How much dynamic range is too much

[Miguelón]

I'm not sure what you mean by "in a time function". That graph shows FFT magnitude, so it's a function of frequency.

Yes, I missed the graph, was looking to the peaks in the first graph when tapping the answer, but was referres to the frequency graph.

As (IIRC) @AnalogSteph or @staticV3 wrote I think is important to listen in an apporpiate silent room to high dynamic range music, if not all of this algorithms are futile I suppose

 

[Miguelón]

in particular

Here is the same principle for a recording chain (from here):

What happens reversed?

If my AOP is 100 dB (calculating 80 dB and 20 more for transients), then should align scale to 0 dBFS to 100 dB SPL.

Then the sensitivity is 106 dB in by default setting, so +6 dBFS should correspond to 106 dB, or 0 dBFS to 100 dB SPL

One of my DACs (WiiM) produces 8 dBU at 0 dBFS so I should reduce its output to around 0.77 V (I have 0.8 V setting in the output voltage, I assume this is a good value, either I have 0.5 V as a possible reasonable alternative to 70 dB SPL + 20 dB transients)

I’m now correct or again messing things up?

In this case with 120 SNR my noose floor is totally imperceptible with respect to ambient or human threshold

EDITED: not very sure about the extra 20 dB for transients listening at an average of 70 (or 80 in a party day) dB. Was suggested by another member but I don't know if its valid for classicsl music as it has more sudden peaks

 

[rdenney]

Going back to the original question:

This paper: https://www.akutek.info/Papers/MS_Sound-Levels-symphony-orchestra-musicians.pdf

...described the actual SPL of orchestra performance, from the perspective of the musicians and audiences, for the purpose of determining risk of hearing damage.

In that paper is this observed histogram:

Note that these are not true peaks, but rather 100ms samples, which is too long for a true peak (and which is the only reason the histogram looks so bell-curvish, despite the slight positive skew). It is limited to a band from 400 Hz to 2.4 KHz. Note that for the average in the middle 60's dB SPL (in the constrained band), the peaks barely reach 100 dB. The quiet bits, however, reach down to around 15 dB, though I suspect where the chart levels off is mostly ambient noise. Where the probability starts to increase is in the low 20's. That represents a dynamic range for 100 ms peaks of about 75-80 dB.

I have measured peaks from my seat in a large performing ensemble, sitting in front of the timpani, at 108 dB SPL A-weighted. Assuming the above scales, the quiet bits would be at about 40 or 45 dB SPLA, which seems to me about right and validates the range in my mind.

I doubt that such ranges are ever recorded as such, which is, in my view, why recordings never sound quite realistic. The times I've made live recordings on media that could capture dynamic ranges greater than about 75 dB, and as long as I did not clip the peaks, the recordings (in headphones) sounded truly realistic. My first experience with that was using a VCR to record sound on VHS using the HiFi depth multiplexing. Part of what made it so was the clarity of the ambient room noise in the recorded signal. This is often gated off in commercial studio recordings, which to my thinking is not a good thing.

When we are listening through speakers, that recorded ambient noise (if it's there) gets mixed with the ambient noise in our listening environment, which may or may not have a salutary effect. Less so with headphones.

So, I would like to see dynamic ranges of something like 75 dB or greater in recorded orchestral music, for when I'm listening only (and not doing something else with music in the background). Even then I'll probably have to gain-ride the playback or use headphones (which I detest unless I'm on an airplane, and even there they are the lesser of evils). If I listen to that recording in the car, the quiet bits are hopeless, even when the loud bits are ear-splitting. But that is true even for regular commercial classical recordings.

For these reasons, mastering applies compression so that people can listen in noisier environments without damaging their (lower-grade) systems, clipping their tiny amps, or dipping down into noise levels low enough that the listener feels the need to gain-ride the playback. Recordings and their dynamic ranges have already been covered in this thread. But this is why distributed recordings never seem to capture live-sound realism.

Rick "answer: >75 dB measured without nuance, and CD's are more than good enough" Denney

 

[olieb]

The quiet bits, however, reach down to around 15 dB, though I suspect where the chart levels off is mostly ambient noise. Where the probability starts to increase is in the low 20's. That represents a dynamic range for 100 ms peaks of about 75-80 dB.

I do not know about the context of these captures, but I would expect a bell curve (to some degree) for the ambient noise too, in particular if the captures are from different performances, ensembles and halls (and seats?). This could mean that actual dynamic is a bit lower when considering the music signal only. 75-80dB seems a lot to me, even if one sits in a very frontal seat.

Part of what made it so was the clarity of the ambient room noise in the recorded signal. This is often gated off in commercial studio recordings, which to my thinking is not a good thing.
When we are listening through speakers, that recorded ambient noise (if it's there) gets mixed with the ambient noise in our listening environment, which may or may not have a salutary effect. Less so with headphones.

Interesting line of thought. The ambient noise will be injected into the room by the speakers though. So if one hears it, this will probably be from the speaker positions. Don't know what is the better solution. But I hate it if the gating itself becomes perceivable. (More a thing for live performances).

 

[Miguelón]

Going back to the original question:

This paper: https://www.akutek.info/Papers/MS_Sound-Levels-symphony-orchestra-musicians.pdf

...described the actual SPL of orchestra performance, from the perspective of the musicians and audiences, for the purpose of determining risk of hearing damage.

In that paper is this observed histogram:

View attachment 433449

Note that these are not true peaks, but rather 100ms samples, which is too long for a true peak (and which is the only reason the histogram looks so bell-curvish, despite the slight positive skew). It is limited to a band from 400 Hz to 2.4 KHz. Note that for the average in the middle 60's dB SPL (in the constrained band), the peaks barely reach 100 dB. The quiet bits, however, reach down to around 15 dB, though I suspect where the chart levels off is mostly ambient noise. Where the probability starts to increase is in the low 20's. That represents a dynamic range for 100 ms peaks of about 75-80 dB.

I have measured peaks from my seat in a large performing ensemble, sitting in front of the timpani, at 108 dB SPL A-weighted. Assuming the above scales, the quiet bits would be at about 40 or 45 dB SPLA, which seems to me about right and validates the range in my mind.

I doubt that such ranges are ever recorded as such, which is, in my view, why recordings never sound quite realistic. The times I've made live recordings on media that could capture dynamic ranges greater than about 75 dB, and as long as I did not clip the peaks, the recordings (in headphones) sounded truly realistic. My first experience with that was using a VCR to record sound on VHS using the HiFi depth multiplexing. Part of what made it so was the clarity of the ambient room noise in the recorded signal. This is often gated off in commercial studio recordings, which to my thinking is not a good thing.

When we are listening through speakers, that recorded ambient noise (if it's there) gets mixed with the ambient noise in our listening environment, which may or may not have a salutary effect. Less so with headphones.

So, I would like to see dynamic ranges of something like 75 dB or greater in recorded orchestral music, for when I'm listening only (and not doing something else with music in the background). Even then I'll probably have to gain-ride the playback or use headphones (which I detest unless I'm on an airplane, and even there they are the lesser of evils). If I listen to that recording in the car, the quiet bits are hopeless, even when the loud bits are ear-splitting. But that is true even for regular commercial classical recordings.

For these reasons, mastering applies compression so that people can listen in noisier environments without damaging their (lower-grade) systems, clipping their tiny amps, or dipping down into noise levels low enough that the listener feels the need to gain-ride the playback. Recordings and their dynamic ranges have already been covered in this thread. But this is why distributed recordings never seem to capture live-sound realism.

Rick "answer: >75 dB measured without nuance, and CD's are more than good enough" Denney

Very interesting, in my pianist student epoch I went to many concerts (most of them to listen, not to play).

After changing my orientation to other career I began to listening a lot of classical at home, because I hadn’t enough time or money to concerts.

Years after, when I go to a concert hall I always find the orchestra like a bit “anemic” in dynamics, and try to get a chair on the first row. But even at this place I find low dynamic from fist minutes.

Quite probably is because too many listenings with altered dynamic at home gave the habit to expect very loud crescendos, specially by headphones which can reach high dynamic range more easily.

 

[rdenney]

I do not know about the context of these captures, but I would expect a bell curve (to some degree) for the ambient noise too, in particular if the captures are from different performances, ensembles and halls (and seats?). This could mean that actual dynamic is a bit lower when considering the music signal only. 75-80dB seems a lot to me, even if one sits in a very frontal seat.

In my experience, I've been on stage when some parts of the performance were extremely quiet in my ears--easily no more than 30 dB and even then in the most sensitive frequencies--and other parts measured as I have reported. That difference would reach 75 dB easily, but, of course, that's on stage. As an audience member, everything is scaled down, but not necessarily linearly--the hall doesn't send sound to the back ensemble row as efficiently as to the audience, if there is a proper shell.

But I have recorded live performances using open-reel tape with a S/N of perhaps 58-60 dB (using a Teac Tascam half-track deck in good nick), HiFi VHS audio that is about 75 dB, and digital with the dynamic range reachable by the microphones and other electronics. The sense of reality for me wasn't reached by the open reel tape, but was with the VHS HiFi tape. That seems to validate that 75 dB range.

Yes, it's a lot. But we also have to recognize that the histogram may not be the same shape in the recording--a recording will compress at the peaks and will thus show more positive skew. That will increase the median with respect to the extremes, particularly the peak extremes, with probably a greater audible effect than shaving on the peakiest parts of the peaks.

Rick "recordings almost never sound completely realistic" Denney

 

[rdenney]

Very interesting, in my pianist student epoch I went to many concerts (most of them to listen, not to play).

After changing my orientation to other career I began to listening a lot of classical at home, because I hadn’t enough time or money to concerts.

Years after, when I go to a concert hall I always find the orchestra like a bit “anemic” in dynamics, and try to get a chair on the first row. But even at this place I find low dynamic from fist minutes.

Quite probably is because too many listenings with altered dynamic at home gave the habit to expect very loud crescendos, specially by headphones which can reach high dynamic range more easily.

Don't confuse dynamic range with loudness. When we listen to amplified playback, we often listen at greater SPL than live performance, particularly with unamplified acoustic music. That raises the low levels, too, but I doubt we notice that as much.

But the main reason playback may sound louder is because the average is higher with respect to the peaks--that's the result of the Loudness War which stemmed from trying to make FM radio signals sound louder without overmodulating the peaks.

Rick "almost never plays back at stage levels, but designed his system to be able to" Denney

 

[olieb]

The sense of reality for me wasn't reached by the open reel tape, but was with the VHS HiFi tape. That seems to validate that 75 dB range.

I am sure you are right with this. But there is a difference in "dynamic" when comparing snippets of 100ms and "dynamic" as in S/N. The peak level in these 100ms can easily be 15 dB higher (but needs not be) than the average (I assume here that this is what the diagram shows. Maybe I did get that wrong).
Wouldn't one need more "technical" dynamic than 75dB for recording 100ms music at 24dB and 100ms music at 100dB if one does not want to have serious distortion/compression of these peaks?

 

[Miguelón]

Don't confuse dynamic range with loudness.

Is the third time I was corrected for the same mistake in this thread… sorry, too many years reading audiophile articles where dynamic range means “from quieter to louder” and the DAC signal is a “stair step” scaled figure.

It takes time to forget wrong concepts…

Thanks for correction

 

[krabapple]

Totally agree, many Deutsche Gramophon recordings of hight talented conductors ruins the symphonies by placing the mics at 2 km away from the scenario, recovering nice eccoes and halo sounding but a sort of musical creme from the orchestra

? I thought DG, from say the 1970s on, and Karajan;s in particular, was famous for its highly multimiked recordings?

 

[MRC01]

... But the main reason playback may sound louder is because the average is higher with respect to the peaks--that's the result of the Loudness War which stemmed from trying to make FM radio signals sound louder without overmodulating the peaks.

Fortunately, classical music is largely outside the blast radius of the loudness wars. Modern classical recordings tend to have more dynamic range than in days of yore, not less. That's not to say classical music recordings don't use dynamic compression, most of them do and experienced listeners can hear it, but they use a lighter hand than the ham-fisted "crank it to 11" we get with pop/rock.

 

[rdenney]

Fortunately, classical music is largely outside the blast radius of the loudness wars. Modern classical recordings tend to have more dynamic range than in days of yore, not less. That's not to say classical music recordings don't use dynamic compression, most of them do and experienced listeners can hear it, but they use a lighter hand than the ham-fisted "crank it to 11" we get with pop/rock.

Even so, I hear substantial modulation "optimization" even on the clean-sounding local FM classical station. This isn't in the recorded signal like it is with brickwalled pop, but it is an example of the same attempt to increase the average output during the quiet bits so that a listener doesn't have to keep their right hand on the volume knob while driving down the road.

Rick "still has a working tuner--actually several" Denney

 

[MRC01]

... Rick "still has a working tuner--actually several" Denney

Hey, I resemble that!
And I hear what you're talking about. I figure they feel they should do it given the nature of FM broadcast and the environments in which people listen to it.

 

[Miguelón]

? I thought DG, from say the 1970s on, and Karajan;s in particular, was famous for its highly multimiked recordings?

And others, but yes Karajan was on the extreme of smoothie’s

The other extreme can be Osmo Vänskä recording of Beethoven symphonies that was suggested by a member up in the thread: the “Allegro con brio” that starts the Eroica (symphony n.3) seems to be recorded to awake an audiophile from his coma…

It worths to take a look on the crescendos, is impossible to an orchestra to produce that without serious wrist injuries

The orchestra is Minessota and the seal is Super Audio CD

 

[Robin L]

Even so, I hear substantial modulation "optimization" even on the clean-sounding local FM classical station. This isn't in the recorded signal like it is with brickwalled pop, but it is an example of the same attempt to increase the average output during the quiet bits so that a listener doesn't have to keep their right hand on the volume knob while driving down the road.

Rick "still has a working tuner--actually several" Denney

That's because you're listening to FM. That always happens on FM, at least these days.

 

[Miguelón]

CDs have enough dynamic range for me...

The storage/transmission medium can't have "too much dynamic range". But the program can... For example, if you listen to classical music in the car you may find yourself adjusting the volume up-and-down. The loud parts can be too loud or you can't hear the quiet parts over the road noise. Or, maybe you simply want to hear the quiet parts without full-orchestra loudness during the loud parts. Or sometimes people complain that they can't hear the dialog in movies without the effects being too loud.


Mostly true, but you can usually hear the signal mixed with noise, even though the signal is quieter than the noise. It depends on the nature of the signal and the nature of the noise.

The noise floor is mostly a problem with quiet sounds or with silence (between tracks, etc.). It's usually not so much the dynamic range or signal-to-noise ratio. It's (mostly) whether the noise is audible or not during silence or during very-quiet parts.


I disagree with that 1000% !!! I grew-up with vinyl and the clicks & pops always annoyed me. I could live with more-constant background noise but we don't have to, and I prefer the dead-silent background.

Records also occasionally have audible distortion and frequency response variations. Overall it's a technically-inferior format. (Some people prefer the sound of vinyl and that's perfectly OK with me.)

Of course it's possible to add pink noise or white noise to a digital file. You can do that with Audacity.


What?... You have a volume control so you can always keep it below the pain level.

I'm agree, CD is enough but perhpas I will vote to a standard of 24/48 to have some extra resolution for room correction and keep with an exact divisor of 24/96 most used in recordings.

Also can be better to add to flims at 24 fps, although I'm not quite sure of the later

Can Dolby or other spatial audio justify the extra resolution or maybe the opposite?

A standard will simplify DAC hardware

 

[Robin L]

And others, but yes Karajan was on the extreme of smoothie’s

The other extreme can be Osmo Vänskä recording of Beethoven symphonies that was suggested by a member up in the thread: the “Allegro con brio” that starts the Eroica (symphony n.3) seems to be recorded to awake an audiophile from his coma…

It worths to take a look on the crescendos, is impossible to an orchestra to produce that without serious wrist injuries

The orchestra is Minessota and the seal is Super Audio CD

Minnesota, the record label is BIS. The Beethoven Symphony cycle of Osmo Vanska/ Minnesota Orchestra is only available as SACDs.

 

[MRC01]

I'm agree, CD is enough but perhpas I will vote to a standard of 24/48 to have some extra resolution for room correction and keep with an exact divisor of 24/96 most used in recordings.

Also can be better to add to flims at 24 fps, although I'm not quite sure of the later

Can Dolby or other spatial audio justify the extra resolution or maybe the opposite?

A standard will simplify DAC hardware

I agree. And as discussed elsewhere here at ASR, 48 kHz sampling also makes it easier to implement a clean and correct DA filter.

 

[krabapple]

And others, but yes Karajan was on the extreme of smoothie’s

But that cames from close miking lots of instruments and creating 'smoothness' at the mixing desk. Though of course, house sound (or echo added during production) contributes too

 

[Miguelón]

But that cames from close miking lots of instruments and creating 'smoothness' at the mixing desk. Though of course, house sound (or echo added during production) contributes too

I suppose is both the conductor and the edition process, never listened Karajan in live concert but had the reputation of a very special sound.

Nothing against, he was one of the greats, and revitalized a lot the classical orchestra for the big public.

if you have some time to listen to Vikingur Olafsson (this time under Deutsche Gramophon) I'm curious to know more about his technics of recordings, I have the impression that places the mic inside the piano.

very special pianist, on the oppsite side of actual performances that exhibit a machine-gun like ultra fast version more on the athletic side.

I'm thinking about Yuja Wang and Lang Lang, more concretely, the public is more attracted by their speed than the expression

Arcady Volodos is even faster, but it belongs to other generation