Someone might enjoy my little A/B comparison between this golden oldie from Yamaha and a modern take in the form of a variant of JMLC Iwata 300 with a slightly larger vertical mouth than the usual JMLC Iwata 300 from Autotech.
Round throat to square sectoral horn from Yamaha from early 1980’s, similar to the Yamaha H-1230, but with 1.5” throat, that required a short 1.5”-1.4” adapter to the JA6681B compression driver. With the DIY mdf adapter, the horn has similar dimensions to the Yamaha H1230 which was designed for the JA6681 :
Impression: goes deeper soundwise, less vertical dispersion obviously. Rollercoaster response below 1000Hz, a bit rough on the ears in comparison, but sweet enough.
Round all around JMLC Iwata 300 with JA6681B:
Impression: Very sweet sound, less deep sounds, but also much less wide dimensions (58cm including lip). More vertical dispersion obviously, that helps to spread the sound more evenly in the room.
Overall: The modern Iwata 300 would have been the clear winner with its round and smooth inside and nice extra vertical dimensions, but only if it was bigger and went deeper. Thanks to the 70cm wide Yamaha horn, I can taste the sweet nectar of pushing the JA6681B down to 300-400Hz. But maybe deeper is just producing distorted garbage anyway. 600Hz is a worthy goal for a problem free presentation.
Additionally, and more importantly in reference to the latest posts - round is the way to go. No more square vintage horns will enter this house.
Powered by V-Fet. Thanks Papa.
Round throat to square sectoral horn from Yamaha from early 1980’s, similar to the Yamaha H-1230, but with 1.5” throat, that required a short 1.5”-1.4” adapter to the JA6681B compression driver. With the DIY mdf adapter, the horn has similar dimensions to the Yamaha H1230 which was designed for the JA6681 :
Round all around JMLC Iwata 300 with JA6681B:
Overall: The modern Iwata 300 would have been the clear winner with its round and smooth inside and nice extra vertical dimensions, but only if it was bigger and went deeper. Thanks to the 70cm wide Yamaha horn, I can taste the sweet nectar of pushing the JA6681B down to 300-400Hz. But maybe deeper is just producing distorted garbage anyway. 600Hz is a worthy goal for a problem free presentation.
Additionally, and more importantly in reference to the latest posts - round is the way to go. No more square vintage horns will enter this house.
Powered by V-Fet. Thanks Papa.
Last edited:
When I made my adapters, I noticed from plastic ebay adapters I had previously, that the weights can become quite substancial with compression drivers. I can see it bending! This is why I opted for mdf and not printing. It required some elbow grease with a file and some pretty wonky drill bits, but I am satisfied. If I keep the old Yamaha horns, I might smooth it with some wood paste
and grind some more.
Am I seeing it right? 4 ways? Woofer, big horns, smaller horns on top and some ribbon tweeter on top of those?
Sorry I am afraid I cannot be of much help to you. All I know is PETG must be printed hot and slow. Being aware that 3D printing can be challenging in its own right, I leave this to others. A friend of mine who runs a mechanical engineering company takes care of this 3D printing. He has a fleet of printers which he changes regularly, and is better versed at using them than I ever would want to be.
PS He offered to give me one of his printers and I kindly refused LOL.
Thank you for sharing these videos. The difference in tonality between these two horns is quite obvious, making it easy to understand how one might prefer A over B - or vice versa.
V-Fet amps are very sweet indeed
Still on my to do list... Spent last week at the lake cabin rewriting the OpenSCAD software that keeps growing, went to a sailing class Saturday, to the Audiophile Cafe on Sunday, and now am wrapping up the software rewrite. Should be testing the new adapter soon, and order Prototype #3 at about the same time. Thanks for your patience ...
This might represent a way of assessing or at least measuring the differences in adapters.
https://pearl-hifi.com/06_Lit_Archi...Keith/Cepstral_Analysis/Cepstral_Analysis.pdf
"3.2 Practical Uses of Cepstral Analysis
The example loudspeaker is one of twenty mid-range loudspeakers which formed the test examples in a listening test conducted as part of a three-year project on mid-range horn loudspeakers [2]. Some of the aims of the test were to find which, if any, of a range of horns sounded similar to example direct-radiating loudspeakers and what physical characteristics make a horn sound like a horn. Measurements were taken of the frequency response functions of all of the loudspeakers and comparisons were made between these measurements and the listening test results. The power cepstra of the loudspeakers were calculated as above and, when compared, were found to reveal many clues as to the similarity or otherwise of the sound of the different loudspeakers. For example, it was discovered that the shorter horns sound more like the direct radiating loudspeakers than the long horns, and this was found to be due to the reflections from the mouths of the short horns occurring at around the same time as those in the cones of the direct-radiating loudspeakers; the reflections from the mouths of the long horns occur later and contribute to the characteristic 'horn sound' even though they are generally lower in level than those of the short horns. During the same horn research, a flare mismatch at the throat of one of the horns when attached to a particular driver was found to be the cause of response irregularities. Cepstral analysis revealed quite strong reflections from the discontinuity and when the horn was attached to another driver both the spikes on the cepstrum and the response irregularities disappeared."
https://pearl-hifi.com/06_Lit_Archi...Keith/Cepstral_Analysis/Cepstral_Analysis.pdf
"3.2 Practical Uses of Cepstral Analysis
The example loudspeaker is one of twenty mid-range loudspeakers which formed the test examples in a listening test conducted as part of a three-year project on mid-range horn loudspeakers [2]. Some of the aims of the test were to find which, if any, of a range of horns sounded similar to example direct-radiating loudspeakers and what physical characteristics make a horn sound like a horn. Measurements were taken of the frequency response functions of all of the loudspeakers and comparisons were made between these measurements and the listening test results. The power cepstra of the loudspeakers were calculated as above and, when compared, were found to reveal many clues as to the similarity or otherwise of the sound of the different loudspeakers. For example, it was discovered that the shorter horns sound more like the direct radiating loudspeakers than the long horns, and this was found to be due to the reflections from the mouths of the short horns occurring at around the same time as those in the cones of the direct-radiating loudspeakers; the reflections from the mouths of the long horns occur later and contribute to the characteristic 'horn sound' even though they are generally lower in level than those of the short horns. During the same horn research, a flare mismatch at the throat of one of the horns when attached to a particular driver was found to be the cause of response irregularities. Cepstral analysis revealed quite strong reflections from the discontinuity and when the horn was attached to another driver both the spikes on the cepstrum and the response irregularities disappeared."
Thank you! Nice find 
.
This method might indeed show the reflections at the adapter interfaces. In any case, it looks like a useful approach to study such time-domain effects. Have Matlab... might give this a shot...
Do you know the author? I ordered the book he co-authored. Any other specific paper of his warrants reading?
.This method might indeed show the reflections at the adapter interfaces. In any case, it looks like a useful approach to study such time-domain effects. Have Matlab... might give this a shot...
Do you know the author? I ordered the book he co-authored. Any other specific paper of his warrants reading?
There is one in this post
https://www.diyaudio.com/community/...w-distortion-with-a-2-way.334757/post-5926313
Page 59 for "Round the Horn" from 1994
https://worldradiohistory.com/Archive-All-Audio/Archive-Studio-Sound/90s/Studio-Sound-1994-03.pdf
Plenty of stuff from the ISVR that I can't remember
His analysis in the latter paper is much more nuanced and complete than in his initial one. Very good stuff. I love that he included an electrostatic in the test. No horns with fins alas, but the results for the cellular Altec are quite a bit different than those of other horns! Lots of good material on cepstrum analysis too. Gives us lots to ponder. Cheers!
I am familiar with Holland and Newell's work in general, and I have their excellent book "Loudspeakers for music recording and reproduction".
That particular article, though (and the related research on horns also published in their book and elsewhere), suffers from some generalizations and jumps to conclusions and sweeping recommendations a bit too soon, IMHO.
For instance, their insistence on an arbitrary "maximum recommended length" of 35cm for all horns (to avoid unpleasant "horn sound" - which they refer to as "archetype C", as opposed to a more neutral "direct radiator sound" = "archetype B") is not justified, and it seems to stem from anecdotal evidence on a few specific sub-optimal devices.
It is also directly contradicted by their own findings in at least one case, where the authors themselves observe "One of the 'long' horns was a Fostex wooden -flared sectoral horn, 440mm in length from throat to mouth. Possibly more strongly than any other horn in the entire test, this horn was judged to sound like the direct radiator cone, archetype B"
About the same Fostex horn, they later say "The horn produced undesirable throat impedance plots, implying an uneven pressure amplitude response when connected to a driver, nonetheless in auditioning before the tests began, the horn was generally considered 'musical', pleasant to listen to, and definitely not horn-like."
To me, this comes across as an admission that their theorised correlation between throat impedance plots and sound quality does not hold water... or at the very least, that there are more factors at play, some of which may end up being more relevant in the grand scheme of things.
@marco_gea We are in agreement on the discussion and conclusion sections. But however imperfect these are, I do find them much improved compared to the earlier paper. My main reservation on this initial paper was taking resemblance to Reference B (direct radiator) to be desirable (over resemblance to other References A, C, and D). Why should Reference B be considered "truer" than the other references, including the electrostatic?
Now the author admits in the more recent paper that ...
"this poses an interesting philosophical point: should a monitor loudspeaker be rejected because it does not pose the midrange problems inherent in most other production loudspeakers?" Kudos here. And kudos for sharing all the raw results.
It is looking at the raw measurements and test results that I find particularly valuable. We can draw our own hypotheses. For instance the long horn speakers were found to be more often similar, than shot horn speakers, to Reference A which was the Quad electrostatic. And of the long horn speakers, the cellular Altec was the one that sounded the most often similar to the Quad!
I've got to admit that, despite other virtues of the JMLC, I find the TH4001 to be less colored. And I've been wondering why. We're talking mild colorations here, but still...
Now the author admits in the more recent paper that ...
"this poses an interesting philosophical point: should a monitor loudspeaker be rejected because it does not pose the midrange problems inherent in most other production loudspeakers?" Kudos here. And kudos for sharing all the raw results.
It is looking at the raw measurements and test results that I find particularly valuable. We can draw our own hypotheses. For instance the long horn speakers were found to be more often similar, than shot horn speakers, to Reference A which was the Quad electrostatic. And of the long horn speakers, the cellular Altec was the one that sounded the most often similar to the Quad!
I've got to admit that, despite other virtues of the JMLC, I find the TH4001 to be less colored. And I've been wondering why. We're talking mild colorations here, but still...