Room gain necessary for a “perceived” flat response?

Y

yepimonfire

Audioholic Samurai
A majority of rooms introduce a mild shelving at the lowest frequencies in room, often equaling a 6dB/octave 3-5dB boost. I recently placed a pair of “full range” speakers in my room (-3dB @32hz), and noticed in measurements the low end was bumped up by about 5dB in measurements. After running Audyssey XT, the corrected response measured fairly flat in comparison, but upon listening and switching between Audyssey on vs off, I immediately noticed the “corrected” response sounded completely unnatural, low notes from a bass guitar sounded thin and weak, whereas the uncorrected response sounded full and natural. Simply crossing the speakers to a sub and boosting the level by a few dB didn’t solve the problem, it just resulted in boomy sounding bass.

I started researching this effect and found that several studies done by harman showed that removing the room gain from the low frequencies in various room correction software resulted in poor subjective performance, resulting in the change being rated as “thin” , “bright” and “harsh”. Based on the results of this, harmans room correction target curve incorporates a slight shelving of bass similar to the effect of boundary gain.


Digging deeper, it’s also interesting to note, that many headphones designed for accurate studio monitoring also incorporate a LF shelf similar to room gain. Three examples here


Considering this, should we really be attempting to completely flatten the bass response? It seems to me, that outside of some minor attenuation of unruly peaks, some amount room gain should be mostly left in tact.
 
William Lemmerhirt

William Lemmerhirt

Audioholic Overlord
I think what you’re discovering is what many refer to as a house curve. Minidsp is brilliant for that. To a lot of people, flat sounds exactly that, flat. IME most people seem to prefer a slow rise in bass response, usually starting around 200hz but that’s not a hard line. I think it’s one reason many guys love dynamicEq. Fwiw, I prefer a house curve for music, but I dislike it in movies as it tends to make every little car door, or footstep absurdly exaggerated. Btw, what were these “full range” speakers you tried out?
 
KEW

KEW

Audioholic Overlord
Just want to make sure you have factored in the SPL and loudness contours.
We know that more bass is needed at lower SPL levels to provide a perceived flat response.
If you are finding you like the FR with shelved bass at reference levels that would be surprising (to me).
However, if you like it at -20dB, that is pretty normal. Since most of us do not listen to reference levels in our homes, turning up the bass makes sense. That is why Audyssey Dynamic EQ is generally preferred since it attempts (and does fairly well) to compensate for this.
 
Y

yepimonfire

Audioholic Samurai
I think what you’re discovering is what many refer to as a house curve. Minidsp is brilliant for that. To a lot of people, flat sounds exactly that, flat. IME most people seem to prefer a slow rise in bass response, usually starting around 200hz but that’s not a hard line. I think it’s one reason many guys love dynamicEq. Fwiw, I prefer a house curve for music, but I dislike it in movies as it tends to make every little car door, or footstep absurdly exaggerated. Btw, what were these “full range” speakers you tried out?
Rp-160m. They roll off anechoic at 45hz, but room gain stretches that down to 32hz. I’m not talking about an extreme house curve that strongly accentuates bass, but the gently sloping response at the mlp introduced by the room. It’s a well known fact that equalizing a loudspeaker flat in a room doesn’t result in anything resembling flat, because what a microphone picks up at the mlp and what our ears and brain hear aren’t the same. The microphone combines all of the sound into the measurement graph, the direct and reflected sound, our brain is able to quickly adapt and “hear through” the contributions made by the room versus the direct sound from the speaker. Considering this, I wonder if strong attenuation of boundary gain above and beyond correcting high Q peaks leads to a perceptually thin response, if the speaker is producing a flat bass response, and the room is helping boost that, perhaps by reducing the bass produced by the speaker in order to make it measure flat is the reason why it’s now perceived thin.

The room gain doesn’t sound unnatural with movies or music, but the “tamed” response does.
 
Y

yepimonfire

Audioholic Samurai
Just want to make sure you have factored in the SPL and loudness contours.
We know that more bass is needed at lower SPL levels to provide a perceived flat response.
If you are finding you like the FR with shelved bass at reference levels that would be surprising (to me).
However, if you like it at -20dB, that is pretty normal. Since most of us do not listen to reference levels in our homes, turning up the bass makes sense. That is why Audyssey Dynamic EQ is generally preferred since it attempts (and does fairly well) to compensate for this.
I never listen at low levels lol.

The ear is perceptually close to flat at about 85dB and above, music is generally played back at 80dB-85dB as movies -10dB to -5dB, which often registers LcFmax peaks of 105dB or even 110dB during big explosions or gun fights. I’m actually not a fan of dynamic eq, unless I were regularly listening below 20dB, which I never do anyways. My system components were carefully chosen with the intention of reference level playback without distortion for a reason :)
 
Last edited:
KEW

KEW

Audioholic Overlord
Interesting!

Check me to make sure I have this straight:
Harman Audio's research is saying that the ideal (as in most preferred) in-room FR would have an 85dB shelf from around 30Hz to 60Hz which would taper down to 80dB at ~150Hz to ~500Hz and then gradually taper down to 75dB at 16000Hz.

I have some questions if the Harman write-up addresses them:
Did they discuss the low end roll-off? Their ideal curve rolls off from around 85dB at 30Hz, down to 77dB at 20Hz. Is this part of their researched FR or does it reflect equipment limitations?
Do you know what content was used for the study...was it music, or movies, etc.?
 
Alex2507

Alex2507

Audioholic Slumlord
Do you know what content was used for the study...was it music, or movies, etc.?
Fast Car by that black chick ... whatshernameagain? Anyway that's from Sean Olive. I crashed the Boston Audio Society talk he gave a few years ago when that 'Learning to Listen' software was put out. He said it was a simple song that people could more easily identify preference with. When asked how loud the playback was he just used the term 'a good volume'. I think it was the 82-85 db range.

I think they just preference tested speakers and concluded that that type of roll off is most preferred and then set out to make such speakers ... if I understand your question about equipment limitations. I think their marketing strategy is not too low and not too loud but plenty good and plenty cheap.
 
S

shadyJ

Speaker of the House
Staff member
A majority of rooms introduce a mild shelving at the lowest frequencies in room, often equaling a 6dB/octave 3-5dB boost. I recently placed a pair of “full range” speakers in my room (-3dB @32hz), and noticed in measurements the low end was bumped up by about 5dB in measurements. After running Audyssey XT, the corrected response measured fairly flat in comparison, but upon listening and switching between Audyssey on vs off, I immediately noticed the “corrected” response sounded completely unnatural, low notes from a bass guitar sounded thin and weak, whereas the uncorrected response sounded full and natural. Simply crossing the speakers to a sub and boosting the level by a few dB didn’t solve the problem, it just resulted in boomy sounding bass.

I started researching this effect and found that several studies done by harman showed that removing the room gain from the low frequencies in various room correction software resulted in poor subjective performance, resulting in the change being rated as “thin” , “bright” and “harsh”. Based on the results of this, harmans room correction target curve incorporates a slight shelving of bass similar to the effect of boundary gain.


Digging deeper, it’s also interesting to note, that many headphones designed for accurate studio monitoring also incorporate a LF shelf similar to room gain. Three examples here


Considering this, should we really be attempting to completely flatten the bass response? It seems to me, that outside of some minor attenuation of unruly peaks, some amount room gain should be mostly left in tact.
You are talking about steady-state curves. Human hearing actually prefers a flat response from the sound source (ie a sound system), but we can tell the response through the room's acoustics. Since rooms tend to boost the low end, that is what we expect to hear, and we can intuit it when something is lacking in bass, even if the response has been equalized to flat at the listening position. The only way a flat response would sound natural is in an outdoor environment or a very large room that doesn't boost bass frequencies.
 
KEW

KEW

Audioholic Overlord
Fast Car by that black chick ... whatshernameagain? Anyway that's from Sean Olive. I crashed the Boston Audio Society talk he gave a few years ago when that 'Learning to Listen' software was put out. He said it was a simple song that people could more easily identify preference with. When asked how loud the playback was he just used the term 'a good volume'. I think it was the 82-85 db range
Cool! Tracy Chapman!
Thanks!
 
KEW

KEW

Audioholic Overlord
I think they just preference tested speakers and concluded that that type of roll off is most preferred and then set out to make such speakers ... if I understand your question about equipment limitations. I think their marketing strategy is not too low and not too loud but plenty good and plenty cheap.
No pipe organ or dubstep there, so I might guess they decided the 30Hz roll-off as more of a practical matter when making an affordable speaker.
IOW, they were not evaluating peoples preferences for musical content in the low 20's!?
 
KEW

KEW

Audioholic Overlord
A shelf, as referred to in a FR graph, would be an interval where the response is essentially flat, but generally above or below the nominal SPL.
If you look at the faded gray line below, there is a +5dB shelf from around 30-70Hz.
 
newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top