Upgrading to Avantgarde Colibri C2

[LTig]

You don't need to explain that to me, I know how normal loudspeakers work. What is not explained is why this horn speaker behaves opposite to a conventional speaker: upwards tilting treble with distance, vs. downwards tilting treble for normal speakers. I have an explanation - the horn and the woofer below it are dropping off at different rates according to distance.

This is correct. I'm not an expert regarding horn technology, so my explanation might be wrong. However: the size of the horns mouth (where it's widest) is larger than the wavelength of the higher frequencies, so for those the horn acts like a 2 dimensional array of small tweeters. Similar to a line array SPL does not go down in square law if you are close enough (you might eventually experience square law if the distance is so large that the array is seen as a point source).

The explanation is NOT simple "directivity", it is something else.
If it WAS directivity, both mid and woofer would be dropping off at the same rate. I still think that my horns (and also the Avantgardes I have heard) have an energy concentrating effect that we are not capturing with conventional measurements. Not to say that it's not measurable, because I have measured it. It's just not described in the literature, as far as I know.

The Klippel NFS measures the 360 degree sphere so it captures everything the speaker emits, and uses the data to calculate the anechoic frequency response at 1m distance (and other curves). What we haven't seen is the anechoic frequency response calculated for different listening distances. It would be interesting to do this for several distances (e.g. 1m, 2m, 3m, 4m, 5m) for a normal speaker and one with a large horn. i bet they would show what you measured.

 

[Keith_W]

Not what I meant. If the microphone was placed at the tweeter height, for example, it would be increasingly off the midrange horn axis as it is moved closer to the speaker. Since the horn is fairly high DI, this may be a substantial source of error. Another problem is the phase relationship between the different "ways" when near vs far. This is why I suggested measuring the midrange horn by itself.

Sure, I have that measurement.

Measurement of the midrange horn, on-axis, at the distances indicated (0cm - MLP; MLP is 3m from the speaker). I have hidden all the measurements except the 10cm and 3m measurement. 15 cycle FDW was applied. Note I am using a 30dB vertical scale so that you can better appreciate the SPL difference. What I see: about 1dB loss at 3m.

Measurement of the woofer, again on-axis, at the distances indicated. Same vertical scale. About 4-5dB loss at 3m (ignoring all the non-minphase behaviour and using the eye of faith!).

 

[Keith_W]

No offense, but I would bet that you have NOT found new physics.

I am not proposing that I have found new physics. I am wondering whether existing physics that describes line arrays, also applies to a horn. The observation is right in front of you. It is either a mistake I made with my measurement, or it is something that needs an explanation. I am not so arrogant as to think I am infallible, which is why I need to repeat this experiment with another horn speaker. If I can find one, and if the owner is happy to let me take those measurements.

 

[thewas]

15 cycle FDW was applied.

If I were you I would repeat those measurements with REW and a time fixed window before the first reflection, unfortunately only usually a couple of ms in normal rooms, you can extend it though with a ground plane placement to remove the room influence which is there with FDW.

 

[Sokel]

If I were you I would repeat those measurements with REW and a time fixed window before the first reflection, unfortunately only usually a couple of ms in normal rooms, you can extend it though with a ground plane placement to remove the room influence which is there with FDW.

One can also apply multi-time windowing across the spectrum with new versions:

 

[Keith_W]

If I were you I would repeat those measurements with REW and a time fixed window before the first reflection, unfortunately only usually a couple of ms in normal rooms, you can extend it though with a ground plane placement to remove the room influence which is there with FDW.

I don't know what you think that is going to show. The point isn't all the dips in the response caused by including reflections in the measurement, the point is the overall volume loss - that is demonstrated in the measurements I already posted. But if it makes you happy, I have applied a shorter window.

So here is the horn at 10cm (red) and MLP (black) with a FDW of 1.

And same thing with the woofer.

 

[thewas]

The point isn't all the dips in the response caused by including reflections in the measurement, the point is the overall volume loss - that is demonstrated in the measurements I already posted.

Because in a non anechoic measurement the volume is a summation of not only the direct sound.

It would be interesting to see these

with sufficient long anechoic gates.

By the way even a low FDW does not guarantee reflection free measurement, that's why I would recommend using REW where it easier to see and place your gating.

 

[illusonic]

Basically the same level at 10cm and 3m? Not even a line array would give such results. And the woofer also shows equal level above 1kHz. Below we may see effects of baffle step. If these were my measurements, I would check if something went wrong with absolute levels.

 

[Flaesh]

existing physics that describes line arrays, also applies to a horn

I believe that a conventional horn is a source of a spherical wave. I see no reason otherwise. A cylindrical wave requires a linear source.
I don't know how to interpret the experiment above. I can't try it myself yet.

 

[bmc0]

Sure, I have that measurement.

I have to suspect level normalization. Perhaps the software does this automatically? The woofer (a direct radiator) is also approximately the same level circa 1kHz.

 

[Keith_W]

I have to suspect level normalization. Perhaps the software does this automatically? The woofer (a direct radiator) is also approximately the same level circa 1kHz.

Acourate does not do anything unless you specifically tell it to. It has not been normalized, and the only manipulation was to apply a FDW. And no, I have not been sneakily adjusting the measurement volume. BUT ... you also have the full range measurement that shows the exact same thing - woofers dropping faster than the horns, see a few posts up. People didn't believe it when I posted that measurement, so they wanted individual measurements of the driver. Now I post individual driver measurements, and you think it's been normalized to a different volume.

Without going into subjectivist mumbo-jumbo, it sounds very different to a conventional speaker. There is an immediacy to the sound of horn speakers. Listen to one, and you will hear it immediately. It's not just my speakers that do that either, I have also heard it with Avantgardes (many models), JBL Everest, Tune Audio Anima, and most recently the Aries Cerat Aurora that I heard a few days ago. Then you start to wonder where you might see it in the measurements. I wonder whether @Chris A might have anything to contribute to the discussion, if i'm not mistaken, he has Klipsch Horns.

 

[Sokel]

Acourate does not do anything unless you specifically tell it to. It has not been normalized, and the only manipulation was to apply a FDW. And no, I have not been sneakily adjusting the measurement volume. BUT ... you also have the full range measurement that shows the exact same thing - woofers dropping faster than the horns, see a few posts up. People didn't believe it when I posted that measurement, so they wanted individual measurements of the driver. Now I post individual driver measurements, and you think it's been normalized to a different volume.

Without going into subjectivist mumbo-jumbo, it sounds very different to a conventional speaker. There is an immediacy to the sound of horn speakers. Listen to one, and you will hear it immediately. It's not just my speakers that do that either, I have also heard it with Avantgardes (many models), JBL Everest, Tune Audio Anima, and most recently the Aries Cerat Aurora that I heard a few days ago. Then you start to wonder where you might see it in the measurements. I wonder whether @Chris A might have anything to contribute to the discussion, if i'm not mistaken, he has Klipsch Horns.

Amir's review for JBL 590 may be a tell but it's contradictory.

While the CEA2034 shows elevation in highs, EIR looks normal (yes I know it's estimated) and most importantly, non-anechoic near field one looks totally normal.

Smaller horn maybe, with different geometry but still horn.

 

[Purité Audio]

I had horns here for years and then the opportunity to compare them side by side to a variety of ‘conventional ‘ passive and active designs, i never found them more immediate, various amounts of colouration odd dispersion yes,’ immediate’ no.
Keith

 

[Mr. Widget]

I had horns here for years and then the opportunity to compare them side by side to a variety of ‘conventional ‘ passive and active designs, i never found them more immediate, various amounts of colouration odd dispersion yes,’ immediate’ no.
Keith

For decades I have had horn systems and was convinced they were more immediate and "lifelike" in their dynamics. I was particularly drawn to the large format horns/drivers from JBL, TAD, Meyer... then recently I swapped in a pair of lowly cone and dome speakers from D&D and really haven't noticed the lack of dynamics I always thought I had heard with Revels and others.

I decided to stop worrying about it and just enjoy the music.

 

[bmc0]

woofers dropping faster than the horns

Only below ~500Hz in the individual measurement. Above that, your measurement of a direct-radiating woofer shows almost zero level difference at 10cm vs 3m. I doubt the measurement because there is no mechanism by which a single direct-radiating driver (or horn, for that matter) of that diameter, measured on axis, could exhibit such behavior over that frequency range in free space without level normalization. It is simply a physical impossibility.

 

[dc655321]

I am wondering whether existing physics that describes line arrays, also applies to a horn. The observation is right in front of you.

It does not, for reasons right in front of you .

 

[Keith_W]

OK I must sheepishly withdraw what I said earlier. I got fed up of you guys and decided to repeat the experiment. "I'll show them!" ... or so I thought.

Tape measure on the floor, distance eyeballed. I wasn't too precise about it, the error is probably +/- 5cm. The mic was pointed at the speaker, and measurements were taken at 50cm intervals. Upon examination, the 0cm and 50cm measurements were obviously unrepresentative (parallax error + weird phase cancellations between drivers) so they were discarded.

Because I was using a loopback timing measurement, we have a rather cool display of how the impulse peak shifts depending on microphone position. Yes, that's about a 2.73 second latency, the reason is because of a complicated measurement circuit which involves getting REW to send output to my FIR convolver.

So here is the 100cm measurement (green) vs. 300cm MLP (purple) measurement with no volume normalisation. Well i'll be damned, it's starting to behave like a normal speaker now. The high frequencies are dropping off faster than the bass.

As an aside you can also see the effect of DSP. The curve is smoothest at the listening position - all those giant bass peaks disappear.

This made me wonder where I went wrong with the previous measurement. So I decided to do an MMM. I swept the mic on a vertical axis (up/down) at 50cm intervals.

Here is 100cm (dark green) vs. 300cm (lighter green), again with no volume normalisation. The rising frequency response above 15kHz might be artefact from wind noise. Well i'll be damned again.

Once the SPL's are aligned we can clearly see the the two curves diverge above 2kHz, and the horn mids and horn tweeters indeed do follow the 1/r^2 fall-off.

I have no idea why my previous measurements were wonky. Apologies to all of you, mea culpa.

 

[Flaesh]

But it remains possible to continue discussing whether this feature of horn (there's also this fuzzy boundary with the waveguides)) speakers exists and what it is.

that conventional speakers like the KEF's don't have.

The directivity of different products is quite different. Hor example, IIRC SEOS24 (wich I usuaslly use) measured by mtg90:

Dynamic capabilities?.. how much exactly
Some special "horn distortions"? The other day, a guest heard them only once in 2-3 hours of listening to music, on vocal in Death Letter (Cassandra Wilson|New Moon Daughter).

 

[Gorgonzola]

They have arrived!

View attachment 543066

Christmas came early this year.

The avantgarde speakers have an amazing build quality, wow, german engineering. With the light in the room their color changes, I love their looks. My kids asked if this was a new sort of toilet.

After using roomperfect on the lyngdorf 1120 and integrating my sub I have arrrived at hifi heaven. For me, a clear upgrade to the Kef LS60. More “live”, punchy sound and voices that are better defined in the soundstage. Voices sound more lifelike and “real”. The music comes in more layers, you can here the different parts better separated from each other.

Another good thing: Tidal playback on the Lyngdorf is faster and more responsive than on the Kefs.

Amazed again about how good roomperfect is. I use dirac and audessey also, but roomperfect just sounds right to me. Bass is tight and sounds just right. Love it.

View attachment 543070

View attachment 543068

View attachment 543067

View attachment 543071


I almost destroyed my SVS sub… Scrolling through the Lyngdorf options, I noted a “full scale” option. After selecting this, my SVS almost seemed to explode, it was suddenly outputting max power. This option will send a maximum signal to the output. Should be accompanied by a a warning I guess.

View attachment 543073

The SVS with the Lyngdorf on top:

View attachment 543075

Very handsome!! At least to look at; not sure about the sound ... but then how could I be? I'm 81.