused by the upward displacement of the basilar membrane in response to fluid waves within the cochlea, we get an event! Somehow the channel opens and from some unbelievable technically adept measurements made previously, we know the amplitude transfer function, and the concuctance change in the membrane to be about 340 pSiemens (or pMhos) So that's about as granular as it gets. We know that the opening of the MET channels in the inner hair cells excites the inner hair cells to release quantal packages of the chemical glutamate that tends to activate the cells that fire nd send electrical responses to the brain. There's that fuzziness again. To complicate matters the outer hair cells are likewise being activated by not only auditory information that reduces sensitivity (compression) or alternatively may facilitate coupling between the inner hair cells and the brain as a result of conscious effort to hear a particular frequency.

All I guess is a way of saying we have mushy granules corresponding to events of maximal sensitivity. We know that maximal stimulation where the hair cell saturates corresponds to a deflection of
transduction transform.PNG

of 200 nm, roughly 1/240 of a human hair. One can see a response to much smaller displacements.
 
Last edited:

danadam

Addicted to Fun and Learning
Joined
Jan 20, 2017
Messages
582
Likes
828
@JRS, or anyone who knows ears or has a good reference for the workings in the sense of setting limitations on the quietest we can hear:

In discussion elsewhere, I found that a not-unreasonable guy differs from my conclusions about limits to what we can detect because he believes everything is continuous, therefore there is no hard limit the the least we can hear—even if we test ourselves, there may be some other person who can hear much lower in volume.
Sounds like something right up @j_j alley. I have a feeling that he already answered that many times, but maybe he'll be willing to repeat it again here :)
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
1,336
Likes
2,688
Location
My office, custom listening room, or anywhere.
Sounds like something right up @j_j alley. I have a feeling that he already answered that many times, but maybe he'll be willing to repeat it again here :)

First, the actual noise at the eardrum due to the molecular nature of air (yes, and atomic for argon) is about 6-8dB SPL 20-20K. It's approximately white noise. (No, the velocity distribution of molecules is not white. Different question.) Now you can't hear that, QUITE, because the energy in any given ERB is just barely too small.


Now, the measured threshold of hearing is about -15dB SPL in a narrow band around your ear canal resonance (which is about 3dB or so higher than the white noise in the same bandwidth).
Theory is not necessary for the statement of threshold of hearing, because it's measured. So, at the most sensitive part of the ear's frequency range, it's a touch above the noise due to molecular movement in the air.

If you look at https://en.wikipedia.org/wiki/Equal-loudness_contour at the zero loudness curve, that's pretty accurate for a person with absolutely no hearing impairment.
 

JRS

Addicted to Fun and Learning
Joined
Sep 22, 2021
Messages
821
Likes
710
Location
Albuquerque, NM USA
It is interesting to note the background noise that the childhood seashell experience makes so abundantly clear is present and apparently ignored as background noise-in fact, its absence is so hmmm disquieting that when exposed to an environment where it is almost within reach of hearing, people run for cover--that's at the Microsoft anechoic chamber where quiet is equal to -20.6, just 4 dB above the current estimate of molecular collision noises. So we apparently can't hear that under "ordinary" conditions, but can't quite prove, and in any event such is masked by sounds from ones own circulation, eye movement, etc.

But that doesn't entire squash the argument that it prevents a signal being heard, as ample experimentation has shown: signal detection actually improves when low level background noise is present. It also seems to me that just as a long period of dark adaptation is needed before one can see a photon, it may be that a similar period of stimulus deprivation may be needed to achieve absolute measure of sensitivity. Now if only one could stop the heart long enough, and make the experience of ultimate quiet less psychologically distressing.... Perhaps an IV cocktail of curare and valium? (Sort of like being intubated because of Covid, and thinking wow, in retrospect those damn masks weren't so bad after all).

In any event, if we take the -8 number you mention and add it to say 125dB as the loudest we can hear without saturation (arguably this is also "felt)--for me, personally, it was a Quicksilver Messenger Concert at a cow palace where I, in all my adolescent bravado, was standing three rows back and they were actually bragging about 125 dB peaks between songs (yep 3 days of tinnitus afterwards--the sound of hair cells dying :eek:), we get what within a couple of bits. All of which is to say there is still wiggle room for the absolute limit of hearing. Granted we may also need to strip Argon from the listening environnment ;), but then oxygen becomes a problem so perhaps dropping the temperature becomes necessary.

In any event, it's not me that needs convincing--though I would still prefer measuring in hard biological terms where that threshold is. It's just a lot more difficult number to measure owing to the complexities of self-noise and necessarily stripping away higher order processing from the picture when the same measurements are made in vitro.
 

j_j

Major Contributor
Audio Luminary
Technical Expert
Joined
Oct 10, 2017
Messages
1,336
Likes
2,688
Location
My office, custom listening room, or anywhere.
It is interesting to note the background noise that the childhood seashell experience makes so abundantly clear is present and apparently ignored as background noise-in fact, its absence is so hmmm disquieting that when exposed to an environment where it is almost within reach of hearing, people run for cover--that's at the Microsoft anechoic chamber where quiet is equal to -20.6, just 4 dB above the current estimate of molecular collision noises. So we apparently can't hear that under "ordinary" conditions, but can't quite prove, and in any event such is masked by sounds from ones own circulation, eye movement, etc.

But that doesn't entire squash the argument that it prevents a signal being heard, as ample experimentation has shown: signal detection actually improves when low level background noise is present. It also seems to me that just as a long period of dark adaptation is needed before one can see a photon, it may be that a similar period of stimulus deprivation may be needed to achieve absolute measure of sensitivity. Now if only one could stop the heart long enough, and make the experience of ultimate quiet less psychologically distressing.... Perhaps an IV cocktail of curare and valium? (Sort of like being intubated because of Covid, and thinking wow, in retrospect those damn masks weren't so bad after all).

In any event, if we take the -8 number you mention and add it to say 125dB as the loudest we can hear without saturation (arguably this is also "felt)--for me, personally, it was a Quicksilver Messenger Concert at a cow palace where I, in all my adolescent bravado, was standing three rows back and they were actually bragging about 125 dB peaks between songs (yep 3 days of tinnitus afterwards--the sound of hair cells dying :eek:), we get what within a couple of bits. All of which is to say there is still wiggle room for the absolute limit of hearing. Granted we may also need to strip Argon from the listening environnment ;), but then oxygen becomes a problem so perhaps dropping the temperature becomes necessary.

In any event, it's not me that needs convincing--though I would still prefer measuring in hard biological terms where that threshold is. It's just a lot more difficult number to measure owing to the complexities of self-noise and necessarily stripping away higher order processing from the picture when the same measurements are made in vitro.

I have to differ a bit with the MS numbers. For the eardrum's area, the number is about +6 to +8 dB SPL white noise.

There were a few people who might have been able to hear it in the old Bell Labs chamber, before it was allowed to slide sideways off its springs into one of the side walls, which greatly upped the noise level. :(
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,844
Likes
13,179
Location
Riverview FL
So, I'm asking whether people think the lowest couple of bits of 24-bit audio can be heard, and why or why not.

First, set a listening level for playback of the loudest parts.

For my convenience, let's pretend it is as loud as someone shouting at you from one meter. Too loud, but, it circles back to something I wrote long ago.

Assuming a set of theoretically perfect listening conditions (that don't exist), the distance of a second shouter's shout reaching you at the level of the last little bit of 24 would be 10,425 miles away...

If "the last couple of bits" is of interest, move that shouter to 5,212 miles distant.

 
Top Bottom