To be honest, I was surprised the results I got from a first try and have been chasing it ever since. Yet to come close. I chalk a lot of it up to dumb luck, but I have done quite a bit of reading on here and other places. I matched the compression driver exit angle to the waveguide and I forget my reasoning for 66 degrees, but it seemed to match that specific woofer perfectly (eminence 15a) and also there is somethign that improves in the midrange when you use a larger CD diaphragm,. I think the 4"ers done cut it for mid range duty. But yeah, I think a got a couple things right on that one completely by accident. I would like to try it again one day. Also the room may have had something to do with it, I have since moved. I dont have pictures of the template, but another thing I suspect that made this waveguide special: (i have no reason to hide any details because I have no plans of any commercial enterprise in audio lol) is that where the horn mouth contacts the air i used a protractor to carry the circle around through the round over back to its backside if that makes sense. I think this has something to do with why the linearity is so good in the crossover region. Anyways, I'm still just a complete noob so I have no real explanations other than that. Oh and another thing I discovered. At least to me, that below around 700hz the ear is not sensitive to diffraction so drooping the waveguide over the woofer to get the c2c distance closer had no ill effects .
It is called LF efficiency is it??? According Bjorn (that guy that people know) there is no connection to exit size and LF performance of a compression driver, stating; "I haven't seen any concluding data that a 2" exit loads lower then a smaller exit if the horn are optimized for both. It may be true, but there are other factors that have much more influence on this. There are compression drivers with 1.4" exit that go very low (100 Hz and 350 Hz)."
I think maybe they crossed bearings and that the difference between loading and LF efficiency has been confused. Its not controversial that a larger radiation size equates to an increase of LF efficiency which in turn lowers excursion, Whether it be the exit size or diaphragm size.
I can definitely relate to that. A proper/complete round over alleviates FR fluctuations nearing cutoff, thats for sure.
Yeah me too, in comparison to the talents on this forum....these guys here are pretty awesome so, they obviously play a factor in our results. One of the reasons I hoped to spread the DIY spirit to the sound engineering community.
Maybe we aren't exactly "noob" .... educated noobs maybe. I definitely kept 700hz in mind though I cannot say I nailed it. I did aim to have the smaller dimension of my Elliptical horn hold directivity down to 700hz. My horn builder was somewhat cunning which is why you don't see me advertise for them in a good light. I wanted the mouths vertical to be over 19" before roll over. My horn is slightly over 19"s tall but, what roll over...I have talked about 3d printing the whole thing, horn and enclosure. If I can come to the understanding of how to do so, I will likely and at the time I will produce a slightly larger horn but to be honest everything sounds pretty good, as is. The horn is some 36" wide so no issues there. Bigger dimensions are very more forgiving thanks to the lack of sensitivity in the ear and probably more importantly, the forgiveness of larger frequency sizes as opposed to the geometry they are interacting with
Dear ambitious noob, Go big.
Correct, a CD driver's exit is its/horn upper limit, i.e. ~13543/pi/1" = ~4311 Hz, etc., in round numbers for comparison, but even this only holds true if it doesn't have an internal conical/whatever WG such as the Altecs, which for 1" exit is actually 0.75" IIRC = ~5,748 Hz, so normally want the horn's initial flare to either match or do as Altec and others did by adding a round to whatever polar pattern square or rectangular adapter was required.
Its usable lower limit is basically a function of its Fs, horn loading (acoustic pathlength, effective terminus area + any local boundaries) with the understanding that the lower one goes, so goes its HF roll off due to its increasing air mass, i.e. be it a simple box or compression horn alignment we're always trading efficiency for BW and vice versa. Not many 'free lunches' in audio speaker design.
Its usable lower limit is basically a function of its Fs, horn loading (acoustic pathlength, effective terminus area + any local boundaries) with the understanding that the lower one goes, so goes its HF roll off due to its increasing air mass, i.e. be it a simple box or compression horn alignment we're always trading efficiency for BW and vice versa. Not many 'free lunches' in audio speaker design.
I don't......just sit in the chair + occasionally friends dance in the room.
But if you're just sitting in a chair, your horn is probably very hard to beat in terms of 'point source performance'.
I've only listened to the Kii, which is a superb studio monitor. I also think it's quite expensive compared to similar sized Genelecs, even a pair of Neumann KH 420s cost about €6.000 less.
Last edited:
Yeah, you got pretty close to the optimum with the first try, which is admirable. It took me about two years of numerical simulations to get to the same place
(the shape, the size, the material...)
I was trying to think of what optimum meant, with out having to ask... I think the reference was to oblate spheroid. Then I thought about how it couldn't be optimized for my listening preference which is called nearfield in the marketing place but some here, have a technical bone to pick with that term, and rather use the words Direct Field. I am still learning the technicality of these words, but I did draw this conclusion;
If you want to design a monitor that is meant for close distance usage, 1m possibly being a status quo.... the CTC spacing has to be within or very close to 1/4WL
If you want to design a monitor that is meant for close distance usage, 1m possibly being a status quo.... the CTC spacing has to be within or very close to 1/4WL
More than 60 years ago, Zeiss Ikon launched a truly brilliant tweeter horn with the phase plug integrated in the throat and an 'alternative' roundover.
Last edited:
News to me, so apparently me too as way back when in the USA, < 4 ft was nearfield and somewhat later when I returned to HIFI/HT design it had become 1 m.
That is probably why I am confused, the definitions are all over. Depending on who you ask and where you look.
Near field - That part of a sound field usually within about two wavelengths from a noise source, where there is no simple relationship between sound level and distance. The area in a room which is in the immediate vicinity of the sound source
I think this is the reason why some do not like the term nearfield monitoring. Under 4ft seems still to fit when using the term to basically describe what...close field? Technically no one is getting nearfield in the HF.
Near field - That part of a sound field usually within about two wavelengths from a noise source, where there is no simple relationship between sound level and distance. The area in a room which is in the immediate vicinity of the sound source
I think this is the reason why some do not like the term nearfield monitoring. Under 4ft seems still to fit when using the term to basically describe what...close field? Technically no one is getting nearfield in the HF.
Yeah I fluked out on a couple things. The horn before you take into account the roundover is around the size of the woofers sd (13") and due to the angle of the horn I was able to droop it over the woofer so that the two 13" circles about touch eachother from the perspective of the listener. I think this is important for a waveguide of this type for the purposes of coherency, (is it lobing?) I think the voicecoils in this configuration acutally lined up pretty good too. The main flaw in the design was the beaming up above around 4-5khz which would sometimes bother me. Very fatiguing. Im not sure if this was an artifact of the compression driver, the horn shape or the material itself. Of course these types of thigns often dont show up in measurements! I did abit of reading on concrete and it has a resonance around 4-5khz as well.Yeah, you got pretty close to the optimum with the first try, which is admirable. It took me about two years of numerical simulations to get to the same place
The beaming should probably be largely attributed to the snout (throat adapter) of the DE750-TN.
In the original DE75, the phase plug ends near the exit:
Despite the 2" throat, beaming can be minimized in combination with an OS(-like) waveguide:
Last edited:
Beaming inherent to a 2" throat starts above ~7 kHz, the driver has little to no effect on directivity below that, so any beaming/narrowing around 4-5k must be the horn, typically a mouth diffraction. If you connect a truncated OS to a circular arc (as I gathered this is what you've done), this is exactly the kind of curvature discontinuity that causes this. This can be eliminated by making the whole profile a single smooth curve, as shown in my papers.
Actually, after thinking about it for a while, I'd say that at 1m the CTC is not really important (as long as the transducers are stacked vertically), and here's why: 1) You can always make it to sum flat at the listening spot and 2) because of the very high direct to reflected ratio the rest will be hardly an issue.
Last edited:
When the speaker dimensions are past 1m and one is sitting 1m away its not far field yet, so probably all kinds of quirks? I've got faint memory reading that far field rule of thumb is roughly 3x longest dimension of the speaker? For 1m listening distance this would mean speakers should be smaller than 33cm. Well, perhaps not something to apply here. Certainly something to check the facts on
I've got faint memory reading that far field rule of thumb is roughly 3x longest dimension of the speaker? For 1m listening distance this would mean speakers should be smaller than 33cm. Well, perhaps not something to apply here. Certainly something to check the facts on
Last edited:
I'll answer to myself:
https://www.prosoundweb.com/far-field-criteria-for-loudspeaker-measurement/
According to the article far field is defined when balloon surface response doesn't change anymore, at least in the measurement context.
I guess it doesn't matter if one listens at far field or near field of a speaker, as long as the response is right. Being very close to a big speaker means even small head movement can change response quite radically if there are multiple sound sources, including baffle edge, so its head in a vice situation.
Quick example: what if a horn or a baffle had poor termination so that sound diffracts at the mouth/edge, secondary sound source forms. If the horn was 1m wide, "edge" 0.5m from the center, and listening distance 1m the edge would be sqrt(50cm^2 + 100cm^2) =~112cm away from ear. If head moves say 2.5cm, one inch so that closer edge is now sqrt(47.5cm^2 + 100cm^2) ~111cm away and the other ~113cm away. This is ~1inch path length difference between the edges and wavelength ~10kHz. Doesn't seem much but the top octave goes woosh woosh within the 1inch head movement. Not sure what it is in reality because the top octave probably doesn't meet the edge but I guess this is what is referred with the near and far field, response at balloon surface not changing radically with small movement of the observer. To remedy the example just use edge / mouth that doesn't form secondary sound source.
So, perhaps extend the 1/4wl thinking from just c-c distance to include all sound sources including physical features of the speaker like any sharp edges on the front. Factor in wavelength they are exposed to.
What this means for practical situation in camplos case is that measuring and tuning the speaker with far field response won't match what the ear hears at near field, right? so one has to measure and tune the system with microphone where the ear is and keep the ear at where the microphone was and it should work.
https://www.prosoundweb.com/far-field-criteria-for-loudspeaker-measurement/
According to the article far field is defined when balloon surface response doesn't change anymore, at least in the measurement context.
I guess it doesn't matter if one listens at far field or near field of a speaker, as long as the response is right. Being very close to a big speaker means even small head movement can change response quite radically if there are multiple sound sources, including baffle edge, so its head in a vice situation.
Quick example: what if a horn or a baffle had poor termination so that sound diffracts at the mouth/edge, secondary sound source forms. If the horn was 1m wide, "edge" 0.5m from the center, and listening distance 1m the edge would be sqrt(50cm^2 + 100cm^2) =~112cm away from ear. If head moves say 2.5cm, one inch so that closer edge is now sqrt(47.5cm^2 + 100cm^2) ~111cm away and the other ~113cm away. This is ~1inch path length difference between the edges and wavelength ~10kHz. Doesn't seem much but the top octave goes woosh woosh within the 1inch head movement. Not sure what it is in reality because the top octave probably doesn't meet the edge but I guess this is what is referred with the near and far field, response at balloon surface not changing radically with small movement of the observer. To remedy the example just use edge / mouth that doesn't form secondary sound source.
So, perhaps extend the 1/4wl thinking from just c-c distance to include all sound sources including physical features of the speaker like any sharp edges on the front. Factor in wavelength they are exposed to.
What this means for practical situation in camplos case is that measuring and tuning the speaker with far field response won't match what the ear hears at near field, right? so one has to measure and tune the system with microphone where the ear is and keep the ear at where the microphone was and it should work.
Last edited: