Hello Camplo
The impulse response will change as you move off axis as will the FR. The initial native IR response will still be intact as the reflections occur after it and you can clearly see this in any in room measurement.
"I would theorize that the approach that causes the least amount of reflections creates the best IR in a room, looking at dead on axis....actually my first description above is likely closer to the truth than your inaccurate oversimplification which could only be possibly true in an anechoic environment"
Why?? The speakers impulse response is independent of the room. The direct sound always gets to you first.
Seems to me like you are combining room and speaker and I see them as interdependent but separate entities.
You separately optimize the speakers impulse response on it's listening axis and treat the room accordingly for the desired response.
Rob
OK Fluid and Rob with great responses....
.
- I don't think this is cut and dry....I mean we put the microphone up as close as possible to a driver to get a measurement devoid of room effects....so a measurement at 6ft is going to include room in it.
- You are correct. I think it is the final perspective to be analyzed. Do you know how many times I've been scolded to not look at Group Delay or Decay in general, regarding bass, so intently because once its in a room, The room dominates among the frequencies its dimensions dictate it dominate. So in this case we are discussing not just areas where the room modes are but the "reflective" area? I don't know if there is a word like "Schroeder frequency", that is meant to point to the area above that term in particular, but, as I described....the polar characteristics of the loudspeaker system become even more critical, when put in a room. SPL (measurements in general) is a combination of direct and indirect energy. A loudspeaker system that maintains a more narrow directivity will generate higher Direct energy, at distance, vs one with a very wide polar. Plain to visualize and understand.... So my thought is that the more narrow the beam, the higher fidelity at listening position.
-Coverage angle is another word for "polar pattern" and we agree here. The wider the coverage angle, the more indirect energy is created. Regardless if the room is treated or not. Wider coverage puts more responsibility on the rooms performance, where as a more narrow coverage angle does not.
Good point, in this application, a Mastering Monitor, "desirability towards ones general liking" is not a Design attribute. Only Sonic Accuracy. It is at this time that someone will start to point to the idea of analyzing a signal devoid of reflections vs the common mans or desired audience's experience, to which the mastering engineer, would have in mind, but plenty of proof exist that various room characteristics have lead to Great mixes and masters, so just add this one to the list, with side effect of higher fidelity in the sweet spot, by the way
.
Last edited:
- This is an over exaggeration but I get your point. The thing is, headphones are considered to be higher fidelity than loudspeakers, in general. It is the loss of room, tactile feedback, among other things (head movement vs source, the way the sound hits the ear vs how the source is meant to be portrayed etc) that spoils the headphone experience.
A loudspeaker system with very strong "headphone" attributes is not a bad place to be, Quite contraire wouldn't you say. This system will produce tactile feedback and a smaller amount of reverb. The pressure waves hit the face from the front instead of directly on the side.... No need for a plug in room simulator or Subpac2, but enough direct energy to be "headphone like" that sounds like winning. Subtracting the room from the equation also facilitates emulation of other rooms using said system. IN other words, you Hear strongly the reverb that is in the signal, and not so much the reverb of the room, you are in.
Compare and Contrast.
Attachments
Last edited:
If you measure it in situ then yes it will be quite hard to have no room in the measurement which is why gating is used to approximate an anechoic response but the accuracy decreases as frequency goes down due to the windowing.
Decay and excess group delay are good in room measurements to look at. They can show if the reflected energy is skewed in frequency too much, too high or too low, and the excess group delay highlights timing problems between speakers.
Once the speaker has been designed and verified with anechoic or similar measurements, then you can look at it in room to see if anything needs to be changed due to the room. Without both sets of measurements it is hard to know where the problems come from.
I think this is a reasonable way to looking at it but you have to be careful because taking this idea to it's final conclusion results in listening in an anechoic chamber or non environment room. Here there is a big difference between what is good for production and reproduction.
Indeed most music mastering is done in rooms that are somewhat like a living room with speakers that are not too different from those that might be used in a good home setup. As a final check and balance to make sure it sounds good in one of it's intended venues.
The familiarity of equipment and room is what helps any one person to make mixes that translate well among different venues.
- So other then preference, why would one chose a system+environment that resulted in less fidelity. That points to an Anechoic environment, as you said, which we know is literally physically uncomfortable for a person....but once again, we are exaggerating....this theoretical system isn't going to be so literally like an anechoic chamber.
Finally that statement is mostly true....but not 100% true...To know a room/system is to know how it "lies".....a A more so neutral response (ie devoid of large peaks and nulls), I think is proven to be the actual goal of FR for this task. With slight character changes to suit preference, secondary....The more a system "lies" to you....the harder it is to get to "know it".
Last edited:
The waveguide polar is more effective for comparison and contrast if we know what it is lol
I just seen a nice looking polar and jumped on it.
S15 Econowave is what its called...Looks to be a 1khz XO with a 4th order slope.
I just seen a nice looking polar and jumped on it.
S15 Econowave is what its called...Looks to be a 1khz XO with a 4th order slope.
Attachments
Last edited:
Only if you take it so literally 😉
I do agree but not everyone feels the same. I like a strong direct field and reduced early reflections. I prefer the clarity and imaging specificity it brings. Others like wide directivity with an expanded but fuzzier soundstage. Have a look at Erin's recent review of the Kef R5 video where he talks about what he likes.
I'm not saying you exaggerate, with intent to do harm, rather, I understand that you do so to make a point. I am only trying to soften the blow I guess because the point needs to be considered thoroughly.
Preference is always a factor, though I am trying to make a point that higher fidelity at this cost will result in better or faster mixes and masters. Theoretically
I made this to start the compare and contrast of a nice performing waveguide and something generic to the horn philosophy though still successful.
These two examples likely have similar dimensions of mouth width. 350hz tractrix and the econowave waveguide pictured above
The larger horn, to me, represents a higher level of performance in a certain light, for horns. I think the Waveguide shown hear has a lot of qualities that are good, that is, higher level for waveguide. The best of the genre improves on FR smoothness? Coverage angle is still similar of the "best" waveguides?"....With a waveguide you cannot run the driver below the cutoff of directivity, with a horn you might be able to. So with a 15" wide waveguide, directivity falls much so after 900hz or so? With a 15" wide tractrix horn, same thing but this horn will play till about 700hz or so? With a 23" wide tractrix directivity starts to fall after about 589hz and it will play down to about 400hz. 2x fc rule....but I am learning that F is still up in the air on what a manufacturer will tell you. Anyway, all that must be considered will matching polars with the woofer, thinking of 15" in general. A waveguide will lean towards a mouth the size of the woofer its matching....a Horn can be mismatched mouth size vs woofer via playing into the area after directivity falls.
The scale on the 200hz is slightly disported I believe, being that its base level is -2.5db vs 0.0db. So It is in effect, slightly wider coverage area than it appears. The general trend should still show. Don't forget, 1khx XO on the Econowave...
Some thoughts; The 800-900hz common XO point for waveguides....is this done with a 1st or 2nd order slope?
Preference is always a factor, though I am trying to make a point that higher fidelity at this cost will result in better or faster mixes and masters. Theoretically
I made this to start the compare and contrast of a nice performing waveguide and something generic to the horn philosophy though still successful.
These two examples likely have similar dimensions of mouth width. 350hz tractrix and the econowave waveguide pictured above
The larger horn, to me, represents a higher level of performance in a certain light, for horns. I think the Waveguide shown hear has a lot of qualities that are good, that is, higher level for waveguide. The best of the genre improves on FR smoothness? Coverage angle is still similar of the "best" waveguides?"....With a waveguide you cannot run the driver below the cutoff of directivity, with a horn you might be able to. So with a 15" wide waveguide, directivity falls much so after 900hz or so? With a 15" wide tractrix horn, same thing but this horn will play till about 700hz or so? With a 23" wide tractrix directivity starts to fall after about 589hz and it will play down to about 400hz. 2x fc rule....but I am learning that F is still up in the air on what a manufacturer will tell you. Anyway, all that must be considered will matching polars with the woofer, thinking of 15" in general. A waveguide will lean towards a mouth the size of the woofer its matching....a Horn can be mismatched mouth size vs woofer via playing into the area after directivity falls.
The scale on the 200hz is slightly disported I believe, being that its base level is -2.5db vs 0.0db. So It is in effect, slightly wider coverage area than it appears. The general trend should still show. Don't forget, 1khx XO on the Econowave...
Some thoughts; The 800-900hz common XO point for waveguides....is this done with a 1st or 2nd order slope?
Attachments
Last edited:
Certainly there is never any intention from me to cause harm. I don't see it as an exaggeration but a description it sounds much more like large headphones to me much like close near field listening does. You of course do not have to agree.
Either can be run below the point where their directivity gives out. A horn is not usually good anywhere near cutoff due to the steeper rolloff and phase change. A waveguide used below its radiation peak will rely much more on the driver alone to set a lower limit. Every case is different and trying to generalise it like this can cause misunderstanding. Both waveguides and horns can be designed to be better than the examples you show. Docali has some great horns and mabat has some great waveguides. The difficulty is in making them.
This is the exact opposite of running them below where their directivity gives out.
Acoustic slope with a horn without any electronic crossover is about 2nd order so you probably end up with third order acoustic at least. Unless you crossover higher than the horn naturally drops out the response. But there you have narrower beam and get directivity mismatch to the woofer. I suppose you can tailor the slope with DSP but I think 1st order is out of question nevertheless.
You'd want the beam widen on the horn/waveguide and narrowing in the woofer to overlap for seamless transition in the xo. Of course you can have the xo somewhere else, where the directivity overlap of the devices is not so nice and the transition will not end up as smooth and there will be a slight step in the directivity index and early reflections. This is a concern if you listen far field where the room affects the sound.
From the few waveguides I have used in a prototype it seems that you want to crossover somewhere around frequency that is one wavelength of the woofer nominal diameter give or take some for smoothest transition depending on the size and pattern of the waveguide/horn. On a 8" woofer I don't get quite as smooth transition to RCF H100 waveguide around ~1200Hz as with a bit smaller STH100 around 2kHz. These are the xo where each of the devices give their smoothest transition to a 8" woofer looking at the DI and early reflections DI. 8" would be wavelength of ~1600Hz.
Based on this limited experience I suspect your smoothest xo with 15" woofer would be around 900Hz give or take depending on your horn/waveguide. I don't know how meaningful "a step" in the DI graphs would be. I suppose it reflects to the sound somewhat but I don't have enough experience to state how much of a difference there would be in each case or is it good or bad compromise for some other qualities.
The STH100 is a tractrix horn with rising DI, and the RCF H100 is almost constant directivity device. Based on this small sample I have reasoned a narrow pattern horn benefits crossover higher in frequency and a wider pattern waveguide benefits crossing over a bit lower in frequency (compared to the woofer size) to get smooth transition. Physically the wide pattern device has a bit larger mouth than the narrow pattern device. I'm not sure if you would get smooth transition with big narrow pattern horn unless having even bigger woofer (maybe array of smaller woofers) to crossover to. You should measure your devices in possession to see the differences in a simulator and then check out which device sounds better.
You'd want the beam widen on the horn/waveguide and narrowing in the woofer to overlap for seamless transition in the xo. Of course you can have the xo somewhere else, where the directivity overlap of the devices is not so nice and the transition will not end up as smooth and there will be a slight step in the directivity index and early reflections. This is a concern if you listen far field where the room affects the sound.
From the few waveguides I have used in a prototype it seems that you want to crossover somewhere around frequency that is one wavelength of the woofer nominal diameter give or take some for smoothest transition depending on the size and pattern of the waveguide/horn. On a 8" woofer I don't get quite as smooth transition to RCF H100 waveguide around ~1200Hz as with a bit smaller STH100 around 2kHz. These are the xo where each of the devices give their smoothest transition to a 8" woofer looking at the DI and early reflections DI. 8" would be wavelength of ~1600Hz.
Based on this limited experience I suspect your smoothest xo with 15" woofer would be around 900Hz give or take depending on your horn/waveguide. I don't know how meaningful "a step" in the DI graphs would be. I suppose it reflects to the sound somewhat but I don't have enough experience to state how much of a difference there would be in each case or is it good or bad compromise for some other qualities.
The STH100 is a tractrix horn with rising DI, and the RCF H100 is almost constant directivity device. Based on this small sample I have reasoned a narrow pattern horn benefits crossover higher in frequency and a wider pattern waveguide benefits crossing over a bit lower in frequency (compared to the woofer size) to get smooth transition. Physically the wide pattern device has a bit larger mouth than the narrow pattern device. I'm not sure if you would get smooth transition with big narrow pattern horn unless having even bigger woofer (maybe array of smaller woofers) to crossover to. You should measure your devices in possession to see the differences in a simulator and then check out which device sounds better.
Last edited:
The experience above is the desired experience...
Theres a 15" wave guide that will play well below 900hz?
Thats about true for horns too...Here I am speaking of 2" exit compression drivers.
"Better"
Well, this is conclusion based on the measurements of two waveguides I have. Eagerly waiting your measurements
I've been going by the simulated but I think I have a good understanding of how things can go. We have one person pointing out this attribute of giant headphones, and you are worried about off axis issues.
Yeah the off axis smoothness has less meaning the more you suppress the effect of the room, loudness of the early reflections compared to direct sound. So you kind of have the headphone effect, or the off-axis matters with gray area in between.
Have you posted the sims? I'd like to take a look at those for self education.
Have you posted the sims? I'd like to take a look at those for self education.
Sure, but you don't get as smooth transition of directivity between them unless the woofer and the waveguide/horn transition to wide polars within the same bandwidth or the transition happens sufficiently outside the crossover region for both devices to not affect the crossover region. You could use steep xo filters as well. This is where I referred earlier, I don't know how much a mismatch would matter.
I see the rationale a deep loading horn would allow this, crossover below pattern control. You might find out it doesn't work out too good or was the docali horn designed for this purpose? What is the relation of losing pattern control to the loading in terms of frequency? If the horn can be tailored to get loading a lot lower than losing pattern control without going into trouble I see this could work. I'd use a shallow wide pattern waveguide with beefy driver (the axi) and crossover frequency that doesn't require loading. This limits the SPL a lot though which might not matter at all in home use.
I reason the benefit here (crossing over omni) is that you don't have to use as big of a woofer (or array of woofers) as the waveguide is. Which allows a crossover lower than a single 15" or 18" or even 21" woofer would normally enable by matching the patterns. Otherwise I see no point doing it.
A thought experiment, I'm not sure if you have posted similar at some point?
Lets imagine we want to control pattern of a system somewhere to 700Hz or so to reduce effect of the room, and omni below that. Then it would be perhaps 24db/oct filters at 350Hz to not have too much effect on the DI. If there was loading required for the compression driver in the 350Hz you need a horn that loads there but loses the pattern around 700Hz. Very deep and relatively small mouth device, a vuvuzela? I think this has been the concern behind most criticism.
You could control the pattern to 350Hz to further reduce effect of the room in which case you'd need to crossover around 175Hz with steep filters not to have a step in DI. But again, not sure if DI mismatch mattered this low frequency at all.
rossover at 350Hz would not get you to point source territory so I see no point why cross this low especially when the pattern control would be only to ~700Hz and you could do it there why not. If crossover was < 200Hz, pattern control to ~400Hz you'd approach point source but I'm not sure if the axis driver would be happy there. I just can't think any benefits crossing below pattern control of a horn/waveguide from the system perspective.
Of course thought experiments are not the truth especially done with not too much experience on the subject. Measurements would show how it works out and listening in situ how it sounds.
I see the rationale a deep loading horn would allow this, crossover below pattern control. You might find out it doesn't work out too good or was the docali horn designed for this purpose? What is the relation of losing pattern control to the loading in terms of frequency? If the horn can be tailored to get loading a lot lower than losing pattern control without going into trouble I see this could work. I'd use a shallow wide pattern waveguide with beefy driver (the axi) and crossover frequency that doesn't require loading. This limits the SPL a lot though which might not matter at all in home use.
I reason the benefit here (crossing over omni) is that you don't have to use as big of a woofer (or array of woofers) as the waveguide is. Which allows a crossover lower than a single 15" or 18" or even 21" woofer would normally enable by matching the patterns. Otherwise I see no point doing it.
A thought experiment, I'm not sure if you have posted similar at some point?
Lets imagine we want to control pattern of a system somewhere to 700Hz or so to reduce effect of the room, and omni below that. Then it would be perhaps 24db/oct filters at 350Hz to not have too much effect on the DI. If there was loading required for the compression driver in the 350Hz you need a horn that loads there but loses the pattern around 700Hz. Very deep and relatively small mouth device, a vuvuzela?
I think this has been the concern behind most criticism. You could control the pattern to 350Hz to further reduce effect of the room in which case you'd need to crossover around 175Hz with steep filters not to have a step in DI. But again, not sure if DI mismatch mattered this low frequency at all.
rossover at 350Hz would not get you to point source territory so I see no point why cross this low especially when the pattern control would be only to ~700Hz and you could do it there why not. If crossover was < 200Hz, pattern control to ~400Hz you'd approach point source but I'm not sure if the axis driver would be happy there. I just can't think any benefits crossing below pattern control of a horn/waveguide from the system perspective.
Of course thought experiments are not the truth especially done with not too much experience on the subject. Measurements would show how it works out and listening in situ how it sounds.
Last edited:
Phase plugs that obscure a section of the mid or low-frequency range driver are a common way to aid in a wider degree of freedom when choosing acoustic crossover point. You’ll see that used a lot in professional audio, where cabinets are more likely to be arrayed such that anomalies in the polar patterns can create an unexpected response when multiple boxes are used.
If you’re clever about it, the edges of the high-frequency horn itself can even serve that function to an extent.
Of course, the argument is that you ideally want a consistent polar response to the lowest frequency possible in a full-range speaker. That way you can avoid an uneven response once the cabinet is placed into a room, or at the very least make the behaviour more predictable and consistent. A speaker with a small horn or waveguide that offers it's stated coverage angles to just under 2 kHz will sound very different to one that maintains the spec sheet values to below 500 Hz in a room with any normal amount of reverberation.
In that pro world, a Norwegian company called NNNN had a short-lived product called the Graft 40 which used a Celestion Axi2050 alongside some clever patented lossy chamber approach for the low section. I believe it got postponed or sent back to the drawing board due to COVID:
Sizing information is still available here:
https://nnnn.no/wp-content/uploads/2021/02/NNNN-Press-Kit-1.4..pdf
That might lend some numbers and visuals to the 'though experiment' @tmuikku
On that note, the 'Schroeder-like' specification you're looking for on the previous page @camplo is possibly the D/R (Direct to Reverberant) Ratio or the Critical Distance, which is where you’re listening to as much of the room as you the speaker.
These are room acoustic parameters that are typically measured for a location using an omni-directional source such as a dodecahedron speaker, placed at several locations and averaged as per ISO 3382-2 and 3382-3. Often the data is compared to simulations using geometric acoustics software, and then again with the desired loudspeaker system or acoustic treatments, either in the model or reality.
For example, a speaker with narrower coverage angles to a lower frequency will have a larger rating for Critical Distance, as the reverberant field will be excited less for a given distance from the source. That may or may not be desirable, depending on the use case. Theatres don't tend to like systems that lack 'envelopment' as an example, but then we're drifting far from audio engineering to building acoustics and theatre design (or lack of).
That way you can separate the influence of the room and the additional equipment to determine a performance improvement or degradation metric. It's not dissimilar to the concept of comparing 'anechoic' data for a loudspeaker to the in-room data, as explained by others in the recent posts here.
Oh, and as someone who's worked a fair bit in anechoic rooms of various types - it isn't as uncomfortable as you'd think. Most HiFi or studio speakers sound pretty strange in there though, since they're designed around their intended final usage location. You might find things feeling a little 'thin', and the aforementioned 'giant headphones' feeling is prominent for anything I've listened to in there. It doesn't mean the data obtained in one is any less useful in practice, as it can be convolved with room impulse responses - generated, calculated or measured.
If you’re clever about it, the edges of the high-frequency horn itself can even serve that function to an extent.
Of course, the argument is that you ideally want a consistent polar response to the lowest frequency possible in a full-range speaker. That way you can avoid an uneven response once the cabinet is placed into a room, or at the very least make the behaviour more predictable and consistent. A speaker with a small horn or waveguide that offers it's stated coverage angles to just under 2 kHz will sound very different to one that maintains the spec sheet values to below 500 Hz in a room with any normal amount of reverberation.
In that pro world, a Norwegian company called NNNN had a short-lived product called the Graft 40 which used a Celestion Axi2050 alongside some clever patented lossy chamber approach for the low section. I believe it got postponed or sent back to the drawing board due to COVID:
Sizing information is still available here:
https://nnnn.no/wp-content/uploads/2021/02/NNNN-Press-Kit-1.4..pdf
That might lend some numbers and visuals to the 'though experiment' @tmuikku
On that note, the 'Schroeder-like' specification you're looking for on the previous page @camplo is possibly the D/R (Direct to Reverberant) Ratio or the Critical Distance, which is where you’re listening to as much of the room as you the speaker.
These are room acoustic parameters that are typically measured for a location using an omni-directional source such as a dodecahedron speaker, placed at several locations and averaged as per ISO 3382-2 and 3382-3. Often the data is compared to simulations using geometric acoustics software, and then again with the desired loudspeaker system or acoustic treatments, either in the model or reality.
For example, a speaker with narrower coverage angles to a lower frequency will have a larger rating for Critical Distance, as the reverberant field will be excited less for a given distance from the source. That may or may not be desirable, depending on the use case. Theatres don't tend to like systems that lack 'envelopment' as an example, but then we're drifting far from audio engineering to building acoustics and theatre design (or lack of).
That way you can separate the influence of the room and the additional equipment to determine a performance improvement or degradation metric. It's not dissimilar to the concept of comparing 'anechoic' data for a loudspeaker to the in-room data, as explained by others in the recent posts here.
Oh, and as someone who's worked a fair bit in anechoic rooms of various types - it isn't as uncomfortable as you'd think. Most HiFi or studio speakers sound pretty strange in there though, since they're designed around their intended final usage location. You might find things feeling a little 'thin', and the aforementioned 'giant headphones' feeling is prominent for anything I've listened to in there. It doesn't mean the data obtained in one is any less useful in practice, as it can be convolved with room impulse responses - generated, calculated or measured.
Last edited:
You've lost me as to what you mean
If the driver can play well enough without much assistance from the guide and very high SPL isn't a requirement I don't see why not. Most 2" cd's would be ok down to 600 maybe 500 Hz. What goes wrong then is that the high end suffers from the larger throat. With the axi it might go lower. Crossing higher to a woofer extends the higher directivity of the waveguide a little lower where running the guide all the way down loses that.
Not really because a good waveguide gives out more gently and a horn designed to load gives out all of a sudden in comparison. The Sims on docalis website would show the relationship between loading and directivity control.