The Shoutometer

[RayDunzl]

That's quite loud in your room, then.

Using 10dB less in this room:

Level C-weighted Fast Maximum - 96
Unweighted peak - 109

The jet part:

I might estimate 119dB unweighted peak at your place...

 

[BYRTT]

For info the jet take off analyzed in JRiver MC read out DR of 18, most of my collection read in area 5-14, but there are few album or tracks more than 14 and in below list there are some examples, unfortunate because of rights i can't share the 32 DR track but think its fun and educating listening, its a Chesky test recording available over at HDtracks.

 

[RayDunzl]

Here's a visual of the recording, in dB and "volts".

The middle of it is -55 to -65dBfs

 

[MRC01]

That's quite loud in your room, then.
...
I might estimate 119dB unweighted peak at your place...

106 dB SPL measured on my meter is the speaker limit; any louder, and the Mag 3.6/R panels start to rattle against the stops, due to those super low frequencies. The amp (Adcom 5800) is close to but still below its rated continuous power; its 1% lights do not flicker.

For the first 30 seconds, the track slowly ramps up from -90 to about -65 dB. So the test is, if you set the max volume to the highest your system, or your ears, can stand: once you hit play, at what point (how early) can you hear the quiet park sounds?

PS: theoretically, each speaker makes 86 dB SPL @ 1 meter with 1.9W of power. The amp is rated at 400W continuous both channels driven, which is 23 dB louder, which is 109 dB SPL. Plus 3 for 2 speakers = 112 dB SPL. So your estimate of 119 dB is in the ballpark but seems a bit high.

 

[RayDunzl]

I'm not sure what you're getting at...

If you turn up the volume enough, yeah, you can hear -85 relative to the new unlistenably loud level.

I played a 1khz tone at 95dB SPL measured at the listening position and then reduced the signal by 85dB (digitally), and it was gone, below the noise floor in the room measurably, below audibility close to the panel speaker.

YMMV of course.

PS: theoretically, each speaker makes 86 dB SPL @ 1 meter with 1.9W of power. The amp is rated at 400W continuous both channels driven, which is 23 dB louder, which is 109 dB SPL. Plus 3 for 2 speakers = 112 dB SPL. So your estimate of 119 dB is in the ballpark but seems a bit high.

My guestimate vs your new numbers:

86dB for one speaker, 92dB for two playing the same signal, add three for Peak vs Fast = 95, add 23dB for the amplifier power increase = 118 vs 119 guessed... Ok, I got lucky.

Original estimate based on my experience looking at Peak vs Fast for general music program using REW and a UMIK-1.

 

[MRC01]

I'm not sure what you're getting at...
If you turn up the volume enough, yeah, you can hear -85 relative to the new unlistenably loud level.
I played a 1khz tone at 95dB SPL measured at the listening position and then reduced it by 85dB, and it was gone, below the noise floor in the room measurably, below audibility close to the panel speaker.
...

Exactly, that's my point! To get a realistic estimate of the usable dynamic range in your listening room, you can't just crank it up until you hear a low-level signal.
With this track, set the loudest part to the maximum volume that you can actually play and hear in your listening room. Then, at that volume, start the track from the beginning and see at what point you can hear it.
Note: we can hear correlated signals below the ambient noise. If the max is 100 dB and the room measures 40 dB when quiet, it doesn't mean you're limited to 60 dB of dyn range.

 

[danadam]

 

[GrimSurfer]

I forget if it were on Secrets of the Dead or maybe Nova. There was a famous speech given by a well known orator in the days of the Revolutionary War in what would become the USA. I seem to recall it was in Boston. Was said the fellow was heard much further away than one would expect, and seems maybe down on the waterfront it was said to be heard. I wished I could find the show. Much more interesting than I am making it sound. In any case, they investigated it and found that yes it was reasonably probable that the speech was heard thru winding streets two miles away or some such crazy distance. The winding streets are a big part of it. Sound was reflecting and spreading less than open air. Plus the orator was a regular speaker with a truly loud voice. They did do measurements to confirm the plausibility of how this would happen.

Not surprised it happened in a coastal city. It could easily be due to a layer of denser, more humid air, above ground level causing ducting.

Ducting is quite common in maritime radar propagation. Obviously, the frequencies are much higher (GHz range) which makes them usually more prone to atmospheric attenuation. Regardless, airborne and surface search radar detection ranges can increase considerably when ducting conditions are present. The same principle applies across the EM range.

It could also be higher air density facilitating sound transmission... the old ear to the rail line trick.

That would be a cool Nova episode... post it if you ever find the link.

 

[MediumRare]

I have a demo disc that starts with sounds recorded in a quiet park at -85 dB, a couple minutes later smoothly transitions to a jet taking off at digital 0 dB. I can hear the full range from my listening position at a single volume setting. But I can't hear someone shouting from 5 or 10 miles away.

My family had a farm in Wisconsin in the 1970s. I recall on a very cold, clear winter morning, hearing clear sounds (conversation, car doors) from a town 2 miles away. Of course, diesel locomotives can be heard from many miles away.

 

[Ron Texas]

I could do a million K's.

 

[DuaneInHiding]

I’m jumping over from a Newbie thread I started a few days ago because this Shoutometer thing has really, really helped me get my head around some big concepts -- "is this DAC really that much better than this other one," "how much of a difference does this amount of noise really make," etc. -- and so I took the concept and did some bad Photoshopping to try and make things even clearer for visual people like myself.

What I’ve realized in this process is that logarithmic graphs in general just break my brain:

To me, a graph like this looks like both points are closer to each other than they are the top. It looks like the orange dot (representing the lowest threshold of the 2nd-to-worst group of measured DACs) is only about 25% noisier than the blue dot (representing the best SINAD score recorded so far).

But, enter the Shoutometer!

Now that I can start connecting spatial relationships on the graph with real-world distances, things begin to take shape.

In my city there are two bridges about a mile apart. I went for a walk today and remembered that this distance in the Shoutometer represents a -66 dB drop. So I stood in the center of one bridge and tried to imagine a person (who I could probably only barely see) shouting at me from the other. And it seemed…possible, but not probable that I could hear them.

Put it this way: if I agreed to meet a friend on Bridge A at noon, and they showed up an hour late but said, “I stood on Bridge B and shouted to you that I was going to be delayed,” I don’t think it would do much to make me feel better. There’s a chance that I could hear something that far away if I was concentrating really hard, but there’s a much bigger chance that it wouldn’t register.

So for me, personally, I think -66 dB represents the boundary of an informal kind of audibility. Definitely not the absolute threshold of human hearing, but down in the weeds where I may or may not hear something. (For perspective, only five DACs tested by this site have ever turned in a SINAD that bad or worse.)

After my walk, I decided to see how far I could push this concept. I used New York City in the hopes that plenty of people have visited there at one point or another.

Pretend there’s somebody standing in Battery Park at the very southern tip of Manhattan, facing north and shouting his head off. 1.25 miles places you about 25 blocks north, on Canal St. Maybe you can hear him, maybe you can’t. That’s -66 dB.

But from what I’ve read here, nobody’s shooting for those numbers. So we move back to the very lowest threshold to get out of the “red” level of SINAD — -85 dB below peak.

That would be like standing in Yankee Stadium and hearing somebody shout to you from Battery Park. Personally, I can’t get myself to imagine that being possible — even though this is still technically way louder/closer than what should technically be “inaudible.”

Finally, we move back to the best DAC measured so far: -121 dB:

That’s like somebody standing on the southern bank of the St. Lawrence River in Canada, facing south and hearing a shout coming from Battery Park. I would concur that this would be inaudible. (And I feel like -120 dB is actually the technical definition of inaudibility?)

So, once again all together:

What’s amazing to me about this is that it basically proves everything true: the cleanest measured DAC is way, way “farther away” than the second-to-worst level. But…they’re probably both too far away for me to hear a guy shout from.

Anyway, that’s what I’ve got. Again, my gratitude to RayDunzi for starting this whole thing. Hope these images help somebody else connect the same dots…

- Duane

 

[MRC01]

The shoutometer is a good example to bring the concept home, yet another reason it seems so crazy is that it assumes a perfectly quiet world with perfect atmospheric conditions (wind, pressure, humidity, temperature) which of course doesn't exist. I suspect also that reality isn't as clean as the math, because I have some wide dynamic range test sounds where in my listening room I can hear at least 80 dB below the peak.

That is, it starts very quiet and gradually (over 2-3 minutes) ramps up to full scale 0 dB. If I set the volume so that full scale is the loudest volume I and my system can withstand (about 105 dB SPL), leave it there and play it from the start, I start to hear the lowest level sounds in the recording when they pass through about -80 dB. So my listening room has about 80 dB of usable range. Of course, music recordings that use this much dynamic range are rare approaching virtually nonexistent. And if they existed, they would be hard to listen to because you'd go from near painful levels to squinting to barely hear it. Furthermore, your hearing can't use that wide a range in a short time window. When sounds get loud, your ears & hearing deploy biological compression mechanisms. So your short time window usable dynamic range is something like 30-40 dB. But the point is, 80 dB of range is a lot, but is audible/usable, perhaps even more than 80 dB for some people, in other rooms quieter than mine.

 

[L5730]

Someone here has said the low level stuff some claim to hear means they could hear an autumn leaf drifting to the ground standing next to a 747 at takeoff. Sort of puts things in perspective like this shoutometer does.

Well, some people see ghosts, so...? (and they don't all see Bruce Willis!)

That sample FLAC file puts things into perspective. What we think is a quiet room is very much not quiet at all. As humans have evolved, we've learnt to become more tuned to certain sounds (human voices) more than anything else, I think. Once basic individual survival was taken care of with farming and fence building, fire creation and containing, then we really had no need to be able to hear the footsteps of a deer or a boar and know the difference. Modern humans are an amusing species - c'mon we really are just one of a kind. I tend to think of humanity in terms of a 'creator' holding the mould for humanity and then looking at what we are now, and then in sheer disbelief dropping it and it smashing into a gazillion shards.

So, hang on. To the OP and any more practical science based folks. If we take any wind out of the equation, and assume a person is standing in an utterly flat desert at 298 K (~ 25°C), is this what the stats are showing, or is it something else? Sound travels better through things than air - even primary school kids know this (two plastic cups and a bit of taught string).
Usually, in science we assume things are in a vacuum to simplify stuff, but sound doesn't travel in a vacuum, does it? Otherwise the phrase "in space, no one can hear your screams" just begs to be nerd corrected .

I do sincerely apologise if I am being troublesome. I'll blame this on whomever invented the wonderfully flavoured and ethanol-laced substance I am drinking.

 

[davidc]

This audio site doesn't let us post audio clips - ah, the irony of it!
Try this: http://mclements.net/audio/sampleDynRange.flac
Here's what it's from, a fun album from the 80s when CD was a new medium: The Digital Domain - A Demonstration
I got this back in the day to play on my first CD player, a Technics SL-P1. The original first track only used 12 bits of dynamic range, and it had a bit of DC offset. That's because when it came out in 1983, many CD players only had 14 usable bits of dynamic range. I modified the track to eliminate the DC offset and boost the dynamic range (make the park quieter).

I thought I remembered what you were describing. I had this disc also.

 

[Robbo99999]

The shoutometer is a good example to bring the concept home, yet another reason it seems so crazy is that it assumes a perfectly quiet world with perfect atmospheric conditions (wind, pressure, humidity, temperature) which of course doesn't exist. I suspect also that reality isn't as clean as the math, because I have some wide dynamic range test sounds where in my listening room I can hear at least 80 dB below the peak.

That is, it starts very quiet and gradually (over 2-3 minutes) ramps up to full scale 0 dB. If I set the volume so that full scale is the loudest volume I and my system can withstand (about 105 dB SPL), leave it there and play it from the start, I start to hear the lowest level sounds in the recording when they pass through about -80 dB. So my listening room has about 80 dB of usable range. Of course, music recordings that use this much dynamic range are rare approaching virtually nonexistent. And if they existed, they would be hard to listen to because you'd go from near painful levels to squinting to barely hear it. Furthermore, your hearing can't use that wide a range in a short time window. When sounds get loud, your ears & hearing deploy biological compression mechanisms. So your short time window usable dynamic range is something like 30-40 dB. But the point is, 80 dB of range is a lot, but is audible/usable, perhaps even more than 80 dB for some people, in other rooms quieter than mine.

So, hang on. To the OP and any more practical science based folks. If we take any wind out of the equation, and assume a person is standing in an utterly flat desert at 298 K (~ 25°C), is this what the stats are showing, or is it something else? Sound travels better through things than air - even primary school kids know this (two plastic cups and a bit of taught string).
Usually, in science we assume things are in a vacuum to simplify stuff, but sound doesn't travel in a vacuum, does it? Otherwise the phrase "in space, no one can hear your screams" just begs to be nerd corrected .

I do sincerely apologise if I am being troublesome. I'll blame this on whomever invented the wonderfully flavoured and ethanol-laced substance I am drinking.

I think part of the reason is that the Shoutometer is overestimating the distances because it's not taking air atmosphere absorption into account so it's making it sound quite a bit more hyped up and unreasonable to expect better SINADS from audio gear. For instance at this link (http://www.sengpielaudio.com/calculator-air.htm ) if you take telephony frequency of 300Hz to 3000Hz to describe a human voice and plug some different frequencies into the calculator at that link you have for instance the following to describe just the effects of the atmospheric air absorption variable part of the overall equation:
0.2dB/100m of attenuation at 300Hz
1.8dB/100m of attenuation at 3000Hz
0.8dB/100m of attenuation at 1650Hz
So we could call it maybe 1dB/100m of attenuation just by the air itself on the identification (not necessarily intelligibility) of a human shout out in the real world.

If we factor the 1dB/100m into the shoutometer table then lets look at the -60dBFS reading that is showing 1km (or 1000m) - so if we factor in the 1dB/100m then we have to add another 10dB of attenuation onto that, so that would make it to be actually -70dBFS at 1km. Another example is the 500m one would turn into -59dBFS. These effects get even greater magnified the further you go down the shoutometer scale as distance is increasing logarithmically. So really the air absorption has a big effect. So I think this puts the table a little more into perspective and might go some ways to explain how you (@MRC01 ) can hear differences when it seems absolutely unreasonable from the shoutometer table. Just some thoughts and rough calculations I've worked out....haven't really worked out all the implications on what we're talking about, but figured it was significant to point out given it's large effect on the table, especially at the lower dBFS.

 

[pjug]

Interesting that the human shouting is only about 78dB at 3 feet
https://www.engineeringtoolbox.com/voice-level-d_938.html

 

[Χ Ξ Σ]

Rob Watts Hearing Acuity

 

[Chromatischism]

We also forget that in our sound reproduction systems, we aren't hearing these things in isolation. Noise or distortion at -80 dB should be completely masked by the fundamental frequencies simultaneously playing.

 

[Hipper]

The shoutometer (how is this pronounced? Shoutometer or Shout'o'meter?) should be applicable for our use, which is in listening rooms. Therefore 84dB should be thought of as 10.2 miles or so in a listening room, not outside.

There is of course a time factor to sound travel. To travel a distance of 10.2 miles at sea level takes sound about 48 seconds.

Anyway, I'd like to order the Shoutometer Deluxe please. I notice that Andromeda is -444dBfs. I know someone who has heard sounds from Andromeda:

 

[RayDunzl]

Measures in terms of signal power...

Full Scale = one unit

-10db = 0.1
-20dB = 0.01
-30dB = 0.001
-40dB = 0.0001
-50dB = 0.00001
-60dB = 0.000001
-70db = 0.0000001
-80dB = 0.00000001
-90dB = 0.000000001
-100dB = 0.0000000001
-110dB = 0.00000000001
-120dB = 0.000000000001
-130dB = 0.0000000000001
-140dB = 0.00000000000001
-150dB = 0.000000000000001
-160dB = 0.0000000000000001
-170dB = 0.00000000000000001
-180dB = 0.000000000000000001