That’s correct but we are talking about the shape of the curves, I.e. how much goes off at what frequency?
-
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!
Should you use Fletcher-Munson loudness compensation?
- Thread starter dshreter
- Start date
People here are proposing that we 'compensate' for it. We don;t.
And whether we 'need' correction for it. When listening at lower than 'reference' level, bass (and treble) will need to be boosted if you want to maintain a perception the correct frequency balance.
That's the fact. The rest is choices.
And whether we 'need' correction for it. When listening at lower than 'reference' level, bass (and treble) will need to be boosted if you want to maintain a perception the correct frequency balance.
That's the fact. The rest is choices.
DanielT
Major Contributor
Adjustable loudness, but does it follow automatically when you change the volume?I love the Yamaha variable loudness control. It's just well done.It can really transform a set of speakers into something else for awhile until the next setting comes down the pipe.
Aha, then maybe it's not so stupid (a matter of taste of course) to have speakers with a "disco bass" that is, an increase in the bass register, if you usually play at low volume? If you do not have loudness functionality on the amplifier. But on the other hand at high volume, it will then be strange. Then too much bass. Pros and Cons.In addition, there are so many flexible EQ options nowadays, that you can "turn the knobs" how you want. The setting options are incredible, that is.
Interesting!
Last edited:
Looking at my Yamaha A-S301 loudness control, as you increase the loudness (from flat) the output level decreases but the bass and treble is boosted (relatively)I made this a while ago.
Most recordings are mixed at average levels between 80 and 85dB (depending on the size of the mixing room and other factors)
I have 'normalized' 80 dB equal loudness contour and then you can see what 'compensation' you would need when you listen to different average levels.
Determining what average level you actually listen to is another matter.
I would say just apply EQ based on the right plot with your average SPL and the 80dB reference. When this sounds better to you and your speakers/headphones are correctly tuned this should work.
It is not by chance that my target curve is matching the 70dB line. I too listen at average levels (active listening) between 70 and 75dB average and well made recordings sound good to me on this target.
With a good imagination you can see the similarity
Weeb Labs
Addicted to Fun and Learning
I have loudness compensated volume control implemented on my primary ADAU1467 DSP unit. It's quite essential, in my opinion.
This is a very minimalistic version of it in SigmaStudio, for those interested.
This is a very minimalistic version of it in SigmaStudio, for those interested.
- Joined
- Jan 15, 2020
- Messages
- 6,904
- Likes
- 16,937
Which is unfortunately quite wrong though (like on most such loudness circuits) as the treble should stay almost the same as shown from @solderdude.
Makes you wonder why Yamaha chose such an implementation, given that it seems quite a deliberate characteristic. It may be time to start playing with a few different variations in DSP that more closely follow the equal loudness curves. This time of year is always a good time to tinker
solderdude
Grand Contributor
Its not about the actual shape of the curves but the relative differences between different Phon level curves.
When you overlay the curves you will see the treble hardly changes.
People prefer bass and treble boost, certainly with lower levels and is why loudness control often is made this way and few complain about it.
When you overlay the curves you will see the treble hardly changes.
People prefer bass and treble boost, certainly with lower levels and is why loudness control often is made this way and few complain about it.
DanielT
Major Contributor
Do you know if Contour has the same settings as Loudness? Maybe it's just another name for Loudness?
Example:
Example:
from the 360
www.manualslib.com
Tape Monitor Switch; Dolby Switch; Contour Switch; Selecting Monophonic Or Fm Stereo Broadcast - Harman Kardon HK630 Instruction Manual [Page 12]
Harman Kardon HK630 Manual Online: tape monitor switch, Dolby Switch, Contour Switch, Selecting Monophonic Or Fm Stereo Broadcast.
Denon Audyssey Dynamic EQ seems to work as intended. I.e. When listening at low volumes it boosts the bass so that the subjectively perceived balance between bass and other frequencies is closer to what would be experienced at higher listening levels. (Sometimes the boost is too much, but the reference level offset is helpful in adjusting this).
However, Dynamic EQ isn't necessarily always what you want. For example, if you are reducing volume to avoid disturbing others in the house, then the last thing you want to do is increase the ratio of bass. To prevent the bass from reaching other rooms you would have to decrease the volume even further, making mid frequency dialogue difficult to hear properly (without resorting to dynamic range compression/ dialogue enhancement/ midnight mode etc).
However, Dynamic EQ isn't necessarily always what you want. For example, if you are reducing volume to avoid disturbing others in the house, then the last thing you want to do is increase the ratio of bass. To prevent the bass from reaching other rooms you would have to decrease the volume even further, making mid frequency dialogue difficult to hear properly (without resorting to dynamic range compression/ dialogue enhancement/ midnight mode etc).
Why would one NOT resort to them? Some might actually work for what you intended (making dialogue easier to hear without pumping up bass). It depends on how they are implemented in your AVR.
"treble hardly changes" , indeed much less then the bass, but still quite audible:
if I look at the graph of the first post:
The curves at 60dB and 80dB have a delta of 20dB at 1KHz but only 16-17dB at 10Khz, so this would need 3-4dB correction at 10Khz.... The relative difference is a little bit more at the 40dB curve, +/-5dB at 10KHz
Audyssey DEQ affects both bass and treble --as it should , as per Fletcher-Munson and later curves
digitalfrost
Major Contributor
Whenever I try this I don't like it. I appreciate in an integrated amp, especially the Yamaha solution with the variable loudness control. But on my main system which is already DSP corrected, it just makes everything worse. Now, maybe by setting a target curve and tweaking the DSP, I already did my own loudness compensation to my preferred listening volume, which could explain why loudness on top of this doesn't work for me.
tvrgeek
Major Contributor
I would not take the precision for granted unless you are an average man from one street in England a long time ago.
I remember them on my old Knight tube amp. You can GOOGLE them. I think Self talked about them and probably something in the various OP-Amp books, Lancaster etc. They are quite effective in background music, but once to am "easy" listening level, not so much. Another useful feature killed by the "purist" garbage. Rumble filters, tone controls, choices of phono eq. How many modern Phono amps have provision for setting input impedance? Even my Hafler did!
I remember them on my old Knight tube amp. You can GOOGLE them. I think Self talked about them and probably something in the various OP-Amp books, Lancaster etc. They are quite effective in background music, but once to am "easy" listening level, not so much. Another useful feature killed by the "purist" garbage. Rumble filters, tone controls, choices of phono eq. How many modern Phono amps have provision for setting input impedance? Even my Hafler did!
I just played with this yesterday.
The Yamaha compensation on the 3080 seems to follow roughly the iso curves and adapts to the chosen volume.
I do like it but the effect is not overly strong.
For that reason, I wanted to compare what level (using input trim) leads to the best curve.
Nevertheless, I was disappointed to find that it is a global modification.
As others have already pointed out, a test tone of -20dB should be modified differently from a tone at -10dB.
This does not seem to be the case.
The modification only depends on the master volume.
Further, but I need to check this again because I measured at low volumes and did not use the preout, it seems that the very low frequencies (<50 Hz) are less boosted than one would expect - maybe to protect the subs?!
Edit:
BTW. Floyd Toole wrote rather positive about such compensations but also pointed out that most existing ones have flaws. He definitely also likes tone controls and even showed that the preferred listening curve that many employ might have been partially linked to the source material not having been played at reference level.
The Yamaha compensation on the 3080 seems to follow roughly the iso curves and adapts to the chosen volume.
I do like it but the effect is not overly strong.
For that reason, I wanted to compare what level (using input trim) leads to the best curve.
Nevertheless, I was disappointed to find that it is a global modification.
As others have already pointed out, a test tone of -20dB should be modified differently from a tone at -10dB.
This does not seem to be the case.
The modification only depends on the master volume.
Further, but I need to check this again because I measured at low volumes and did not use the preout, it seems that the very low frequencies (<50 Hz) are less boosted than one would expect - maybe to protect the subs?!
Edit:
BTW. Floyd Toole wrote rather positive about such compensations but also pointed out that most existing ones have flaws. He definitely also likes tone controls and even showed that the preferred listening curve that many employ might have been partially linked to the source material not having been played at reference level.
This is exactly how it should be. The correction should be based on (preferably calibrated) gain and not on the signal itself. A dynamic loudness correction based on the signal itself does not make much sense.
You are almost right, but there is some variation that should be handled. Let me give you an extreme example - others are less pronounced.
I want to play 4 tones and the sound is mastered for 80 dB.
I want to play
1k Hz at 80
Then a sound at 20Hz that is perceived at the same loudness, thus 120 dB by following the curves (I know it is out of range but it simplifies the example)
Then I want 1k at 20 dB
And again a sound at 20Hz matching this previous tone in loudness, thus we need 90dB.
Now someone takes the remote and lowers the level by 20 dB.
Then the first 1K sound is at 60dB.
The equivalent 20Hz loudness is now 110dB. So, we know that the loudness compensation should be a +10dB at 20 Hz... but what about the other two tones?
So the 1k tone is just at the hearing threshold, the equivalent 20Hz tone is 72dB. Nevertheless, it was 90dB, got lowered to 70dB by the remote and is now compensated to 80dB because of the +10. That is a mismatch of 8 dB and would be perceived roughly at the same level as a 10 dB 1k Hz tone, which feels roughly twice as loud as it should.
Again, the variation is most of the time small. Nevertheless, it is often in the order of a couple of decibels.
Similar threads
