lyd&akustikk
Member
How does a smaller horn compare to a very large horn?
While the JBL 2384 horn isn’t a small horn in many opinions with a dimension of approximately 76 cm width, 38 cm height and 28 cm depth, it’s actually quite small acoustically speaking. Directivity control from a horn demands size. The horn we’re comparing it to is 110 cm wide, 80 cm tall and 54 cm deep.
JBL 2384 has many traits with the horn in the JBL M2 monitor.
In regards to directivity, it measures better than the horn in M2 because it's bigger.
Horizontally, the width of 76 cm together with a wider beamwidth of 100°, the horn is able to maintain the directivity quite low in frequency. However, it’s not super uniform. Between 700 Hz and 900 Hz is in the area of 70-80° and above 4500 Hz is becomes narrower and is generally uneven in the highs.
Vertically the horn is definitely very small with a height of 38 cm height. The vertical beamwidth is as narrow as 50° and it maintains that high in frequency.
However, due to the low height, the directivity collapses high in frequency, thus the spectral content changes much because of the reflective energy from the ceiling and floor. The 2384 loses its directivity already at 1500 Hz and quite abruptly below this frequency. Already at 1000 Hz, the beamwidth has increased to 80° and continue to increase considerably below this.
The horn we’re comparing 2384 to has a beamwidth of approximately 85x55° and is extremely uniform. In other words, it’s narrower horizontally and minimizes room reflections to a greater degree. Vertically is has a similar beamwidth to the 2384 but because it’s more than double in height, it’s able to maintain a constant directivity much lower. To about 640 Hz and with a gradually loss below that. The large horn maintains the horizontal directivity more or less all the way up and doesn’t narrow in the highs.
The woofer used for both horns is a 15” woofer with low mass, sensitivity of 99 dB but still decent Xmax. Especially in the midrange it’s much more open and detailed compared to the well know JBL 2226 woofer. Enclosure is vented with large and long vents with net volume about 95L.
No scientific comparison
Comparing two different horns like this is not an easy task and no ABX (blind test) was performed.
For the JBL 2384 a JBL 2450SL (acuaplas coating) compression was used. And Radian 951BePB with the large horn. Some may object and to the fact that large horn had a compression driver with beryllium. But in our experience this doesn’t make a big difference. A comparison between Klipsch K-402 with Radian 950BePB (beryllium) vs a prototype horn similar to the large horn here with JBL 2450SL made it clear that the horn is far more important - where the prototype horn with a “poorer” driver than used in the K-402 horn still overall sounded much better. Especially when using active crossover and where EQ minimizes the difference. The beryllium diaphragm basically make things smoother and is something that is more experienced as less listening fatigue over time.
When equalizing the horns, several choices needs to be made. The large horn doesn’t really change in the response whether you measure it right outside the horn lips, at 50 cm distance or at longer distances as long as the room doesn’t affect the measurements. The JBL 2384 horn on the other hand changes already at 40 cm distance compared to right outside the horn lips. I ended up equalizing the 2384 at about 40 cm distance.
In order to get both horns on same ear level, either the JBL 2384 needed to be lifted or the woofer raised from the floor. I started raising the horn, but found out that this didn’t work that well.
Integration in the crossover region with the woofer became more troublesome. Having them closed space together worked much better. Therefore, the woofer had to be raised but this changed the woofer response higher in frequency and new EQ had to be applied to the woofer.
Both horns were equalized flat at 40 cm distance and same shelving for the response to fall towards the treble was applied to both. They were level matched.
When comparing, mono was primarily used. Only some stereo at the very end. One clear benefit with mono is that is goes much faster switching from one speaker to the other, which is important when AB testing. A drawback is less spaciousness when listening in mono. Some marks were made in the floor, and the speaker was moved to this point in the test. Despite of using mono, it still takes time to switch from one speaker to the other and upload a new filter in the DSP.
Crossover for JBL 2384 and the woofer was 900 Hz with 4th LR. Crossing over lower would have reduced the sensitivity. Crossover for the large horn and woofer was either 600 Hz or 700 Hz, also 4th LR.
The response in the listening position was quite similar. However, the large horn was somewhat flatter in the area of 1 kHz and 2.5 kHz and there were some other minor differences. I don’t believe one should correct for this in the listening position because it’s how the horn reacts with the surfaces in the room. Besides, it’s not minimum phase behavior - meaning equalization will not work well.
Electronics and acoustics:
- Hypex DLCP as DAC/pre and active crossover with IIR. Perfect time alignment between woofer and horn.
- Vera Audio P150/600 RS power amplifier with world class performance and measurements. One channel feeding the woofer and one feeding the horn.
The DLCP measures very well but do lack a bit of transparency vs a SOTA measuring DAC.
The room is treated acoustically with broadband absorption in ceiling and on side walls, some sheep wool rugs were on the floor, and RPG diffusers on the rear wall. Distance to the diffusers was minimum 237 cm. It’s very likely that with no or less acoustic treatment, the difference between the horns would have been greater. Acoustic treatment will minimize deviations of the off-axis response.
Listening impressions
Listening back and forth with different types of music, made it very clear that the differences were large between the two. The JBL 2384 sounds very none fatiguing but lacks the clarity, openness and details of the larger horn. It sounds like good hifi but never realistic (or an illusion of vocals and instruments). Compared to what one listens to at hifi exhibitions and hifi stores, the quality is way behind most of that. It’s very odd too see super expensive speakers using old fashioned passive crossovers with many compromises. Running active crossover with the JBL 2384 and a quality woofer it’s really a cheap way to get a good sounding setup.
Listening to great female voices like Mary Black, Alison Krauss and Alanis Morrissette or the male voice of Bjørn Eidsvåg makes a great distinguish between the two though. Switching to the large horn is a bit like pulling back the drapes and getting a much clearer view into the music material. Suddenly you experience Alison Krauss’ voice is her actual voice. And the flute on Sondre Bratland’s “Fløytelåt” gives you the impression that you’re listening to the real instrument. Many have experienced that good horns can sound almost spooky realistic, but the JBL 2384 doesn’t seem to achieve this. When listening to Sissel Kyrkjebø high pitched voice in “Sarah’s Song”, the higher frequency content is dull with the JBL 2384 compared to the large horn.
But it’s not only about clarity, details, resolution and high frequency energy. It’s also about a coherent sound stage and size. One would maybe think that when crossing the large horn at 700 Hz and the JBL at 900 Hz and the latter being spaced much closer to the woofer, the JBL 2384 would at least sound as coherent or “together” as the larger horn with greater spacing between the drivers. But that’s not the case. Quite the opposite. The larger horn sounds with both vocals and instruments more coherent. Which again leads to a larger sound stage and more like the real deal.
Listening to the large horn moves you with both engagement and with beauty of the music. The JBL 2384 while sounding good and none fatiguing ends up not moving you emotionally in any similar capacity, and is somewhat boring when you have heard the larger horn.
At the end I also switched to a front firing speaker (not a horn) with constant but wider directivity, with a very coherent sound stage and no vertical lobing. While this speaker has much lower sensitivity (something many claim to be important for engaging sound) and the woofer is also smaller I also found this speaker to be more engaging to listen to compared to the JBL 2384. Mainly because it’s sounded much more coherent, though it doesn’t have the resolution, detail and realism of the large horn. Vocals and instruments sounds like a oneness without any break up like real music does, and which is is very important to me. This is a speaker by the way that we hope to release in 2022.
Conclusion:
Someone asked me once when I shared a picture of a large horn if we couldn’t design a small horn too. While we can and also will design a smaller horn compared to the large horn here, it’s very unlikely that we’ll every release a horn that is as small as the JBL 2384 in height and needs to be crossed over as high as 900 Hz area. The drawbacks and compromises are too big, and there are other designs than horns that are better then.
Great horn speakers need to be of a certain size to be of high quality combined with a uniform directivity in both planes. If they are, they can be incredible.
While the JBL 2384 horn isn’t a small horn in many opinions with a dimension of approximately 76 cm width, 38 cm height and 28 cm depth, it’s actually quite small acoustically speaking. Directivity control from a horn demands size. The horn we’re comparing it to is 110 cm wide, 80 cm tall and 54 cm deep.
JBL 2384 has many traits with the horn in the JBL M2 monitor.
In regards to directivity, it measures better than the horn in M2 because it's bigger.
Horizontally, the width of 76 cm together with a wider beamwidth of 100°, the horn is able to maintain the directivity quite low in frequency. However, it’s not super uniform. Between 700 Hz and 900 Hz is in the area of 70-80° and above 4500 Hz is becomes narrower and is generally uneven in the highs.
Vertically the horn is definitely very small with a height of 38 cm height. The vertical beamwidth is as narrow as 50° and it maintains that high in frequency.
However, due to the low height, the directivity collapses high in frequency, thus the spectral content changes much because of the reflective energy from the ceiling and floor. The 2384 loses its directivity already at 1500 Hz and quite abruptly below this frequency. Already at 1000 Hz, the beamwidth has increased to 80° and continue to increase considerably below this.
The horn we’re comparing 2384 to has a beamwidth of approximately 85x55° and is extremely uniform. In other words, it’s narrower horizontally and minimizes room reflections to a greater degree. Vertically is has a similar beamwidth to the 2384 but because it’s more than double in height, it’s able to maintain a constant directivity much lower. To about 640 Hz and with a gradually loss below that. The large horn maintains the horizontal directivity more or less all the way up and doesn’t narrow in the highs.
The woofer used for both horns is a 15” woofer with low mass, sensitivity of 99 dB but still decent Xmax. Especially in the midrange it’s much more open and detailed compared to the well know JBL 2226 woofer. Enclosure is vented with large and long vents with net volume about 95L.
No scientific comparison
Comparing two different horns like this is not an easy task and no ABX (blind test) was performed.
For the JBL 2384 a JBL 2450SL (acuaplas coating) compression was used. And Radian 951BePB with the large horn. Some may object and to the fact that large horn had a compression driver with beryllium. But in our experience this doesn’t make a big difference. A comparison between Klipsch K-402 with Radian 950BePB (beryllium) vs a prototype horn similar to the large horn here with JBL 2450SL made it clear that the horn is far more important - where the prototype horn with a “poorer” driver than used in the K-402 horn still overall sounded much better. Especially when using active crossover and where EQ minimizes the difference. The beryllium diaphragm basically make things smoother and is something that is more experienced as less listening fatigue over time.
When equalizing the horns, several choices needs to be made. The large horn doesn’t really change in the response whether you measure it right outside the horn lips, at 50 cm distance or at longer distances as long as the room doesn’t affect the measurements. The JBL 2384 horn on the other hand changes already at 40 cm distance compared to right outside the horn lips. I ended up equalizing the 2384 at about 40 cm distance.
In order to get both horns on same ear level, either the JBL 2384 needed to be lifted or the woofer raised from the floor. I started raising the horn, but found out that this didn’t work that well.
Integration in the crossover region with the woofer became more troublesome. Having them closed space together worked much better. Therefore, the woofer had to be raised but this changed the woofer response higher in frequency and new EQ had to be applied to the woofer.
Both horns were equalized flat at 40 cm distance and same shelving for the response to fall towards the treble was applied to both. They were level matched.
When comparing, mono was primarily used. Only some stereo at the very end. One clear benefit with mono is that is goes much faster switching from one speaker to the other, which is important when AB testing. A drawback is less spaciousness when listening in mono. Some marks were made in the floor, and the speaker was moved to this point in the test. Despite of using mono, it still takes time to switch from one speaker to the other and upload a new filter in the DSP.
Crossover for JBL 2384 and the woofer was 900 Hz with 4th LR. Crossing over lower would have reduced the sensitivity. Crossover for the large horn and woofer was either 600 Hz or 700 Hz, also 4th LR.
The response in the listening position was quite similar. However, the large horn was somewhat flatter in the area of 1 kHz and 2.5 kHz and there were some other minor differences. I don’t believe one should correct for this in the listening position because it’s how the horn reacts with the surfaces in the room. Besides, it’s not minimum phase behavior - meaning equalization will not work well.
Electronics and acoustics:
- Hypex DLCP as DAC/pre and active crossover with IIR. Perfect time alignment between woofer and horn.
- Vera Audio P150/600 RS power amplifier with world class performance and measurements. One channel feeding the woofer and one feeding the horn.
The DLCP measures very well but do lack a bit of transparency vs a SOTA measuring DAC.
The room is treated acoustically with broadband absorption in ceiling and on side walls, some sheep wool rugs were on the floor, and RPG diffusers on the rear wall. Distance to the diffusers was minimum 237 cm. It’s very likely that with no or less acoustic treatment, the difference between the horns would have been greater. Acoustic treatment will minimize deviations of the off-axis response.
Listening impressions
Listening back and forth with different types of music, made it very clear that the differences were large between the two. The JBL 2384 sounds very none fatiguing but lacks the clarity, openness and details of the larger horn. It sounds like good hifi but never realistic (or an illusion of vocals and instruments). Compared to what one listens to at hifi exhibitions and hifi stores, the quality is way behind most of that. It’s very odd too see super expensive speakers using old fashioned passive crossovers with many compromises. Running active crossover with the JBL 2384 and a quality woofer it’s really a cheap way to get a good sounding setup.
Listening to great female voices like Mary Black, Alison Krauss and Alanis Morrissette or the male voice of Bjørn Eidsvåg makes a great distinguish between the two though. Switching to the large horn is a bit like pulling back the drapes and getting a much clearer view into the music material. Suddenly you experience Alison Krauss’ voice is her actual voice. And the flute on Sondre Bratland’s “Fløytelåt” gives you the impression that you’re listening to the real instrument. Many have experienced that good horns can sound almost spooky realistic, but the JBL 2384 doesn’t seem to achieve this. When listening to Sissel Kyrkjebø high pitched voice in “Sarah’s Song”, the higher frequency content is dull with the JBL 2384 compared to the large horn.
But it’s not only about clarity, details, resolution and high frequency energy. It’s also about a coherent sound stage and size. One would maybe think that when crossing the large horn at 700 Hz and the JBL at 900 Hz and the latter being spaced much closer to the woofer, the JBL 2384 would at least sound as coherent or “together” as the larger horn with greater spacing between the drivers. But that’s not the case. Quite the opposite. The larger horn sounds with both vocals and instruments more coherent. Which again leads to a larger sound stage and more like the real deal.
Listening to the large horn moves you with both engagement and with beauty of the music. The JBL 2384 while sounding good and none fatiguing ends up not moving you emotionally in any similar capacity, and is somewhat boring when you have heard the larger horn.
At the end I also switched to a front firing speaker (not a horn) with constant but wider directivity, with a very coherent sound stage and no vertical lobing. While this speaker has much lower sensitivity (something many claim to be important for engaging sound) and the woofer is also smaller I also found this speaker to be more engaging to listen to compared to the JBL 2384. Mainly because it’s sounded much more coherent, though it doesn’t have the resolution, detail and realism of the large horn. Vocals and instruments sounds like a oneness without any break up like real music does, and which is is very important to me. This is a speaker by the way that we hope to release in 2022.
Conclusion:
Someone asked me once when I shared a picture of a large horn if we couldn’t design a small horn too. While we can and also will design a smaller horn compared to the large horn here, it’s very unlikely that we’ll every release a horn that is as small as the JBL 2384 in height and needs to be crossed over as high as 900 Hz area. The drawbacks and compromises are too big, and there are other designs than horns that are better then.
Great horn speakers need to be of a certain size to be of high quality combined with a uniform directivity in both planes. If they are, they can be incredible.
Last edited: