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

[kemmler3D]

I think compression on music recordings ia about making it listenable on modest equipment like little smart speakers and in noisy environments, earbuds on a 'plane or a busy family home and not wanting to annoy other family members.

Mostly agree, but even normal "really good" systems aren't really usable for more than 80dB of dynamic range above noise in a real room. Even 50dB is a lot to ask from a typical listening situation...

 

[Frank Dernie]

I will confess to being slightly OCD about getting 24bit files a few years back. Took me a while to realise that the quality of the actual recording/mastering is way more important than the bit depth of the playback file.

Yes well IME the quality of the recoding makes more difference than any of the electronics in any music system. Some speakers maybe not, but also IME I would rather listen to a fabulous performance of a piece on modest equipment than an uninspiring one on the best hifi available.

 

[Chr1]

Indeed.

The downside of this hifi hobby is that half of the new wave from my youth has become virtually unlistenable now!
Ho hum.
(First world problems)

 

[Frank Dernie]

Mostly agree, but even normal "really good" systems aren't really usable for more than 80dB of dynamic range above noise in a real room. Even 50dB is a lot to ask from a typical listening situation...

Oh absolutely.
The comment of mine you quote was a reply to somebody posting about dynamic range compression, not HiFi music requirements.
I usually check the dynamic range at concerts with a sound level meter...

 

[Mart68]

Also lets accept that plenty of people on other sites believe LPs are superior SQ to digital and even the most cloth eared would not be able to believe that if more than 11-bits or so was actually needed for music.

Ironically they're often the same people who say that digital is unlistenable unless it's at least 24/96

 

[ZolaIII]

I agre to disagree. To a 16 bit with dithering is certainly enough for music or little bit more than 18 bit simple as it is is enough even for THX regarding cinema 24 dB DR to 88 dB calibration point. More than that is insane and good luck finding speakers/headphones. Bit width and packing is architecture influenced and for signal manipulation or better volume control's it's normal to use even double precision FP and 32 bit integer (for and to simeles) to DAC but for storage and transport 19 bit is enough (shaving and reconstruct add nulls on decode). On the other hand when having wide bus and keeping it synced is too expensive (from engineering point) having tap based for transport is OK filtered of course and very high at DSD128 level cut in 16 bit streams. Remember that on any interconnect you will lose couple dB of DAC SINAD and how not only speakers and amps will have higher THD but even very low distortion headphones but we are less sensitive to it in lows and highs anyway.
On the another hand it works for music on another scale the note one and to main tones and their harmonics and neadles to say little is left behind 12 KHz all together. And so on.

 

[Multicore]

I will confess to being slightly OCD about getting 24bit files a few years back. Took me a while to realise that the quality of the actual recording/mastering is way more important than the bit depth of the playback file.

That's correct. And it's also worth remembering that the quality of the recording/mastering isn't worth a damn if the musical performance and material is bad. In other words, what are we really trying to do here?

 

[Chr1]

Yes, but here we're into the murky realms of subjectivity...

 

[Neddy]

Mahler 6. Let us give thanks we don't have to endure this every day.

View attachment 416744

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.

 

[Frank Dernie]

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.

My wife is a professional musician and is now deaf.
The sound level in the middle of an orchestra is unbearable IMO, luckily not so for the audience so not needed at home either.
Most musicians I know wear ear plugs nowadays, too late for my wife.
We have a Steinway Model B in our music room and even with the main lid shut I find that too loud at the keyboard.

 

[kemmler3D]

My wife is a professional musician and is now deaf.
The sound level in the middle of an orchestra is unbearable IMO, luckily not so for the audience so not needed at home either.
Most musicians I know wear ear plugs nowadays, too late for my wife.
We have a Steinway Model B in our music room and even with the main lid shut I find that too loud.

Pretty much case in point as to why "realistic" dynamic range isn't necessarily practical. Live music in person (especially if there are drums) is often well beyond safe long term levels.

 

[Frank Dernie]

Pretty much case in point as to why "realistic" dynamic range isn't necessarily practical. Live music in person (especially if there are drums) is often well beyond safe long term levels.

That is not the problem IME.
The dynamic range of the music they play is nowhere near 16 bits so is not a problem to record it is just that, in order for the music to be audible throughout the concert hall, it is of necessity horribly loud for the musicians themselves, even the bits that are fairly quiet for the audience are quite loud on stage.

The dynamic range of the music isn't the problem for the musicians themselves, it is the level itself.

 

[Neddy]

From comments he's made, his generation (of concert musicians) were 'too late' to recognize the risk, or have technical solutions available, and so now suffer.
I think he mostly played cello, but sitting in the midst of the horn section for the french horn work may have been worse on his ears- just a guess.

And he is a (now retired) expert in control theory (and the like), so should have been aware of the risk he was taking.
Interestingly, he seems perfectly fine with the 'price' he paid, and still loves to play when he can...though I think he may spend more time with teaching some youngsters (anyone under 80) now.
I also suspect that not being able to hear the constant yackety blather from most folks fits his pyschology well, too.
(I lost hearing in one ear for a while this year, and found the quiet on that side oddly comforting...)

I had a baby grand in my apartment for a few years.
I could not believe how loud that thing could get. Respect (fear?) for the other tenants forced me sell it soon after.
Ah...the Keith Jarrett Solo years.....

 

[fatoldgit]

Like @sergeauckland says, no point whatsoever.
I bet most of the audiophiles, who waffle on about dynamic range, have never heard the real thing.

You don't need 120db DR or even 96db DR to play Diana Krall or Patricia Barber...which seems to be audiophlle common fodder.

Peter

 

[EERecordist]

I have only been in a anechoic room once, it was in Japan. Maybe the next extreme wealth audio acquisition are extremely quiet rooms. My house is in a suburb and has a noise floor of about 29dB(A). Microsoft has the quietest measured anechoic room:

"HBK and BlackHawk Technology Inc. measured the noise floor of Microsoft’s quietest anechoic chamber at –20.6 dB(A) SPL. The quietest level of noise theorized by mathematicians is Brownian motion – the movement of particles in a gas or liquid – at –23 dB. The measurement method was specified by Guinness and used a two-microphone coherent power measurement technique with two Type 4955 low-noise microphones. The acousticians measured the same overall dB(A) multiple times."

In a concert hall, or even a recording studio, you may have the sound of the HVAC system, all the audience or musician rustling, coughing, and so forth. Your best microphones have self noise which has its own way of measuring. In close microphone placement on each instrument you may be able to exercise the full microphone range, then the preamp range, and the A-D converter range.

An isolated sound conditioned mixing control room might be 20dB(A).

To the original question, It's hard to get even close to needing 24 bits in home listening. If you are in a stadium or outdoor concert, the SPL at the speaker may be high, but the ambient noise floor at the listening position is also high.

 

[Ropeburn]

Sva

This thread is full of good explanations... If there is any difference between a 16 and 24 bit version of a track, the differences only start to appear below -96dBfs... If you've ever heard anything that quiet in a recording, your system was probably turned up to a sc

Indeed.

The downside of this hifi hobby is that half of the new wave from my youth has become virtually unlistenable now!
Ho hum.
(First world problems)

May I suggest the new new wave at this point?

Cyberwave. With updated aesthetics and sound quality.

 

[LionIT]

Normally I listen to music at about 85dB SPL. When I lower the volume by -35dB I hardly hear anything... so about 8 bits are more than enough to me.

 

[Ropeburn]

To reinforce what many have said in recent posts: we as humans with our fleshy audio sensors need surprisingly little sound level and dynamics for a satisfactory emotional response from our brains.

Take the 16-bit resolution where 96dB is maximum and 0 is absolute silence. Now, where is this "absolute silence" for our ears? Depending on circumstance and background (real life) noise, it can very well be some 30-40 real life decibels. Below which we hear exactly nothing. Add 96dB dynamic range, or even just the 80-90 that the average hifi/studio system could do 30 years ago, and you have anything well in excess of 120-130dB, the absolute pain threshold. From absolute silence to pain level, that's super satisfactory already and live concert "getting your socks blown off" category. In other words: the height of human experience.

You can play music very loud on any reasonably good older system just fine, without EVER running out of relevant dynamic range. Our recording and playback devices have been better than our (limited) ears for decades in that regard. A reminder: vinyl was already good enough for that - enjoying loud music -, with its mere 50-60dB dynamic range under laboratory conditions. IRL with some dirt and cartridge wear more like 20-30. And it still sounds fine and very enjoyable to this day.

Because our ears stay as limited as they always were.

 

[Frank Dernie]

Normally I listen to music at about 85dB SPL. When I lower the volume by -35dB I hardly hear anything... so about 8 bits are more than enough to me.

At an enthusiasts show several years ago @Pluto converted a 96/24 file from Eric Whitacre's Waternight to 8-bit with noise shaping and most of the people listening didn't believe they were listening to 8-bit, some were angry and refused to believe it. Without noise shaping the hiss was a bit much in quiet bits.

 

[LionIT]

A recent article from Izotope, very concrete and informative.
As can be seen, in the production phase it can be useful to have higher bit depth (and sample rate) to keep a bit of headroom.
For streaming services that offer 24-bit audio, it can only be explained with marketing to me.
Lossless is already hardly justifiable given the sophistication of modern lossy encoders.
High sample rates and bit depths inevitably involve higher costs to provide the service (storage, bandwidth, electrical consumption).
Spotify, with its market share, I think is an example of the fact that all of this is a waste of resources.
But I understand that to make a profit you have to differentiate yourself and/or invent the needs... like MQA.