It does not take much thought to realize that the higher the DI the slower the reverberation field will build. In the Griesinger paper the first data plot clearly shows that all reflections < 10 ms (very early) interfere with imaging. The fewer of these there are the better.
We clearly do not have the same goals.
We clearly do not have the same goals.
I don't believe that is true. For reverberant build up and decay, speaker directivity has no effect in a diffuse room. Room RT is the same for trombones (high d.i.) and piccolos.
Your latest notion is that a more directional system will buy a larger Initial Delay Gap. Clearly the first several room reflections are easy to predict and adding patches of treatment can buy as much delay as you want. As a practical matter I would venture that one can absorb a much greater proportion of energy at a wall bounce than you can reduce via directivity.
I really don't have a problem with the CD horn, 2-way system approach that you are promoting, it is just that you constantly make claims of unique superiority that just aren't justified. Define the Direct to Reflected ratio that you think is best, the spectrum of the reverberant field and direct field, and let me play with a little room treatment and a wide dispersion speaker. I guarantee that I can match your spec and better your IDG.
This is not an empty claim as I am currently sitting in front of a system that allows me to independently set direct sound level and curve, and reverberant field level and curve, plus reverberant delay. That plus some room treatments have let me kill the first reflections and increase the IDG. The sound can range from overly dry and near headphone quality to more sound wash than you would want on any music.
How about some in-room impulse responses to back up your claims?
David
David,
I'm a bit jealous of the tools you have access to that allow you to independently set direct sound level and curve, reverberant field level and curve, plus reverberant delay. Few have access to being able to change all those parameters fast enough to retain a valid acoustical memory (or measurement) the changes impart.
I agree with your points above, using some small acoustical "clouds" in my control room absorb the wall and ceiling bounce from the monitors, the small room size would not allow or warrant speakers large enough to have pattern control adequate to avoid those problems.
For my live sound work, where I don't have the luxury of room treatment, using CD horns allow the system to achieve a similar Direct to Reflected ratio as in my control room.
Art
They are fairly standard DSP tools with some of our own enhancements to allow independent parametric EQ but with one button switching between 2 different states.
I've attached some impulse responses to show the difference between forward firing and reverberant field elements, plus the effect of using reflectors to deal with specific wall reflections.
The first shows the energy from the reverb elements with and without some reflectors added to tame the earliest hard reflections. Orange is with the reflectors in place. The blue poking out is the reflections with the specific reflectors removed. Note that the direct energy (not shown here) is starting at the 0ms point.
The second curve is a reverberant decay (log level vs. time) and shows the relative energy of the direct field chain (brown) and the later reverberant field chain (orange). There is about 15dB difference between the two over time although the greater initial energy in the direct response means that they are both about the same steady state level.
Also note that the rate of decay (the room RT) is not effected whether driven by the wide dispersing rear elements or the fairly directional front element.
Fun stuff!
Attachments
But what (for me) truly matters is the total D/R ratio, or, conversely, the reverberation distance vs. directivity. If correctly interpret your picture, it suggest that, for roughly the same steady-state initial energy, the reverberant field generated by the (directional) direct chain is 15dB lower than the one generated by the wide rear-radiating element. This can only be due to directivity, as the more directive source is pumping less energy in the room relative to the direct level.
This is in line with the math Linkwitz is doing here wrt reverberation distance vs. directivity:
Room Acoustics
Due to the slop it's hard to see, but interpolation + translation does suggest that the RT would be slightly larger for the dispersing element.
Attachments
Indeed and relatively straightforward to reproduce with typical DIY kit. Not sure it will fit in the current project but perhaps the next.
Are you sure about that? Once an isotropic stage is reached, I agree source directivity is of no consequence. However, I believe with a high directivity speaker it will take longer before a more or less diffuse sound field is realized.
yes, this is what I usually refer to as the Hopkins/Stryker relationship (from a BB&N paper, I believe). If you know speaker Q or d.i., room constant (from absorption, alpha) and distance you can calculate direct level and reverberent field level. Plot it vs. distance and you have the draw-away curve and see critical distance. As I have pointed out numerous times, the reverberent field level, once direct level is fixed, can be whatever you want if you can adjust listener distance, room acoustics, or speaker Q (any factor will do).
i'm not seeing that as the lines look fairly parallel. For my experiments my real concern is the separation rather than room RT.
No, not 100% sure...
I measured RT once in my living room. Twice, actually, with the speakers in the normal orientation and a second time firing straight into our very heavy velvet plus lining drapes. RT measured the same both ways.
I know that if you measure RT they expect a non directional speaker some distance away. If yo are close to the source, you can see the direct sound and its instantaneous drop as it it turned off. At a lower level the reverberent field starts to drop.
But RT is defined as the rate of reverberent field decay and would ignore a drop from a direct field level. The reverberent energy decay rate would be the same (diffuse field, isotropic level, etc.). Yes, it seems like speaker directivity and aiming would make a difference to the specific build up and decay of the field, but with the thousands or millions of ray bounces in effect at one time I don't think it matters.
For Earl's case of the IDG from the first few ray bounces, RT isn't the issue. RT is the statistical diffuse case view.
David
One way to look at things is to think the decay, if its dense and lateral, aids perception of spaciousness, and the level of first reflections after ITD delay determines the liveliness and perceived size of the room. Without a sufficient ITD gap (>5-10ms) there will be loss of intelligibility and as Heyser found, 'time smear distortion' of images.
How it is achieved, with wide dispersion and absorption/redirection or narrow dispersion and less absorption doesn't seem to matter - the end result matters. Narrow dispersion implies more 'efficiency', so it is beneficial for those who are into 'dynamics' - but not always. Thats why JBLs new 120degree coverage pro monitors are interesting.
How it is achieved, with wide dispersion and absorption/redirection or narrow dispersion and less absorption doesn't seem to matter - the end result matters. Narrow dispersion implies more 'efficiency', so it is beneficial for those who are into 'dynamics' - but not always. Thats why JBLs new 120degree coverage pro monitors are interesting.
I do not think this is correct. Attached is the RT measured in the same room, with the same speaker (CD with high DI), just the speaker in a different location. There are already some clear differences just by this change in location. I will try to make another with a more wide directivity speaker, and see what I get...
Attachments
Careful, you don't want to be seen agreeing with me, but yes, I agree with you.
Agree with your point, however if we look at the three variables for a typical home listening environment:
1. listener distance: is rather constrained by practical reasons, most people listening from ~3m
2. loudspeaker directivity: a change in DI from omni to dipole yields a critical distance that 1.73 larger than monopole or, conversely, a D/R ratio at 3m of -7.7dB as compared to -12.4dB for monopole (as per Linkwitz, for a room with alpha = 18% T60 = 630 ms)
3. room acoustics: critical distance ~ 1 / sqrt(T60). Achieving the same increase in critical distance as above (or the same D/R ratio at 3m) would require a T60 of one third of the original room.
Wouldn't that be quite drastic to achieve, meaning a LOT of treatment ?
So, while your arguments stand, it does look like controlling the DI is a more practical choice for most people - a point where I agree with Dr. Geddes - no matter how dangerous to agree with him these days
.However, where I do disagree with him is how much DI is optimal ? Dr. Geddes says the more the better and aims at 90 degree coverage. To me (and others) that is too much and dipole pattern is sufficient. If we extrapolate Griesinger's findings (which were done in large halls), a D/R ratio of ~-7 dB as above for dipoles at 3m sufficient. An even higher D/R starts lacking spaciousness.
Thanks for thr RT curves. They are very interesting.
I'm sure you realize that the considerable differences below 250 Hz are the modal nature of your room and that speaker placement is making the difference. Reverb time is only properly applied to a diffuse space and the commonly defined Schroeder frequency lets you calculate over what frequency range RT is appropriate.
Above the Schroeder frequency you are seeing differences of a little more than 10%. this is fairly typical.
The notion being offered by Earl and Keyser is that we can deal with a few of the early wall bounces (by aiming a directional speaker or local absorbtive treatment) and effectively alter the RT of the room. I still do not believe this is true.
Imagine the extreme case of a series of parabolic reflectors that guide a beam of sound around the room, avoiding the listener for a number of wall bounces. After the last reflector the sound is allowed to hit its first diffusing wall and from that point a normaly diffuse sound field builds up. Surely that is proof of a reduced build up rate. Keyser then points out that an isotropic field is ultimately built up and it will decay normally when the sound source is turned off. Earl points out that his more directional speaker will achieve the same thing.
We are confusing a delay in build up with a changed rate of build up. My series of parabolic reflectors is no different than putting a half second of electrical delay in series with the speaker. For half a second after turning it off, the room sound continues and then begins to fall. Have we changed RT? No, RT is a measure of slope while level is falling and a delay in onset of this fall is ignored in the measurement. Many acouticians like to use the metric of the time to fall from -5 to -35 (times 2) as it gets clear of any direct sound drops (or in our case delays) in field fall off.
In truth we can not seperate the sound build up from the sound decay. If we watched in detail the build up of a sound field, we might see the initial arrival of the direct sound and then, say, the first 6 reflections. Energy and timing build up from those 7 and subsequent arrivals. When we turn off the source after a steady state is reached, the direct sound goes away instantly (after the time of flight), then the first reflection, then the second, the third, and so on.
That is, the sound field turns off exactly as it turned on. Even in our diffuse (final) field the lights go out in the same sequence as they went on. Now we normally look at this on a logarithmic plot in dB so it appears that the field builds quickly and decays slowly, but in straight energy units the build and decay are mirrors. If ultimately the sound field builds to a diffuse field then the decay still matches the build up.
This is all to say that we can come up with oddball cases that take us away from a diffuse case and warp measurements of RT. For example acousticians discuss the case where two rooms have the same average absorption. In one room the absorption is well distributed and traditnal RT calcs give a good answer. In the second room the absorption is not so well distributed and two reflective walls face each other. In this case sound continues to bounce between those walls and RT is longer than calculated.
The initial claim was that the sound field would build more slowly if a directional speaker was used. I would say that you might delay the build up slightly by missing a first reflection, but using a broad dispersion speaker and absorbing that particular bounce gives you the same thing. After the first bounce speaker directivity is innefective and the reverberent field builds (and necessarily decays) at the same rate.
This is also why no equation for RT contains any term for source Q.
David
Not if the room itself is fairly reverberant.
You see Dave this is where you are not listening as I never made such a claim.
I never confused them - I was always talking about the build up, not the decay as you have implied. You got them confused.
And quite honesty I don't see much practical difference in "delay in build up" and "rate of build up" - I'll take either, but both are probably effected by directivity.
Keyser was thinking that sound build up and sound decay could be separated, yet directivity would have no impact on sound decay. I was explaining that the two are inseparable.
I explained the difference between a delay of sound build up (creating by an a artificial dry path) and changing the rate of build up (and hence decay).
"Probably effected by directivity"? As I said, look for an equation for RT that includes speaker directivity. You won't find one. If it doesn't impact sound decay then it doesn't impact the build up. Energy build up and decay are mirrors. You can create some artificial dry sound paths (without diffusion) and delay the onset, but once the sound hits the normal walls and starts to spread, the rate of build up and decay is defined by the room and not the speaker.
The explanation was easy enough to follow if your interest is in understanding rather than obfuscating.
Regards,
David
Are you sure this is correct? I would assume that if you aim a narrow sound "beam" at a flat wall, it will reflect again as a narrow "beam". Unless the wall is a diffuser off course, but in my living room I have a lot of flat surfaces...
Fairly constrained but most listeners could easily move to 70% or 140% of their current distance if they desired. That would be comparable to a 3dB change in d.i.
I'll assume your math is correct. How I would look at it is that the useful range of d.i. for most speakers on the market is from 3 to 6dB for very wide dispersion (mid and high respectively) to a max of about +12 for the more directional horns. Your dipoles are nicely in the center of that range. If level is set by the direct field then the reverberant field would directly change by that range of d.i.
A simple way to look at it is that the reverberant field, I believe, goes up and down at 10 log the ratio of any added treatment (reverberant field drop equals 10log (absorption after/absorption before)). Treatment isn't free but room treatment is rewarding and adds the benefit that room modes will be smoother and the room will be quieter.
So, while your arguments stand, it does look like controlling the DI is a more practical choice for most people - a point where I agree with Dr. Geddes - no matter how dangerous to agree with him these days
Possibly. My point has always been that if we could decide what direct-to-reflected ratio is desirable, we can arrive at it by any number of speaker directivity, room deadness and listener distance combinations. If we thought that reverberant level must be reduced to way below what is typical today, then higher d.i. speakers might be essential. As Earl seems to be advocating using a more directional system in a less well damped room, it isn't clear how that differs from using a normal speaker in a normal room.
David
Last edited:
That is completely wrong. RT does not depend on directivity but the rise of the reverberation field will. While it is true that the build up and decay in terms of individual reflections are the same, the initial conditions are completely different and as such the results will be completely different.
If you over simplify the problem to just looking at D/R ratio then they may not be different, but that is an over simplification of the real issues.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.