G weighting is specifically for infrasound, as it says. That’s more useful for characterising industrial nuisance noise, rather than audio. But as you can see, the A curve is -50 dB at 30 Hz. A system with subs can be running at 100 dBA, but be much more annoying to neighbours compared to a system without subs doing the same level. Hence the use of multiple readings at events and the like.
The B curve actually matches our 100 phon hearing better than the A curve, but you won’t find it used anywhere. C perhaps overestimates perception, but is a good approximation for ‘annoyance’ in practice.
You can easily make your own weighting curve, if that’s what you want. You can even trace the contour in VituixCAD and export it as a text file, then import it to whatever software you want.
The appropriate curve is the one for the phon level you’re going to be listening at, which is where the problems begin for this simple idea of ‘EQ to match the contour’.
The phon rating is the equivalent of a pure tone at 1 kHz, so for a 100 phon curve, you’d start with a pure tone at 100 dB(Z) SPL at the listening position. So far, so good. If you want to do this the easy way, just take the measured trace, and automagically fit it to your ‘ELC target’ curve with an FIR filter. Taps be damned for now, have a go at it and tell us what you think.
Perhaps there’s a reason this isn’t often done. For a start we don’t listen to pure tones; music is far more dynamic than that. So is pink noise, for this same reason. M-Noise is even closer to real music, and goes from 12 dB crest factor at the low notes to almost 20 dB CF at the highest.
Perhaps this is becoming clear - if you tune a loudspeaker to have the same magnitude response as the 100 phon ELC, it’ll be wrong for peaks which either go to 120 dB, or wrong when you reduce the system level for 80 dB RMS, 100 dB peak.
It’s also telling to look at the response curve for the highest rated speakers on a site like AudioScienceReview. Many of them can be downloaded and opened in REW. Check the on-axis and listening window responses, as well as the predicted in-room responses. You can compare them to the ELC traces for the 1 kHz SPL level shown on the plot.
This is all for listening purposes too. Don’t you want a flat two-channel transfer function when you’re producing? I’m sure you said this currently theoretical box is aimed as studio monitoring, in the near field no less. So what goes in should come out, no boosts or dips, maximum ‘revealing’ mix, or have I got that wrong? I don’t think even the popular SoundID thing from SonarWorks tries to match an ELC, and aims for flat response with linear phase.
The B curve actually matches our 100 phon hearing better than the A curve, but you won’t find it used anywhere. C perhaps overestimates perception, but is a good approximation for ‘annoyance’ in practice.
You can easily make your own weighting curve, if that’s what you want. You can even trace the contour in VituixCAD and export it as a text file, then import it to whatever software you want.
The appropriate curve is the one for the phon level you’re going to be listening at, which is where the problems begin for this simple idea of ‘EQ to match the contour’.
The phon rating is the equivalent of a pure tone at 1 kHz, so for a 100 phon curve, you’d start with a pure tone at 100 dB(Z) SPL at the listening position. So far, so good. If you want to do this the easy way, just take the measured trace, and automagically fit it to your ‘ELC target’ curve with an FIR filter. Taps be damned for now, have a go at it and tell us what you think.
Perhaps there’s a reason this isn’t often done. For a start we don’t listen to pure tones; music is far more dynamic than that. So is pink noise, for this same reason. M-Noise is even closer to real music, and goes from 12 dB crest factor at the low notes to almost 20 dB CF at the highest.
Perhaps this is becoming clear - if you tune a loudspeaker to have the same magnitude response as the 100 phon ELC, it’ll be wrong for peaks which either go to 120 dB, or wrong when you reduce the system level for 80 dB RMS, 100 dB peak.
It’s also telling to look at the response curve for the highest rated speakers on a site like AudioScienceReview. Many of them can be downloaded and opened in REW. Check the on-axis and listening window responses, as well as the predicted in-room responses. You can compare them to the ELC traces for the 1 kHz SPL level shown on the plot.
This is all for listening purposes too. Don’t you want a flat two-channel transfer function when you’re producing? I’m sure you said this currently theoretical box is aimed as studio monitoring, in the near field no less. So what goes in should come out, no boosts or dips, maximum ‘revealing’ mix, or have I got that wrong? I don’t think even the popular SoundID thing from SonarWorks tries to match an ELC, and aims for flat response with linear phase.
Lots of interesting info, and thanks for sharing....
When we talk about this peaks vs rms the idea of a 115db at 1 meter comes up....This idea that 20-30db of headroom is needed for accurate playback.....If 85db is status quo....I've built this...I have this 115db of headroom (probably higher). When viewing the RTA through REW I don't anticipate any of that theoretical info to come true unfortunately and I wonder if either my RTA is subpar for reading true peaks, or if the theoretical has meet real life and real life is more forgiving.
I wonder if Peak being 20-30db over average is a misnomer or being recited incorrectly, or just completely misinterpreted by myself.
The Peak, as I see it, being the highest part of the Mix, normally resides on the Sub bass side...And when visiting the Bass heavy rap material, it normally is 20-30db over the midrange.....No less, the Pink Noise compliments this slope of the music, as we know it is a representative of what we hear is close to neutral. The 80db ISO curve and the Pink noise do have things in common. I think at lowest bass loudness perception and things falling of the wall go in different directions and confuse the topic.
Crest Factor is what I need and want to understand very well....I am sure it is the key to using visuals like the RTA effectively in the studio.
I would imagine that certain genres of music have similar crest factors per spectrum region....or at the very least, there are trends that are useful to know about.
Seems crest factor gets smaller, moving away from bass, and that makes sense to me, considering what Ive said so far....
Crest Factor is like a translation of headroom needed per frequency, it seems.
I really would like to actually see this 20-30db peak concept play out in real life
Please submit material demonstrating why anyone needs 20-30db headroom over 700-7000hz let alone any part of the spectrum, outside of bass.
I think when we say the peak can be 20-30db over average.....we a not considering that this peak is going to happen only in bass region and no where else
From there I had the thought that using pink noise or simular, to draw a Slope.....and use that Slope to figure headroom needs, if we are saying 20-30db, make the the pink noise slope the transfer function the headroom is based off....and even then....thats likely more than enough headroom for anything but bass....
According my reading our ears aren't much use below 100hz in judging Level. If thats the case, then I should be able to approach 100hz visually, with enough understanding, and get a mix in the ball park, every time, using visuals, and chasing a scientific neutral. and then ears for minor correction
I believe that the Haptic experience of bass is relevant to judgment of bass levels, and I am not certain this is a thing that microphones can adjust for easily....I think this experience is directly tied to direct energy levels, and the only software I've heard that might separate direct from reflective energy is software @gedlee developed but I have not yet reviewed. The concept that; even at low volume, that there is another information vector, a long the haptic receptors....how to accurate explain this phenomena that I can now verify is real, from my own personal experience with the newly assembled woofer section.
I think this leads us to Velocity vs Vector.
When judging peaks vs rms is REW good enough for this or do I need to run this through my DAW and visit a true peak limiter or something simular?
When we talk about this peaks vs rms the idea of a 115db at 1 meter comes up....This idea that 20-30db of headroom is needed for accurate playback.....If 85db is status quo....I've built this...I have this 115db of headroom (probably higher). When viewing the RTA through REW I don't anticipate any of that theoretical info to come true unfortunately and I wonder if either my RTA is subpar for reading true peaks, or if the theoretical has meet real life and real life is more forgiving.
I wonder if Peak being 20-30db over average is a misnomer or being recited incorrectly, or just completely misinterpreted by myself.
The Peak, as I see it, being the highest part of the Mix, normally resides on the Sub bass side...And when visiting the Bass heavy rap material, it normally is 20-30db over the midrange.....No less, the Pink Noise compliments this slope of the music, as we know it is a representative of what we hear is close to neutral. The 80db ISO curve and the Pink noise do have things in common. I think at lowest bass loudness perception and things falling of the wall go in different directions and confuse the topic.
Crest Factor is what I need and want to understand very well....I am sure it is the key to using visuals like the RTA effectively in the studio.
I would imagine that certain genres of music have similar crest factors per spectrum region....or at the very least, there are trends that are useful to know about.
Seems crest factor gets smaller, moving away from bass, and that makes sense to me, considering what Ive said so far....
Crest Factor is like a translation of headroom needed per frequency, it seems.
I really would like to actually see this 20-30db peak concept play out in real life
Please submit material demonstrating why anyone needs 20-30db headroom over 700-7000hz let alone any part of the spectrum, outside of bass.
I think when we say the peak can be 20-30db over average.....we a not considering that this peak is going to happen only in bass region and no where else
From there I had the thought that using pink noise or simular, to draw a Slope.....and use that Slope to figure headroom needs, if we are saying 20-30db, make the the pink noise slope the transfer function the headroom is based off....and even then....thats likely more than enough headroom for anything but bass....
I am applying these thoughts too.....measurements and behind the scenes....sound engineering, we are in agreeance that the monitoring should have a 0 transfer function or something of neutral nature.
According my reading our ears aren't much use below 100hz in judging Level. If thats the case, then I should be able to approach 100hz visually, with enough understanding, and get a mix in the ball park, every time, using visuals, and chasing a scientific neutral. and then ears for minor correction
I believe that the Haptic experience of bass is relevant to judgment of bass levels, and I am not certain this is a thing that microphones can adjust for easily....I think this experience is directly tied to direct energy levels, and the only software I've heard that might separate direct from reflective energy is software @gedlee developed but I have not yet reviewed. The concept that; even at low volume, that there is another information vector, a long the haptic receptors....how to accurate explain this phenomena that I can now verify is real, from my own personal experience with the newly assembled woofer section.
I think this leads us to Velocity vs Vector.
Last edited:
Why?? snares, rim shots, cymbals cowbells and god knows how many other percussive instrument's not counting a whisper to a scream. You can get a hell or a lot of dynamic contrast on an acoustic guitar. What about horns like a trumpet or a sax! A piano??
Rob
Last edited:
7 years of pro lessons by the age 14, life time of vocals, play piano/organ to please myself.
Favorite Genre; swing and bebop....and can play it at professional level, as well as, classical, which is easier for a percussionist, anyway. Elvin Jones is your Daddy...Brubeck and his band (and the like) are visitors, imitators...I consider my level of skill higher than his drummer who struggle to keep time in "Take 5"...consider he struggled playing in 5... thats not saying much is it
Love the song by Future; Wicked....played at loud volumes....
Abuse is an understatement I started at 7 years with a full kit and no ear plugs. I have learned later on....I also think I may have stronger ears due to early abuse...I only wonder this because my brain software/ear training still puts my hearing ability seemingly above normies, with my tinnitus and past and future abuse in the mix.
my listening is pretty eclectic, But Centers Around Jazz/Rap&Rnb/EDM/Gospel/Rock/Classical/Folk/Alternative/Country
Favorite Genre; swing and bebop....and can play it at professional level, as well as, classical, which is easier for a percussionist, anyway. Elvin Jones is your Daddy...Brubeck and his band (and the like) are visitors, imitators...I consider my level of skill higher than his drummer who struggle to keep time in "Take 5"...consider he struggled playing in 5... thats not saying much is it
Love the song by Future; Wicked....played at loud volumes....
Abuse is an understatement I started at 7 years with a full kit and no ear plugs. I have learned later on....I also think I may have stronger ears due to early abuse...I only wonder this because my brain software/ear training still puts my hearing ability seemingly above normies, with my tinnitus and past and future abuse in the mix.
my listening is pretty eclectic, But Centers Around Jazz/Rap&Rnb/EDM/Gospel/Rock/Classical/Folk/Alternative/Country
Last edited:
Well then, Thanks for walking with me through this thought train, as I try to debug my understanding....the first thing I notice is....I've never really mastered the concept of RMS...All I ever look at is PEAK.
So with that said I have placed my thinking in a strange place.
I am not seeing a 115db at 1 meter need for anything. I don't think anyone is arguing a need for this now that I see that no one is arguing for a need for this...
This meter is in 1/3 octave...Not sure what the units are, it says db, but who knows what that means...standard metering though.....I put it on the master channel...this is signal not audio from loudspeaker.
I hope this isn't an example of good sound engineering by the way...I can here the squeeze the snare is causing via a triggered compressor somewhere on the mix.....
Without getting into the settings of the RTA of above, I can see the headroom we are talking about over RMS.
The 1/3octave RTA suggest that 20-30db can be used against, a Pink Noise like slope, transfer function.....
So with that said I have placed my thinking in a strange place.
I am not seeing a 115db at 1 meter need for anything. I don't think anyone is arguing a need for this now that I see that no one is arguing for a need for this...
This meter is in 1/3 octave...Not sure what the units are, it says db, but who knows what that means...standard metering though.....I put it on the master channel...this is signal not audio from loudspeaker.
I hope this isn't an example of good sound engineering by the way...I can here the squeeze the snare is causing via a triggered compressor somewhere on the mix.....
Without getting into the settings of the RTA of above, I can see the headroom we are talking about over RMS.
The 1/3octave RTA suggest that 20-30db can be used against, a Pink Noise like slope, transfer function.....
Last edited:
You've got it backwards yet again ! Crest factor with music gets higher as frequencies increase.
https://m-noise.org/
I see that now.... I think I typed that out of error....the loss of RMS in teh higher spectrum is what I feel, supports the idea that higher frequencies are more harsh to our ears, Safety wise. In the study of alpha vs beta vs gamma, the high frequencies had more ability to do damage to human cells, while easier to reflect while the lower frequencies had the ability to penetrate but did less damage? You see some of this trend play out in sound where treble bouces off the wall while bass travels through the wall.....but as far as ear damage, it may be apple and oranges.
Is there a study of Crest factor per genre?
I feel like Peak signal has been my struggle with sound engineering and not so much rms. It is likely a matter of proper eq. AS I sought to find ways to make bass level judgment as easy as possible, I just learn that our ears suck at judging bass levels regardless of SQ.
I think there are tricks to increasing level perception of bass but I haven't made them up yet lol
Maybe adding temporary high levels of harmonic distortion to the bass signal will help....
Otherwise, it seems it can't be done well, without visual aid....and if one has to use a reference, to get it right thats, a sign of weakness.
Is there a study of Crest factor per genre?
I feel like Peak signal has been my struggle with sound engineering and not so much rms. It is likely a matter of proper eq. AS I sought to find ways to make bass level judgment as easy as possible, I just learn that our ears suck at judging bass levels regardless of SQ.
I think there are tricks to increasing level perception of bass but I haven't made them up yet lol
Maybe adding temporary high levels of harmonic distortion to the bass signal will help....
Otherwise, it seems it can't be done well, without visual aid....and if one has to use a reference, to get it right thats, a sign of weakness.
Last edited:
Camplo does the RTA have an instant integration time?
No so i doubt you could ever have any usefull information about signal dynamic using a frequency analysis tool... ( remember that the way you look at signal influence what you can see...)
About low end i think you spent too much time listening for 808 drop bass ( like i did) and have loose of focus that eccept for some rare instruments the bass signals are mainly transients.
How an RTA could help you to judge rms? Focus on Peak rather than rms... educated through digital and not analog gear.
No so i doubt you could ever have any usefull information about signal dynamic using a frequency analysis tool... ( remember that the way you look at signal influence what you can see...)
About low end i think you spent too much time listening for 808 drop bass ( like i did) and have loose of focus that eccept for some rare instruments the bass signals are mainly transients.
How an RTA could help you to judge rms? Focus on Peak rather than rms... educated through digital and not analog gear.
The wiki page is correct not a typo, just not shown on the graph
"The G-weighting function is used to determine annoyance due to infrasound. The G-weighting for the determination of weighted sound pressure levels of sound or noise, whose spectrum lies partly or wholly within the frequency range from 1 Hz to 20 Hz, has been standardized in ISO 7196, (1995) [2]. G-weighted sound pressure levels are denoted LpG and are measured or estimated in dB(G)".
https://docs.wind-watch.org/Roberts-2004-Ecoaccess-guidelines-assessment-low-frequency-noise.pdf
Why not just follow the good advice of Bob Katz on this topic?
https://www.digido.com/portfolio-item/level-practices-part-1/
https://www.digido.com/portfolio-item/level-practices-part-2/
https://www.digido.com/portfolio-item/level-practices-part-1/
https://www.digido.com/portfolio-item/level-practices-part-2/
Camplo, about abuse: sorry i translated an expression we have in french, so i'm not sure if it was taken as what i meant...
I wanted to say that as a drummer you know instruments with very wide dynamic range ( even more if you are into bebop or jazz): ghost notes to full blast on a snare will give you high dynamic contrast into mid/high freq range.
Think about same things with a kick and then you know why you need so much power for the low end despite the freq band doesn't have high rms content ( with acoustic music).
Of course for 'modern' productions you won't need this much dynamic but if you work with non treated recording you'll need it.
I don't meant to judge you either with analog/digital education: it's a fact, people who haven't used analog gear often doesn't understand how to use vu meter and rely on peak.
In my view this is wrong: peak meters have been implemented to protect the gear, not as an useful tool for mixing.
We have now less issue with digital artefact ( we know what to do to lessen them at least) so better focus on RMS rather than peak as this is closer to the way our brain integrate level.
About low bass and how to make them 'pop'... distortion is your friend as well as eq. Mainly eq...
I wanted to say that as a drummer you know instruments with very wide dynamic range ( even more if you are into bebop or jazz): ghost notes to full blast on a snare will give you high dynamic contrast into mid/high freq range.
Think about same things with a kick and then you know why you need so much power for the low end despite the freq band doesn't have high rms content ( with acoustic music).
Of course for 'modern' productions you won't need this much dynamic but if you work with non treated recording you'll need it.
I don't meant to judge you either with analog/digital education: it's a fact, people who haven't used analog gear often doesn't understand how to use vu meter and rely on peak.
In my view this is wrong: peak meters have been implemented to protect the gear, not as an useful tool for mixing.
We have now less issue with digital artefact ( we know what to do to lessen them at least) so better focus on RMS rather than peak as this is closer to the way our brain integrate level.
About low bass and how to make them 'pop'... distortion is your friend as well as eq. Mainly eq...
the analog version of this worked very well as long as you didn’t over do it
https://www.avid.com/plugins/aphex-aural-exciter-and-big-bottom-pro-bundle
MountainMan Bob,
It works wonder for individual tracks within a mix but it's a nono for mastering ( it lacks subtility for a whole mix).
Aphex 'aural enhancers' implement distortion... and eq.
We have now multiband treatments to do this and they are easier to manage than the 'big bottoms' or 'aural enhancers' which were very good for their time!
It works wonder for individual tracks within a mix but it's a nono for mastering ( it lacks subtility for a whole mix).
Aphex 'aural enhancers' implement distortion... and eq.
We have now multiband treatments to do this and they are easier to manage than the 'big bottoms' or 'aural enhancers' which were very good for their time!
Last edited: