Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Mickey Tee]

Hi,

Just wondering if anyone has any thoughts on Fletcher-Munson or Robinson-Dadson curves with respect to mixing/ mastering music.

Given the bump at approx 1.5khz in the curves, then ideally, for human hearing, you would want this area to be slightly lower in amplitude than say 950hz and 2.5khz? That is, taking into account the natural way our ears work, in order to make a sound more even and less "harsh" (as the octave 1khz - 2khz can be quite harsh and strident) you would deliberately dip this frequency range, if it is not dipped already?


Anyone?


Thanks for your help,

Mickey

[bruno putzeys]

My eyes are less sensitive on the red and blue ends of the spectrum, but does that mean I get more faithful colour rendition if I boost the red and blue channels on a video monitor? Or would it just give me an oddly pink outlook on life?
Your ears hear the real world through those same frequency curves. That's the reference you have. So any change you make to the signal  will still sound like a deviation from neutral.  The only thing that matters about f-m curves for mix/mastering is the *difference* between the curves at different loudness levels.

[Mickey Tee]

Bruno Putzeys wrote on Wed, 02 July 2008 13:26

My eyes are less sensitive on the red and blue ends of the spectrum, but does that mean I get more faithful colour rendition if I boost the red and blue channels on a video monitor? Or would it just give me an oddly pink outlook on life?

If I can split your reply in two:

Yes, yes, very good; I don't mean to suggest for a second that you should use a sonogram or an analyser to make or master music. I was just making a connection between two different scientific (ie. testable) phenomena/ principles, and looking to see what people (who know a lot more about this/ have more experience with this field than I do) think about it.

Bruno Putzeys wrote on Wed, 02 July 2008 13:26

Your ears hear the real world through those same frequency curves. That's the reference you have. So any change you make to the signal  will still sound like a deviation from neutral.  The only thing that matters about f-m curves for mix/mastering is the *difference* between the curves at different loudness levels?

Nicely put tbh, and all I can really respond is that this infers that you should listen to your mixes and other projects at a wide variety of volumes.

[bblackwood]

I started a thread about this a couple of years ago (you can view that thread here) and I'm more convinced every day that my gut feeling is correct...

[Mickey Tee]

Mickey Tee wrote on Wed, 02 July 2008 12:21

Hi, Just wondering if anyone has any thoughts on Fletcher-Munson or Robinson-Dadson curves with respect to mixing/ mastering music. Given the bump at approx 1.5khz in the curves, then ideally, for human hearing, you would want this area to be slightly lower in amplitude than say 950hz and 2.5khz? That is, taking into account the natural way our ears work, in order to make a sound more even and less "harsh" (as the octave 1khz - 2khz can be quite harsh and strident) you would deliberately dip this frequency range, if it is not dipped already? Anyone? Thanks for your help, Mickey

Just to correct myself, the bump is at 3.5k; the curves have to be read upside down. Someone on "whatever works" just pointed this out lol, otherwise bass response would be overwhelming even at 10db lolololol...

[Andy Simpson]

When we consider the equal loudness curves, it becomes clear that the level at which we monitor directly affects the frequency balance that we favour.

In other words, if you equalise an orchestra recording to sound 'right' at some arbitrary listening level, it cannot sound 'right' at any other listening level, even with the same speakers.

With the current fashion for compression in both mixing & mastering, this problem of specific listening level becomes magnified by the RMS gain involved (recordings are less scaleable), since this RMS gain is usually not uniform across the spectrum.

Does compression cause recordings to 'translate' less well?

In my experience it does, for the reason described above.

Andy

[Bob Olhsson]

It's colossally important to always check signal processing and balances at a variety of different listening levels in addition to having a repeatable mark on the volume control.

Modern studies suggest a better model of how it works is like a two band compressor where the band below 300 to 800, depending on the individual listener, is bypassed and the high band is amplifying the low levels and attenuating the high levels. You can hear this obvious effect by turning the volume up and down.

Also beware that a number of self-powered monitors so some pretty weird things when you do this.

[Andy Simpson]

Andy Simpson wrote on Sun, 06 July 2008 13:33

When we consider the equal loudness curves, it becomes clear that the level at which we monitor directly affects the frequency balance that we favour. In other words, if you equalise an orchestra recording to sound 'right' at some arbitrary listening level, it cannot sound 'right' at any other listening level, even with the same speakers. With the current fashion for compression in both mixing & mastering, this problem of specific listening level becomes magnified by the RMS gain involved (recordings are less scaleable), since this RMS gain is usually not uniform across the spectrum. Does compression cause recordings to 'translate' less well? In my experience it does, for the reason described above. Andy

I rarely quote & reply to myself but in this case I'll do so.

On the subject of equalising to counteract equal loudness affects, there is a further problem I forgot to mention.

If we are talking about an orchestral recording with PP & FF passages, we have a further issue of the equal loudness effect having different consequences at each level of performance.

In other words, if we equalise the PP passage to sound correct at some arbitrarily reduced gain listening level, we cannot expect the FF passage to also sound correct with the same equalisation.

The more processing you do, the more of a lottery the result will be.

Andy