Should you use Fletcher-Munson loudness compensation?

[dasdoing]

sometimes you just have to turn it on and off and listen and decide what sounds better

that is something "cientific audiophiles" should do much more than they do

 

[Dj7675]

I'm not 100% sure, but I think this is what Denon Dynamic EQ is doing. I find that I do like it. I'm not sure exactly what it's doing (can't see a REW graph of it or anything) but I think it's boosting the low frequencies as the volume decreases from "reference," which I think is around 85dB. To my ears music and movies sound better with it on, but I do the vast majority of my listening below reference. It pays to test out all of the features and extras on your AVR! I'm a big fan of audio science and graphs. I like REW graphs, Audyssey XT32 etc. but sometimes you just have to turn it on and off and listen and decide what sounds better!

A nice visual of DEQ if you haven’t seen it of what it does at different volume settings. You can also use reference level offset which helps if there is too much bass/treble boost at lower volumes.

 

[luft262]

A nice visual of DEQ if you haven’t seen it of what it does at different volume settings. You can also use reference level offset which helps if there is too much bass/treble boost at lower volumes.

View attachment 184480

I measured Dynamic EQ with REW to see what it's doing. Here is what I got. I still think it sounds better with it on, lol! Red is no dynamic eq, blue is dynamic eq on, but at 15dB reference level offset. Green is a reference level offset of 0dB.

On a side note. I just got the SB 3000's and was considering sending them back for the PB3000's, but based on the graph I probably don't need any more bass hu? What do you think?

 

[Sal1950]

On a side note. I just got the SB 3000's and was considering sending them back for the PB3000's, but based on the graph I probably don't need any more bass hu? What do you think?

Looks like you're good to me.

 

[dasdoing]

A nice visual of DEQ if you haven’t seen it of what it does at different volume settings. You can also use reference level offset which helps if there is too much bass/treble boost at lower volumes.

View attachment 184480

that's how it should look like.
btw, it's important to get a feel for the reference level. using a fixed level will not work universaly. using it to adjust the bass to liking at a given listening level is a good starting point

 

[GalZohar]

I measured Dynamic EQ with REW to see what it's doing. Here is what I got. I still think it sounds better with it on, lol! Red is no dynamic eq, blue is dynamic eq on, but at 15dB reference level offset. Green is a reference level offset of 0dB.

View attachment 184504

On a side note. I just got the SB 3000's and was considering sending them back for the PB3000's, but based on the graph I probably don't need any more bass hu? What do you think?

I think it mostly depends if the lack of power between 30-40Hz compared to 50Hz+ is due to the sub or due to room modes. If it's the former than the ported one might work better, but if it's the latter then nothing will help other than moving the sub or adding another at a different location.
Usually the ported ones will have the output advantage at 20Hz and up. In any case you should measure the response without any loudness/DEQ enabled to really see what your system is capable of. Preferably also measure without Audyssey at all.

 

[luft262]

I think it mostly depends if the lack of power between 30-40Hz compared to 50Hz+ is due to the sub or due to room modes. If it's the former than the ported one might work better, but if it's the latter then nothing will help other than moving the sub or adding another at a different location.
Usually the ported ones will have the output advantage at 20Hz and up. In any case you should measure the response without any loudness/DEQ enabled to really see what your system is capable of. Preferably also measure without Audyssey at all.

I've done those measurements (without DEQ or Audyssey) before. I could post them if you want. I posted the DEQ measurements here, because they seemed applicable to the Fletcher/Munson discussion. As I turn down the volume from reference with DEQ on the gap between the sub and main frequencies widens. The point I was trying to display is that the 10dB boost in bass over the mains might be perfect at reference, but as the volume decreases a wider gap in dB might be preferable to some and without DEQ there is no way to automatically achieve that as the volume decreases unless you want to manually boost the bass gain on your subs as you turn the overall volume down on the AVR. It was just an observation I was interested in showing/discussing further, but I see what you're saying.

 

[Lttlwing16]

FWIW, here is a good write up on loudness compensation that I hadn't seen posted yet:
https://mycourses.aalto.fi/pluginfile.php/666520/course/section/128564/Lauri Myöhänen_1929989_assignsubmission_file_LoudnessCompensation_myohanen_final .pdf

I like the idea of DEQ, but found it to heavy handed on music and distort the surround sound experience with it's surround level boosts. I, with another user, found allowing Audyssey full control of the entire FR narrowed the soundstage for music, so I limit below 700hz.

For loudness I was thinking of 1) conducting my own loudness compensation house curve with some test tones or 2) just copying one of the ISO226:2003 curves for around 70 phon, as I believe @solderdude mentioned is the average comfortable level for most folks. Way I see it I probably will rarely listen above that level, and below that I probably don't need the added bass anyway.

My question is this: what to do with 3khz-10khz? ELLC adds db's but Audyssey's reference curve rolls it off. Since I have to use reference curve with the curve editor perhaps I should add it back, or just limit the range with movies as well.

Here is the normalized ELLC graphs from above article:

 

[GalZohar]

I think the Audyssey rolloff is supposed to match with what they think your room is likely to have. Just like the midrange compensation is for what they expect your speaker to have. In reality you should probably adjust the curve to your speakers and room or just limit the correction frequency. Of course the limit will prevent from applying loudness correction to high frequencies. But in any case getting a DRC to correctly calibrate high frequencies seems like a challenge you should deal with first.

If you think DEQ is heavy handed on the bass you can always reduce it with theoffset (or source or speaker levels for a more granular tuning). However, according to the article at least the lowest frequencies are about the same: 2.6db boost per 5db master volume.

With music the main problem seems to be widely spread mastering levels and actual SPL in the source. Maybe it would be better to leave a constant curve and adjust the volume to the desired SPL in hopes the difference between that and the mix level wouldn't chang too much. I am really not sure if there is any good approach here.

For movies I am not sure how much I like/dislike the surround boost. My surrounds are quite behind me due to room limitations so some compensation might not be so bad? Compared to sound from the sides which shouldn't need compensation? As for the bass it seems fine in movies with DEQ (watching at -25 to -18 master volume), but only the mix engineer would know if that actually sounds right...

I am also not sure about best ways to choose target curves (loudness aside). Seems each have their own way (for example see Anthem Arc default curve parameters keeping room gain). Until I upgrade from Audyssey XT there is little I can try there anyway dueto limited bass correction capabilities, although a small and wide bass boost might be do-able.

 

[bkdc]

Just need to confirm the units on the Y axis. If you get a hearing test (audiogram), it is marked “hearing level (db)” instead of SPL (db) as the hearing test is normalized for a perfect human ear where 0dB at each frequency is perfect for a standard human ear. Thus a perfect ear has a ruler flat result. I think it would be useful to show measurements with a HEARING LEVEL scale instead of a sound pressure level scale as this shows loudness already adjusted for the human ear.

The recording engineer has should have already accounted for human hearing when he creates the recording master. If he is older, he sometimes does not trust his own ears and will depend on a visual representation of levels.

I am a fan of loudness adjustments when it comes to protecting me against sudden excessive noise. Otherwise, I am a fan of HIGH HIGH HIGH dynamic range. This is typically only seen in classical music as most recording engineers know that the general public listens to music on the go (gym, cars, cafes) where there is a significant level of background noise.

 

[GalZohar]

The main thing is that the engineer, who did already account for human hearing when creating the recording master, did it at a certain listening level. If you listen at a different level, the high and low frequencies will be out of balance for you. These curves are trying to compensate for that, so that stuff will still sound as good as possible at your selected listening level.

 

[Soundmixer]

It is clear we have insensitivities at low frequencies and a bit at high frequencies as well. My question is how do we translate this research from headphones to speakers in actual rooms? The Fletcher/Munson curves were established through headphones, not speakers in rooms.

 

[Soundmixer]

The main thing is that the engineer, who did already account for human hearing when creating the recording master, did it at a certain listening level

The Engineer accounted for HIS human hearing when creating a master. The end-user does not know what level the recording master used to monitor the process, it is not listed anywhere. There are no standards in music production or post-production, so everything is basically a crap shot.

 

[JeremyFife]

Hi, novice question for you: does this mean I should make REW measurements at a lower volume level than the general recommendation. I typically listen at low levels (thin walls / sensitive family)
Or, are the recommended REW settings a good reference point for me to start with?

 

[ZolaIII]

It is clear we have insensitivities at low frequencies and a bit at high frequencies as well. My question is how do we translate this research from headphones to speakers in actual rooms? The Fletcher/Munson curves were established through headphones, not speakers in rooms.

You are wrong it's frontal presentation how ever ISO 226 used compensated hedaphones for presentation.

http://www.lindos.co.uk/cgi-bin/FlexiData.cgi?SOURCE=Articles&VIEW=full&id=17

Read part 4.
How ever there ware concerns among which some are expressed in

https://www.researchgate.net/publication/258687253_Some_Reasons_to_Revise_the_International_Standard_ISO_2262003_Acoustics-Normal_Equal-Loudness-Level_Contours

And there is ISO 226 2007 revision which again did address some.

 

[ZolaIII]

Hi, novice question for you: does this mean I should make REW measurements at a lower volume level than the general recommendation. I typically listen at low levels (thin walls / sensitive family)
Or, are the recommended REW settings a good reference point for me to start with?

Do the measurements to the listening position and desired targeted loudness level and do correction from there.

 

[Soundmixer]

You are wrong it's frontal presentation how ever ISO 226 used compensated hedaphones for presentation.

Pump your brakes here. I cannot be wrong since I asked a question, not give an answer. My question stands(since you didn't effectively answer it) is based on this.

"Frontal presentation, from a single central loudspeaker, can be expected to show reduced sensitivity to high frequencies, which are partially masked by the head, and presentation using two loudspeakers, as for stereo will reveal more complicated differences related the HRTF (head related transfer function) which is also dependent on elevation of the sources and plays a major role in our ability to locate sounds. :"

In light of this complication, and that (once again) headphones were used for compensation, how can we apply FM to our loudspeakers in our own rooms? Most people are taking the FM measurements literally when there is an obvious disconnect here.

 

[sweetmusic]

A thought: when speaker reviewers comment that a particular pair of speakers sounds especially good for low-level listening, maybe what they really mean is that it has a bass boost.

Lots of bookshelf speakers have a bit of a bass bump. I always thought it was to compensate for the lack of low-end extension, and maybe some psychoacoustic testing found that people preferred a bass bump when there's less low end. Now I wonder if the bass bump is also there because people are more likely to listen to bookshelf speakers at lower volume levels, and there might be an equal loudness compensation effect there too. Wdyt?

 

[krabapple]

A thought: when speaker reviewers comment that a particular pair of speakers sounds especially good for low-level listening, maybe what they really mean is that it has a bass boost.

Lots of bookshelf speakers have a bit of a bass bump. I always thought it was to compensate for the lack of low-end extension, and maybe some psychoacoustic testing found that people preferred a bass bump when there's less low end. Now I wonder if the bass bump is also there because people are more likely to listen to bookshelf speakers at lower volume levels, and there might be an equal loudness compensation effect there too. Wdyt?

That sounds logical. Another thing to consider it that speaker reviewers can be far from rigorous about reducing variables. If one set of bookshelf speakers is placed closer to a nearby wall than another, it could get reported as better for 'low level listening' due simply to boundary reinforcement of bass.

 

[dasdoing]

A thought: when speaker reviewers comment that a particular pair of speakers sounds especially good for low-level listening, maybe what they really mean is that it has a bass boost.

Lots of bookshelf speakers have a bit of a bass bump. I always thought it was to compensate for the lack of low-end extension, and maybe some psychoacoustic testing found that people preferred a bass bump when there's less low end. Now I wonder if the bass bump is also there because people are more likely to listen to bookshelf speakers at lower volume levels, and there might be an equal loudness compensation effect there too. Wdyt?

makes total sense