Is there any music that actually requires 24 bits for replay?

[Tell]

It's a good moment to meditate on this chart from Amir's deck for AES.

View attachment 416742

All the fussing over and debating the just-measurable microscopic differences in the blue and green is kinda like our version of listening to the differences between power cables, except with an AP instead of ears.

I really think that Amirs rating of DACs is off, everything should be to the right, so everything in the "Poor" category should be "Fair" instead (well except that very bottom I guess), and what's in "Excellent" should be called "God like" or something in the similar vein. I mean it feels like people are afraid of getting DAC in the current yellow category which they really shouldn't be since most won't hear any difference between those and the one in the absolute top.

A lot of this is for marketing purposes in the same way that 4K video resolution is pointless on a 50” TV and 8K is silly, but makes people go, ‘oooooh’.

The audiophile community who eschew science still believe that the “staircase” is audible, so to them, the more bits and samples the closer we get to analogue. This ignorance has been exploited by hifi manufacturers for decades now and continues to this day.

I do get what you mean, but it's not exactly the same way since 4K even at 55" do make a quite clear visible difference from 1080p if you have good eyes, and when I'm sitting around 1.5m from my 55" even 8K would make a noticeable difference.

 

[pablolie]

My brother (an EE prof) is also a gifted musician, and, after nearly 7 decades of playing semi-pro with local orchestras (french horn and cello), he's almost completely deaf now.

Unfortunate but the culprit is excessive SPL... not DR.

 

[Killingbeans]

I went to my local used record store yesterday and most of the shoppers were teens, and twenty something, buying records not CD's. Whether there after the cool album covers or, "chasing sound" I don't know.

Being a teenager is all about searching for identity and a place in society. People, who listen to vinyl and idolize it, have recently shifted from being a minority group to a somewhat respected tribe, and so, teens naturally gravitate towards vinyl as a safe compromise between individualism and social acceptance.

Or maybe I'm overthinking it. Could be that vinyl has simply become a fad and nothing more

 

[Saponetto]

Being a teenager is all about searching for identity and a place in society. People, who listen to vinyl and idolize it, have recently shifted from being a minority group to a somewhat respected tribe, and so, teens naturally gravitate towards vinyl as a safe compromise between individualism and social acceptance.

Or maybe I'm overthinking it. Could be that vinyl has simply become a fad and nothing more

First theorem is almost actual.

 

[Cbdb2]

Is it not the case that 24 bit has advantages when using DSP?

Yes to a point. But taking original 16 bit and converting it to 24 bit for processing than back to 16 bit for playback is the same as starting with 24bits.
I believe most DSP units now do this.
Edit: Should have read more posts.

 

[Blumlein 88]

In answer to the thread question, no music requires 24 bits for replay.

How much is needed is a bit trickier. First off a single noise level in your listening room is deceptive. Noise is pinkish. Our ears split up the sound into 30 or so bands (somewhat simplified explanation). ERB, effective rectangular bandwidth for each band. It is why we can hear a little bit below noise floors. Even more notable in the range of our most sensitive hearing between 3-5 khz. Rooms with a noise level well up into the 30 dbSPL range will be at or below 10 dbSPL in that 3-5 khz range. With healthy hearing we'll hear that range down near 0 db SPL in any slightly quieter than normal room.

So you might need 20 bits if your gear could play to 120 db SPL cleanly. Next however is dither. It allows you to capture below the 96 db range of 16 bit. Shaped dither can let you record and replay things down to -110 db or more. Of course most systems are unlikely to do more than 105 dbSPL cleanly. So 16 bit would be enough.

Now going the other way, are there recordings with 24 bits in dynamic range? No, other than maybe total electronic creations. When I tried to find the widest range recordings there were a few that reach -70 db. Though like noise in rooms in our most sensitive range some could get down around 90 db. Some are below that at high frequencies, but then our hearing is not sensitive at high or low frequencies. Usually (probably always) the recording itself is the bottleneck for dynamic range. You have self noise in microphones, amplified noise in microphone preamps and usually higher than either is environmental noise where the music was performed.

You also have masking and the fact we cannot hear at the very lowest levels in the presence of louder things at the same time. The fellow who wrote Audio Diffmaker, at Liberty Instruments, has offered some Brahm's recordings where he also mixed in a marching band recording. I forget if the band was down 40 db or 60 db from the main recording. I think it was 60 db down. Somewhere on ASR is a link to this. I can save you the trouble unless you want to convince yourself, but you cannot hear whether the marching band is there or not.

So how much is enough? 60 db (10 bits), 72 db (12 bits) to be sure? Well dithered 8 bit is better than most people imagine. I'm not sure what the very finest LP noise level is. On some very excellent TTs and good clean vinyl you can barely hear the surface noise in silent portions even at loud listening levels and once any music starts you cannot hear it at all. So for playback 16 bit is plenty. One more thing we really don't need to worry about.

 

[oPossum]

Real cannons.
16 bits.
It's fine.

 

[pablolie]

Real cannons.
16 bits.
It's fine.

Nobody needs to hear a cannon in full resolution at full SPL. 1812's cannon shots are one of those audio myths to me. "Oh I heard the unmistakable bark of a true Canon de 12 Gribeauval blabla"... who has a reference to claim that? What would the neighbors say...

 

[oPossum]

Nobody needs to hear a cannon in full resolution at full SPL. 1812's cannon shots are one of those audio myths to me. "Oh I heard the unmistakable bark of a true Canon de 12 Gribeauval blabla"... who has a reference to claim that? What would the neighbors say...

OK. So 16 bits are too much.
How about 12???
I think Hip-Hop would be fine with 8.

 

[Blumlein 88]

OK. So 16 bits are too much.
How about 12???
I think Hip-Hop would be fine with 8.

Hip hop needs about 3 bits.

 

[Chr1]

Shoegaze: 5

 

[Ropeburn]

Nobody needs to hear a cannon in full resolution at full SPL. 1812's cannon shots are one of those audio myths to me. "Oh I heard the unmistakable bark of a true Canon de 12 Gribeauval blabla"... who has a reference to claim that? What would the neighbors say...

Take it from a retired soldier who has attended to more than one night shooting in his heyday (meaning pitch black silent night, not even crickets daring to mouth up), with loads of rifles and machineguns blaring into the night with tracer rounds, plus various grenades thrown and Panzerfausts firing battlefield illumination. Talk about "dynamic range"!

No, we (our ears) do not appreciate cannon fire. We prefer much lower peak levels for a satisfying and very impressive sound experience. Infact you don't even want to experience the average battle rifle firing without hearing protection. It's 160dB or more. It's painful and immediately damaging to the unprotected ear.

 

[Blumlein 88]

I've worked where there were buildings that ran around 105 to 108 dbSPL A weighted. You could stand it for a while. Some worked in it. I used hearing protection myself. The sound of this mechanical gear was the very example of a cacophony. The thing was if you used some foam ear plugs dropping the level 25-30 db, there was a wealth of fine detail in the various mechanical sounds. While we can hear at higher levels that isn't normal, and not what our hearing is optimized to do. It was a dramatic difference in what you could discern at the lower level. When loud you could make out little.

I've read accounts from an amp designer at a pro audio company. They did DBT tests. He said listeners were most discerning when music was an average level of 70 db. He said whenever test listeners turned it up, especially if they had it at 80 db or more the results became mostly garbage. Our ear has a muscle to limit sensitivity starting around 75 db. So this whole thing is a bit complicated, but the good news is our gear other than transducers exceeds our hearing. Many common audiophile concerns are wasted effort, and worry.

 

[Ropeburn]

That is the thing indeed. While the average human ear can certainly tolerate and to

I've worked where there were buildings that ran around 105 to 108 dbSPL A weighted. You could stand it for a while. Some worked in it. I used hearing protection myself. The sound of this mechanical gear was the very example of a cacophony. The thing was if you used some foam ear plugs dropping the level 25-30 db, there was a wealth of fine detail in the various mechanical sounds. While we can hear at higher levels that isn't normal, and not what our hearing is optimized to do. It was a dramatic difference in what you could discern at the lower level. When loud you could make out little.

I've read accounts from an amp designer at a pro audio company. They did DBT tests. He said listeners were most discerning when music was an average level of 70 db. He said whenever test listeners turned it up, especially if they had it at 80 db or more the results became mostly garbage. Our ear has a muscle to limit sensitivity starting around 75 db. So this whole thing is a bit complicated, but the good news is our gear other than transducers exceeds our hearing. Many common audiophile concerns are wasted effort, and worry.

So
high sound levels for a very short time

I've worked where there were buildings that ran around 105 to 108 dbSPL A weighted. You could stand it for a while. Some worked in it. I used hearing protection myself. The sound of this mechanical gear was the very example of a cacophony. The thing was if you used some foam ear plugs dropping the level 25-30 db, there was a wealth of fine detail in the various mechanical sounds. While we can hear at higher levels that isn't normal, and not what our hearing is optimized to do. It was a dramatic difference in what you could discern at the lower level. When loud you could make out little.

I've read accounts from an amp designer at a pro audio company. They did DBT tests. He said listeners were most discerning when music was an average level of 70 db. He said whenever test listeners turned it up, especially if they had it at 80 db or more the results became mostly garbage. Our ear has a muscle to limit sensitivity starting around 75 db. So this whole thing is a bit complicated, but the good news is our gear other than transducers exceeds our hearing. Many common audiophile concerns are wasted effort, and worry.

That's perhaps the most overlooked thing in audio. We always talk about source, amplifier, even speaker distortion. But what about ear distortion? Granted, it's practically impossible to measure, and individually varying too. But it's still a real thing. Personally I have rather sensitive ears that audibly distort earlier than others' - but I have no way to measure and quantify it.

Edit: reply formatting fucked up, I hope you get the general idea

 

[Matt_Holland]

I do get what you mean, but it's not exactly the same way since 4K even at 55" do make a quite clear visible difference from 1080p if you have good eyes, and when I'm sitting around 1.5m from my 55" even 8K would make a noticeable difference.

Copied from HERE:

“On computer displays there is definitely something to say for 4K. You can display a lot of information simultaneously and you usually only have to focus on a small area at the time, which means the higher detail really has added value. Furthermore the short viewing distance allows a wide field of view without the need for extremely large displays. 8K might have its uses with very specific applications, but in general it would be excessive.

With televisions it’s a different story. Many people probably aren’t even making full use of their FHD TV yet. To really profit from 4K you’d need an extremely large screen, or sit extremely close. And 8K is just plain ridiculous. For a 250 cm viewing distance you’d need a 595 x 335 cm screen. There aren’t that many people with a wall that big in their house and even if you had, you’d need a pretty impressive beamer and a very large projection screen (they obviously don’t make TV’s that big).

One of the reasons that 4K televisions sell relatively well might be that in the store people tend to look at it from a very short distance, at which they could easily see that 4K is sharper than FHD. If they would look at it from the same distance as the actual distance they would view it from at home, many would not be able to tell the difference (if all other aspects of the image reproduction were identical for both displays). Manufacturers know this, so from a marketing perspective 4K is very clever.”

 

[Robin L]

I've got 16 recordings with too wide a dynamic range - the lowest levels are buried in environmental noise if the peak levels are tolerable.

 

[Multicore]

I really think that Amirs rating of DACs is off, everything should be to the right, so everything in the "Poor" category should be "Fair" instead (well except that very bottom I guess), and what's in "Excellent" should be called "God like" or something in the similar vein.

Maybe so. But the most important thing for Amir is that he be consistent. Then, no matter what words you might prefer, you can easily translate his measurements into your preferred vocabulary.

I think Robert Parker has bad taste in wine but he is very consistent so his wine reviews are still useful because I can translate what they mean.

 

[Ropeburn]

Maybe so. But the most important thing for Amir is that he be consistent. Then, no matter what words you might prefer, you can easily translate his measurements into your preferred vocabulary.

I think Robert Parker has bad taste in wine but he is very consistent so his wine reviews are still useful because I can translate what they mean.

Really, Amir's and ASR's standards are high enough in that regard. From over 30 years of listening experience, my (very average) ears say: the "good" category translates to "excellent" - and above that, only practical things matter.

 

[krabapple]

There are those who have the opinion that the audio system should be able to record the entirety of the dynamic range of heathy young human ears.

You mean a recording audio system. Most home audio systems aren't used for for that.

Anyway as far as what's 'required' to record the corner cases without distrotion, Fielder wrote the book (article ) on that.

Dynamic-Range Issues in the Modern Digital Audio Environment​

• L. Fielder
• Published 1 May 1995
• Engineering, Physics
• Journal of The Audio Engineering Society

The peak sound levels of music performances are combined with the audibility of noise in sound reproduction circumstances to yield a dynamic-range criterion for noise-free reproduction of music. This criterion is then examined in light of limitations due to microphones, analog-to-digital conversion, digital audio storage, low-bit-rate coders, digital-to-analog conversion, and loudspeakers. A dynamic range of over 120 dB is found to be necessary in the most demanding of circumstances, requiring the reproduction of sound levels of up to 129 dB SPL. Present audio systems are shown to be challenged to yield these values

Future MQA huckster Bob Stuart later leveraged this to famously argue in JAES that we 'need' music to be supplied as 24 bit audio for home playback.

The 'need' is there during recording and production, for headroom convenience , to avoid quantization errors becoming audible, and for the rare case that there really is >96dB of dynamic range to be captured.

But for music you buy or download, not so much.
(That said, your AVR is probably doing its DSP on your audio converted to 24bit, which makes sense for the same reason 24 or 32 bits are useful during production)

 

[Multicore]

Really, Amir's and ASR's standards are high enough in that regard. From over 30 years of listening experience, my (very average) ears say: the "good" category translates to "excellent" - and above that, only practical things matter.

That's up to you. I can't tell the difference between 128 kbps compressed and CD so I'm not the one to ask. And yet I still think it's more important that Amir remain consistent on his reviews because that simplifies our individual tasks translating those numbers into our own language. We are all subjectivists in the end and should not shirk our duty in interpreting measurements to decide how they relate to our own specific purposes.