It seems that most of our listening environments have a base level of noise of about 40dB. Assume that listening becomes painful at 110dB. That is 70dB, 26 dB less than the dynamic range of a CD. If am listening at 110dB peaks, doesn’t the softest 26dB simply disappear into the noise. LPs sound better because the softest sounds are above the average noise floor, while the peaks remain below the pain level.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
How much dynamic range is too much
- Thread starter Rayadc
- Start date
Amirm has a great video about the loudness of sound : https://www.audiosciencereview.com/forum/index.php?threads/music-how-loud-is-loud-video.22434/
It is not as straightforward to state that one listen to a specific SPL level, such as 110 dB.
That kind of statement are meaningless without specifying the duration of the peak SPL and the distance between the sound source and the microphone and the type of acoustic where the measurement is taken. 110 dB measured over a 1 ms time span at 3 meters in an closed space is not the same thing as 110 dB over a 100 ms time span at 10 meters in the open air.
As far as the dynamic range of a particular record, it all boils down the the quality of the sound recording. A dynamic recording is not necessarily loud to listen to. That recording of some percussions out of a test CD would not fit on a vinyl, by any conceivable means, also some very brief peaks are clipped despite the low recording level :
means :
It is not as straightforward to state that one listen to a specific SPL level, such as 110 dB.
That kind of statement are meaningless without specifying the duration of the peak SPL and the distance between the sound source and the microphone and the type of acoustic where the measurement is taken. 110 dB measured over a 1 ms time span at 3 meters in an closed space is not the same thing as 110 dB over a 100 ms time span at 10 meters in the open air.
As far as the dynamic range of a particular record, it all boils down the the quality of the sound recording. A dynamic recording is not necessarily loud to listen to. That recording of some percussions out of a test CD would not fit on a vinyl, by any conceivable means, also some very brief peaks are clipped despite the low recording level :
DVDdoug
Major Contributor
- Joined
- May 27, 2021
- Messages
- 4,003
- Likes
- 5,490
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.
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 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.
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.
Last edited:
- Thread Starter
- #4
Of course not, 110 dB at 10 meters is less than 91.
Wayne A. Pflughaupt
Senior Member
The noise floor is inherent in the media and (as @DVDoug noted) is typically only evident when the signal fades to silence.
However, no one is making recordings with continual program content just a few dB above the media's noise floor – at least I hope they aren’t. If you have a recording like that, it’d probably be best to listen to it with isolated headphones.
Regards,
Wayne A. Pflughaupt
Whenever LPs sound better than digital mediums, this is not the reason. Whenever LPs sound better than a CD release, it's because the master is different. As an example, see The Mars Volta - Noctourniquet. The (at least initial) CD releases were mixed wildly different from the LP release, often digitally clipping. It did not sound good.
A good digital capture of the LP sounded a lot better than the actual digital releases, but only due to the fact that the digital releases were a disaster. And with a digital recording of the LP release you could digitally de-click the release, making the digital LP release sound better than the actual LP release.
That said, I am often amazed at how well LP sounds to this day. A pristine LP still sounds good to my ears, it's just that compared to digital it's more than a little bit behind the times. But my ears can't pick up the elevated noise floor and limited dynamic range when listening to random records via speakers.
- Joined
- Feb 23, 2016
- Messages
- 23,392
- Likes
- 43,684
Listening to some modern recordings apparently 6 db is too much. They don't let it get beyond that.
Seriously, one issue with the noise floor is we don't hear the noise floor across its whole bandwidth. Our ear breaks it into ranges. Read up on ERB, effective rectangular bandwidth. In simplified form out hearing does something like a 30 band FFT. So we can hear slightly below noise floors. The important part is you need to look at the noise spectrum. In that 3-5 khz range where our hearing is most sensitive a room with a 20 khz wide noise level of 40 db SPL very likely has a noise level in that 3-5 khz range of maybe 10 db SPL or a touch less. So we can hear noise down near that or close to 0 db SPL even.
The flip side is there is masking and where a recording is close to silent or fading out to silence we might hear these low levels, but while the music is playing very much it will mask noise anyway. If you've heard LPs, notice how the lead in groove has noise you can easily hear, but once the music starts you cannot hear it.
There is then the issue of noise in the recordings. Other than electronically created music, even in studios the noise level is enough it may be the limiting factor. As dithered CD can reach more than -100 db levels accurately and the facts of all this mean 16 bit if done well is enough not to be an issue. Beyond that is just icing on the cake.
Seriously, one issue with the noise floor is we don't hear the noise floor across its whole bandwidth. Our ear breaks it into ranges. Read up on ERB, effective rectangular bandwidth. In simplified form out hearing does something like a 30 band FFT. So we can hear slightly below noise floors. The important part is you need to look at the noise spectrum. In that 3-5 khz range where our hearing is most sensitive a room with a 20 khz wide noise level of 40 db SPL very likely has a noise level in that 3-5 khz range of maybe 10 db SPL or a touch less. So we can hear noise down near that or close to 0 db SPL even.
The flip side is there is masking and where a recording is close to silent or fading out to silence we might hear these low levels, but while the music is playing very much it will mask noise anyway. If you've heard LPs, notice how the lead in groove has noise you can easily hear, but once the music starts you cannot hear it.
There is then the issue of noise in the recordings. Other than electronically created music, even in studios the noise level is enough it may be the limiting factor. As dithered CD can reach more than -100 db levels accurately and the facts of all this mean 16 bit if done well is enough not to be an issue. Beyond that is just icing on the cake.
Maybe.
Here, if the mesurement is reasonable, I see 40dB SPL at 35hz, but much lower in the range where most folks ears are more sensitive.
Even so, the unweighted measure is 53.29db right now.
My room is not super quiet, with a computer (fan) and a floor fan gently blowing air (powered with a variac), and whatever else the evening suburban ambience provides.
"Threshold of hearing" says you are unlikely to hear that low rumble, and only might hear the higher frequency stuff.
Of course you do hear something in a quiet space. Even I do, deaf as I am.
--
Make your loudness peaks (at digital 0dBfs) 112db and you've "stretched" each of 16 bits to 7dB (instead of 6)
Last edited:
- Joined
- Aug 18, 2020
- Messages
- 2,880
- Likes
- 4,753
No,
1) because the noise floor in a room is usually dominated by very low frequency sounds which most of us don't notice or even hear (Fletcher–Munson curves), the noise floor is better where the majority of the musical detail is.
2) because we can comprehend intelligible sounds lower than the noise floor. It's not like a concrete floor!
What does means in practical terms? I'm a little bit lost in the thread and the subject is of my interest.
If an average listener is listening at 60 to 70 dB SPL, appart from peaks and quiet classical tracks that means having noise at - 55 to - 65 dB from the signal on the 1000 Hz range will be inaudible?
(seeing on the above FL curves that the audible threshold is +5 dB at 1 kHz)
Galliardist
Major Contributor
In most cases, music will mask the noise at those levels. And it's not so much quiet tracks, as the frequency range that is playing. With a turntable/LP, the noise is often also at lower frequencies. Solo instrument with no bass will show up noise because it isn't masked. Even then, the low level noise need not be obtrusive. You're still better off with digital and losing the other compromises inherent in LP, though, as long as the master is OK.
The range used for LP is still a good target range for domestic recordings, though. Not because of pain levels and such, but simply because it allows us to hear the music sufficiently well in a domestic environment without ruining our neighbours' lives. And classical CDs, largely, play within a dynamic range on a par or slightly greater than LP for that reason.
Too many digital masters of "commercial" music are not made for us - but for listening in difficult conditions, because so many people listen while travelling in different ways, where background noise levels are higher.
It's the dynamic range - and often the noise - of the recording that counts, rather than the medium, especially when using speakers in a room.
- Joined
- Aug 18, 2020
- Messages
- 2,880
- Likes
- 4,753
My first point is about what room background noise sounds like. We are most sensitive to any sound source (music or otherwise) from 500Hz to 4kHz. We are extremely insensitive to sub 50Hz and above 10kHz (Fletcher–Munson curves). Simply put, the noise in a room is normally a deep rumble which most of us hardly notice, but a calibrated microphone or SPL meter does detect and give a result for. So, let's say the microphone records room noise with a single result = -50dB. This is almost certainly rumble which, thanks to Fletcher–Munson feels like -70dB to us. If we look only at the same room noise 500Hz to 4kHz it's probably measuring -60dB or better. So a single value of room noise is useless unless it's heavily weighted in favour of the way we actually hear.
In my setup (not turntable but a streamer which I use at very low volumes) I theoretically have 121 dB of dynamic range, but should reduce around 50 dBFS to enjoy my active monitors at their original tonality (I can lower their sensitivity by - 12 dB but it doesn't sound good to my ears, and still have around 40 dBFS to reduce).
This let a 70 dB dynamic range system, which is precisely the maximum average sound pressure at I listen to.
So according to FL curves and the "hissing test" (totally inaudible at my listening distance, people coplaint to hiss in my monitors but should approach to 50 cm to perceive it), my setup is good for domestic uses or am I missing something?
Sorry but still confused: when industry recording standard is - 50 dB room noise that means in all frequencies weighted? Or is a peak?
Minimum detectable levels for white noise are roughly
+4 dB SPL via speakers and
about or slightly below 20 dB SPL via headphones.
(Our brain correlating sounds from both ears with the help of head movements is dropping the effective hearing threshold a lot. Meanwhile, uncorrelated noise in left and right ears just blends in with the general internal racket from blood circulation and whatnot. We have learned to ignore it as there are no natural external noise sources like that.)
About that much below average signal level, yes.
With an average signal of ..., this translates to a dynamic range requirement of...
-10 dBFS (modern pop/rock) --> 66-76 dB
-15 dBFS (1980s pop/rock) --> 71-81 dB
-26 dBFS (dynamic classical) --> 82-92 dB
You can probably tell why this makes the practice of brickwalling look stupid. Nothing wrong with compression if you insist, but there is no reason why peak levels have to be so high that there are overs, let alone on a regular basis. Even 16/44 (CD) audio provides plenty of dynamic range assuming proper dithering. The classical department may have to resort to shaped dither.
+1 just what i was about to say op sugest a single figure of 40dB not very useful without knowing the spectrum , your room is usually much much better in the sensitive midrange .
The noise figure are usually HVAC and traffic rumble
Which are average signal levels? I was sure I'm below average because having a quite sensitive monitors.
Post edited: I've just red that operating level is usually - 12 dBFS but refers to professional environments
Last edited:
True, if digital really sounds worse than it's LP counterpart they F*cked up, and probably on purpose at that, there's is no reason at all why digital should sound worse than analog, except with intent!
ZolaIII
Major Contributor
- Joined
- Jul 28, 2019
- Messages
- 4,958
- Likes
- 2,940
It's getting programe level above residential noise and towards as we hear it (psy). DR is from there including peeks in ± dB or LUFS. Too much is 24 dB as that's as four times as loud but again for some purposes used and needed as more realistic explosions and so on (THX cinema large ones). Even 9 in good not too complex music mix is fine with me. 12 is enough for most things and for very complex and layered in space again 16~17 is tops. Programe level is fine from 60 dB (even with residential noise at 40 dB and we perceive loudness never the less the same) up and no need going above 85/88 mono/stereo to listening area (white noise reference calibration point) for controled reproduction process (home, studio listening). There are applications when you do as you want to; trow on longer distance for targeted listening area and/or still have huge DR. But that's huge hals, stadiums and open spaces. Worst is with night clubs where target is usually in the middle of a dance floor and residential noise is huge so pushed to 100 dB programme. As they won't adopt that as only space to take your money and closer you get to a surce it's worse.
Similar threads
- Locked
