[sonitus mirus]

Again, none of this data can be used in this discussion. Please see this footnote for reasons why in your above graph:



As you see, it is a slow reading meter which means it is not peak SPL numbers. And further, A-weighted which is not what we store in your music files. We store actual/peak values of audio samples, not any kind of weighted or averaged values. As such, the channel must deal with the peak numbers as the highest value which as a rule are 5 to 10 db higher than average numbers indicate.

How would anything you just posted change the fact that your list showed a maximum exposure of 12.5% of an hour at 120 dB? A peak SPL makes no sense for an exposure at over 7 minutes. How can any peak last 7+ minutes that is not continuous?

 

[amirm]

1. Exactly, hallelujah brother! Notice I've highlighted the "could be". If the orchestra are dead and make zero noise, if we're in a venue with zero noise, if we take the damage thresholds of human hearing, if we only use the mic with the largest dynamic range and "close mic" with them, use zero noise amps/mic pre-amps, etc., then YES, WE COULD achieve the numbers you are talking about and we would all be wrong about everything. HOWEVER, ...

As I noted, in my last post, and have said over and over again, you cannot use peak SPL numbers with regards to what causes hearing damage. Hearing damage metrics use averages and importantly include a duration which is orders of magnitude longer than what is in dynamic music.

Now, if you sit there in mastering and boost the hell out of average loudness level and makes them just a hair less than peaks, then sure, your ears will bleed and complain. The solution to that is not to advocate for less bits but to fix mastering.

So there is no "HOWEVER" here. After so many posts we still see the same misunderstandings about SPL numbers repeated.

Even if we go along with what you are saying, it is dead against your original post that started this thread:

I know that some people are going to say this is all rubbish, and that “I can easily hear the difference between a 16bit commercial recording and a 24bit Hi-Rez version”. Unfortunately, you can't, it's not that you don't have the equipment or the ears, it is not humanly possible in theory or in practice under any conditions!!

That is wrong, isn't it? You are not only allowing the theory to show that, but also in practice in some cases. Nothing pointing to impossibility.

Are we going to see an edit to that first post that evidence to the contrary has been provided?

 

[gregorio]

I found conflicting data with the link to noise exposure as provided earlier from OSHA.

Noise exposure levels are designed for industrial purposes, where the noise is fairly constant, say in a factory, and even for that they're not very good and some countries have laws lower than the OSHA recommendations. None of this is really applicable to music listening except obviously that any duration of 140dBSPL is dangerous. What we do have that is relevant to this line of discussion is some studies which show hearing damage of orchestral musicians, one of which I posted a few pages back but has been singularly ignored by certain parties with an agenda. Obviously, listening to orchestral music at the actual real life levels that an orchestra produces is potentially dangerous, much less so for an audience who get lower peaks SPLs than those close to/in the orchestra, far more compressed music genres would likely be more dangerous still at those same peak levels, as there are likely to be far more of them! For anyone out there, even if you have the equipment and inclination to do so, do not listen to music at 120dB SPL peak levels!!!!!

G

 

[amirm]

How would anything you just posted change the fact that your list showed a maximum exposure of 12.5% of an hour at 120 dB? A peak SPL makes no sense for an exposure at over 7 minutes. How can any peak last 7+ minutes that is not continuous?

It shouldn't make sense because that is not what I said! I have repeatedly said that the exposure metrics are average SPLs. But that aside, I also wanted to indicate that there is a duration and hence that chart. There are different charts by the way with different numbers so if that is what you are observing, that is fine. It is just not applicable to discussion at hand.

 

[amirm]

For anyone out there, even if you have the equipment to do so, do not listen to music at 120dB SPL peak levels!!!!!

You say that based on what peak SPL measurement or study?

 

[sonitus mirus]

It shouldn't make sense because that is not what I said! I have repeatedly said that the exposure metrics are average SPLs. But that aside, I also wanted to indicate that there is a duration and hence that chart. There are different charts by the way with different numbers so if that is what you are observing, that is fine. It is just not applicable to discussion at hand.

Ah, ok. Thanks. I admit, I was confused.

 

[amirm]

1. Our goal is to create the best experience we can BUT the best experience is ABSOLUTELY NOT a 0dB noise floor and peaks of 120dB!!! HOW ON EARTH can a potentially dangerous/damaging dynamic range be "the best experience"?

Once again, all of our impressions of what is "dangerous" is reading SPL charts online and not realizing they are a) averages b) weighted c) for wideband noise and not tones in music and d) they are rough figures and not based on any specific research. Importantly unless someone here has actually used a peak measuring technique -- which I assume no one has -- you are most likely exposed to such "dangerous" SPLs and actually liked it! I know I have.

In my entire library of research papers, I think I only have 3-4 papers that use peak SPL measurements. Every other place such data is averaged and hence not appropriate for this conversation.

On the noise floor, again, we cannot speak of single value SPL numbers. We must, must look at the spectrum. SPL numbers of 40+ db can be dead silent to us as a result! As I show in my article on quietness of our listening rooms based on Fielder's reference room, people have rooms that are this quiet:

The solid line is threshold of hearing. As we see, at 20 Hz it can be as high as 65 db and we still consider that dead silent!!!

The best measured room easily qualifies for such. This research is from 1990s. Today, there are many state of the art rooms built at far lower noise levels. So if survey was done today we would uncover many such rooms. And regardless, as long as these rooms can be built, then we owe it to those people to use a distribution format that delivers noise free experience to them.

As an aside, notice how the threshold of hearing at mid-frequencies where we are most sensitive is actually a negative SPL! Not zero. Our hearing is quite sensitive when it comes to that region (likely for the need to understand each other's voices).

Furthermore as I have explained, noise coming from speakers is point source and is more audible than ambiance which is diffused. As such, it is not sufficient for it to be the same as ambient noise to be inaudible. It needs to be lower.

 

[71 dB]

71, the dB scale is logarithmic, so the difference between 60dB and 120dB is a LOT more than double. And most LPs have dynamic ranges under 50dB... a lot of them are quite a bit lower than that.

Good call bigshot, but let me assure you logarithmic dB scale is very familiar to me. I mean double on logaritmic scale, 2 x 60 = 120. On linear scale 120 dB is of course a million* times more acoustic energy than 60 dB. Sorry, if I assumed too much other readers to know dB scale. I was trying to say that if you attenuate the loudest sounds of a vinyl 60 dB so that they became as quiet as the noise floor, it's easy to understand that the noise floor becames insanely quiet when attenuated 60 dB too and that's roughly the noise floor of CD with shaped dither. The noise floor of vinyl is in the "midway" of CD noise floor and 0 dBFS sounds on logarithmic scale.

Vinyl dynamic range being 60 dB is the best case scenario. Sometimes vinyl is given even 70 dB, but I don't believe that.


* 10^(60/10) = 10^6 = 1.000.000

 

[71 dB]

I was tempted to make the same remark as a joke. because I'll go on a limb here and bet that 71dB knows how dB work

That's a good bet man! The day I don't know how dB scale works is the day to die, because I have become too senile to live anymore.

 

[71 dB]

Again, none of this data can be used in this discussion. Please see this footnote for reasons why in your above graph:



As you see, it is a slow reading meter which means it is not peak SPL numbers. And further, A-weighted which is not what we store in your music files. We store actual/peak values of audio samples, not any kind of weighted or averaged values. As such, the channel must deal with the peak numbers as the highest value which as a rule are 5 to 10 db higher than average numbers indicate.

This is a good remark. Also, people want to listen to music at pleasing levels, not at the highest possible levels our ears can possible take. A pleasing level is of course subjective. Teenagers want as loud as possible, 160 dB of bass that makes their first car pulse of the EDM beats! Old people want their evergreens softer because their ears hurt easily.

 

[bigshot]

Listening to music with 120dB peaks probably won't damage your hearing, but it won't be comfortable. If there is a snare drum that hits 120dB with each hit, you'll be wincing with each one, and the music behind it will be so loud, you won't be able to hear anyone in the room speak to you, even if they are sitting right next to you. It's a great idea in theory to be able to reproduce 120dB, but it isn't a good idea to actually listen to music like that. If you doubt that, get a nice SPL level meter, set it properly and try it yourself. Listen to a whole album that way. I bet you don't even get through 30 seconds.

 

[71 dB]

As an aside, notice how the threshold of hearing at mid-frequencies where we are most sensitive is actually a negative SPL! Not zero. Our hearing is quite sensitive when it comes to that region (likely for the need to understand each other's voices).

Yes, between 3 and 4 kHz we have the ear canal resonance at 1/4 of wavelength. Due to this resonance, sound pressure levels below 0 dB are audible, but it work in the upper end too: The threshold of pain is lowered too for the same amount and goes under 120 dB at this frequency range. The dynamic range of hearing is pretty constant 120 dB between 800 and 4000 Hz and elsewhere it is less, for example at 100 Hz only about 80 dB!

 

[RRod]

Here's an example reverb tail from a non-live recording made at the symphony hall previously discussed. I've put random names on these test files (24/44100), but rest assured one of them is the original > 16-bit source. Once y'all have found a minimal level where you can hear some differences, I'll put up the loudest part of the album. For my own part, in my 35dBA bedroom, listening on my PM-3s (which isolate well >1kHz), with the volume cranked as loud as I care to for this album, I can maybe hear just a bit of difference in one file; whether that holds up in ABX I haven't tried.

 

[old tech]

Gnds on logarithmic scale.
Vinyl dynamic range being 60 dB is the best case scenario. Sometimes vinyl is given even 70 dB, but I don't believe that.
* 10^(60/10) = 10^6 = 1.000.000

70db would be a best case scenario and within the realms of possibility. Depending on the quality of the pressing, the recording, the position of the groove on the record and the frequency, 70db is possible. That of course also assumes that the T/T & cart is able to track that specific "sweet spot" with some degree of accuracy.

 

[bigshot]

i have a lot of audiophile vinyl, and I doubt any of it goes much beyond 50. thats in practice though, not in theory.