Thoughts on room modes and EQ
Oct 20, 2008 13:38:43 GMT -7
Post by bumpyride on Oct 20, 2008 13:38:43 GMT -7
I mentioned some of this in a different thread but felt that it was important enough to start some discussion here. To start off with here's a wikipedia entry from the article on resonant room modes that touches on some of the issues I'd like to discuss:
"Attempts at evening out or equalising resonant modes ultimately fail to produce 'solid' bass notes in some rooms. This is because in an ideal listening environment, the note from a bass guitar for example, begins suddenly (ie, high risetime) and has a relatively slow tail off (ie, decay time). When a system has been more or less improperly is equalised to reduce the effect of a peak in room bass response by attenuating the level at the frequency of the peak, the rise time of bass notes at that frequency will be affected, and the room mode will still store energy at that frequency, also interfering with those bass notes decay times. At the extreme, these effects cannot be avoided and will cause audible problems (eg, loss of impact). Loudspeakers are often wrongly blamed for this effect. While it is true that speakers, especially poorly designed bass reflex systems, contribute to sluggish bass, a better speaker can't really improve a bad listening room. Recently, attempts have been made to cancel the reverberant field, using digital techniques that create inverse signals corresponding to the room reverberation, but they work best for a single listening position, as the room modes are not really affected."
from en.wikipedia.org/wiki/Resonant_room_modes.
There's a distinction to be made between a room modal resonance and the natural frequency of a subwoofer. Linear signal processing can effectively change the resonant frequency of a subwoofer and its Q; it cannot, however, eliminate room modes. There are things being examined that I'll discuss later but I'd like to possibly insult everyone for a moment to discuss why equalization should be used gently - particularly on the subject of transient response or 'impact.'
The biggest part of this distinction I'm talking about is that the resonant modes of a room have a finite rise time with regard to the music being played. When the subwoofer pumps the room at a frequency that corresponds to a standing wave, the energy reflects off the boundaries of the room and comes back to the sub in phase with the subsequent cycles at that frequency being played by the subwoofer. The result is that the intensity of the standing wave grows as it's being driven by the subwoofer over several cycles until the energy being put in matches the fraction of energy absorbed by each reflection - given by the Q of the resonance.
Every sound that comes through the subwoofer will excite these modes to some extent or another, the key here is that the room response seen in a sine sweep with a microphone only represents the effect on steady signals. The contrasting case would be an impulsive sound like a bass drum or just an electronic impulse signal. In this case, if the signal is short enough then the intensity measured by a microphone won't be affected by the room mode but it's decay will be. That's not to say that room modes don't play heavily into the room's impulse response, but merely that the initial impact of the impulse isn't affected in amplitude by them, which may be an important distinction for psychoacoustics.
Anyway, I'm taking a long time to say what everyone knows and that's that the best way to treat room modes is to eliminate them all together. If you can decrease the Q of room resonances then you've increased their rise time so they're less likely to be driven fully by unsteady musical signals and also the effect when they are excited is less dramatic. But I'd also like to make the point that the common practice of using EQ to flatten the subwoofer response can have significant effects on music played through it, especially when it's used on a high Q resonance. The goal should be a balance between maintaining initial attack and leveling longer notes (bass drums versus cellos for example).
There's another important point, though, and where I believe using EQ has the most benefit and that's for flattening 'room gain.' Room gain doesn't have any of the issues regarding transient response that a standing wave does and so flattening bumps caused by it is definitely the right way to go.
This is where I think the Audyssey EQ system, for example, and others like it do about everything that should be done in terms of equalizing speakers in a listening room. It uses impulse response tests and it averages several measurements within the room. That's a simple way to be sure that the system is addressing room gain kinds of issues and not attacking a standing wave that it can't do anything with.
There are newer processers like the ANTI-MODE 8033 that attempt to cancel the reverberant field by creating a modal model of the listening room. Going back a few paragraphs I mentioned that at a standing wave frequency, reflected energy comes back to the sub in phase with subsequent cycles and gets reinforced. These processors use their counter-model for the reverberation of the room to play an inverted signal that will cancel the reverberant field at the listening position. This approach can only tackle one listening position and expanding the area of cancellation can be difficult (it's like active noise cancellation where the noise is the room's reverberant response). The measurements for these systems look good, but I'd have to wonder about the transients because the system is pumping more energy into the room at the resonant frequencies. They cancel for the first reflections but I'd have to wonder about the long tail of a note as the initial sound and the counter-signal bounce around the room together. I didn't want to spend much space on this but I saw someone asking about it in another thread, and I figured, while I was on my soapbox...
While I'm writing my dissertation here I might as well ask if anyone has any experience with ASC's Isowall system. I mentioned that absorption is best when it comes to room modes and getting absorption at low frequencies is very difficult (stating the obvious). The only approach is to use a lot of surface area and having every wall in the listening room represent a damped, soft boundary could give good absorption. In theory, anyway, and that's why I'm asking. I haven't found any measurements to show that systems like the Isowall actually lower the Q of room modes, and I was curious if anyone here has experience.
Anyway, that's my penny's worth of thoughts. I'm mostly thinking out loud here as I contemplate my own IB installation, so I'd like to know what you think about any of it. Thanks for reading if you made it this far.
"Attempts at evening out or equalising resonant modes ultimately fail to produce 'solid' bass notes in some rooms. This is because in an ideal listening environment, the note from a bass guitar for example, begins suddenly (ie, high risetime) and has a relatively slow tail off (ie, decay time). When a system has been more or less improperly is equalised to reduce the effect of a peak in room bass response by attenuating the level at the frequency of the peak, the rise time of bass notes at that frequency will be affected, and the room mode will still store energy at that frequency, also interfering with those bass notes decay times. At the extreme, these effects cannot be avoided and will cause audible problems (eg, loss of impact). Loudspeakers are often wrongly blamed for this effect. While it is true that speakers, especially poorly designed bass reflex systems, contribute to sluggish bass, a better speaker can't really improve a bad listening room. Recently, attempts have been made to cancel the reverberant field, using digital techniques that create inverse signals corresponding to the room reverberation, but they work best for a single listening position, as the room modes are not really affected."
from en.wikipedia.org/wiki/Resonant_room_modes.
There's a distinction to be made between a room modal resonance and the natural frequency of a subwoofer. Linear signal processing can effectively change the resonant frequency of a subwoofer and its Q; it cannot, however, eliminate room modes. There are things being examined that I'll discuss later but I'd like to possibly insult everyone for a moment to discuss why equalization should be used gently - particularly on the subject of transient response or 'impact.'
The biggest part of this distinction I'm talking about is that the resonant modes of a room have a finite rise time with regard to the music being played. When the subwoofer pumps the room at a frequency that corresponds to a standing wave, the energy reflects off the boundaries of the room and comes back to the sub in phase with the subsequent cycles at that frequency being played by the subwoofer. The result is that the intensity of the standing wave grows as it's being driven by the subwoofer over several cycles until the energy being put in matches the fraction of energy absorbed by each reflection - given by the Q of the resonance.
Every sound that comes through the subwoofer will excite these modes to some extent or another, the key here is that the room response seen in a sine sweep with a microphone only represents the effect on steady signals. The contrasting case would be an impulsive sound like a bass drum or just an electronic impulse signal. In this case, if the signal is short enough then the intensity measured by a microphone won't be affected by the room mode but it's decay will be. That's not to say that room modes don't play heavily into the room's impulse response, but merely that the initial impact of the impulse isn't affected in amplitude by them, which may be an important distinction for psychoacoustics.
Anyway, I'm taking a long time to say what everyone knows and that's that the best way to treat room modes is to eliminate them all together. If you can decrease the Q of room resonances then you've increased their rise time so they're less likely to be driven fully by unsteady musical signals and also the effect when they are excited is less dramatic. But I'd also like to make the point that the common practice of using EQ to flatten the subwoofer response can have significant effects on music played through it, especially when it's used on a high Q resonance. The goal should be a balance between maintaining initial attack and leveling longer notes (bass drums versus cellos for example).
There's another important point, though, and where I believe using EQ has the most benefit and that's for flattening 'room gain.' Room gain doesn't have any of the issues regarding transient response that a standing wave does and so flattening bumps caused by it is definitely the right way to go.
This is where I think the Audyssey EQ system, for example, and others like it do about everything that should be done in terms of equalizing speakers in a listening room. It uses impulse response tests and it averages several measurements within the room. That's a simple way to be sure that the system is addressing room gain kinds of issues and not attacking a standing wave that it can't do anything with.
There are newer processers like the ANTI-MODE 8033 that attempt to cancel the reverberant field by creating a modal model of the listening room. Going back a few paragraphs I mentioned that at a standing wave frequency, reflected energy comes back to the sub in phase with subsequent cycles and gets reinforced. These processors use their counter-model for the reverberation of the room to play an inverted signal that will cancel the reverberant field at the listening position. This approach can only tackle one listening position and expanding the area of cancellation can be difficult (it's like active noise cancellation where the noise is the room's reverberant response). The measurements for these systems look good, but I'd have to wonder about the transients because the system is pumping more energy into the room at the resonant frequencies. They cancel for the first reflections but I'd have to wonder about the long tail of a note as the initial sound and the counter-signal bounce around the room together. I didn't want to spend much space on this but I saw someone asking about it in another thread, and I figured, while I was on my soapbox...
While I'm writing my dissertation here I might as well ask if anyone has any experience with ASC's Isowall system. I mentioned that absorption is best when it comes to room modes and getting absorption at low frequencies is very difficult (stating the obvious). The only approach is to use a lot of surface area and having every wall in the listening room represent a damped, soft boundary could give good absorption. In theory, anyway, and that's why I'm asking. I haven't found any measurements to show that systems like the Isowall actually lower the Q of room modes, and I was curious if anyone here has experience.
Anyway, that's my penny's worth of thoughts. I'm mostly thinking out loud here as I contemplate my own IB installation, so I'd like to know what you think about any of it. Thanks for reading if you made it this far.
