Good questions. Challenging and long to answer, and perhaps not satisfactory to most readers.
Let's start off with the easy one:
1. Max power handling depends on the test signal standard and how it's applied. For decades it was mainly a test of durability, using pink noise, and implemented with a filter to protect the driver. As usual, with standards, there is never one standard.
IEC RMS power handling (IEC 268-5, cl 18.4 Rated noise test) –
100 hours(!) continuously, with an IEC shaped pink noise signal.
Typically gives the lowest, worse case scenario value
IEC Long Term Power Handling –
signal for one minute and pause for 2 min, 10 times alternating. Total test time is 28 min. (IEC 268-5, cl 18.2)
IEC Short Term Power Handling –
signal for one second and pause for 1 min, repeat for 60 times. Total test time one hour. (IEC 268-5, cl 18.1)
AES (1984) Power handling= voltage ^2 / Zmin, using Pink noise signal with crest factor 2; bandpass filtered at 12dB per octave with Butterworth filter response as specified by manufacturer. The rated power of the device shall be that power the device can withstand for
2 hours without permanent change in acoustical, mechanical, or electrical characteristics greater than 10%.
AES Music program -
twice the continuous AES power rating.
With an arbitrary high pass of a frequency eg. 2000, 1000Hz, the manufacturer must be willing to push the unit within an inch of it's life... well, because, that is what the old standards require. ie.
100 hours or 2 hours of continuous pink noises, without permanent damage
You can see where the challenges lay. You (or I) do not have much interest in how much power it can handle. What we're really interested in is how loudly it can play, cleanly, at say 1m, so we can makes educated estimates for our intended application eg. centre speaker for cinema speaker. The power handling is just a proxy.
But music is very dynamic, and whilst these pink noise signals may be useful stress tests for drive units, it will under-rate the maximum SPL with music or speech, which is what listeners are most interested in.
An emerging standard is the AES75-2023. This uses Music-Noise; which is a noise signal that more closely emulates the dynamic characteristics of music:
Reference:
https://audioxpress.com/news/meyer-...l-for-real-world-loudspeaker-spl-measurements
With the Purifi tweeter not yet in mass production, I'm not certain we'll know, in the short term, what the maximum long term power handling is. And even if we did, it doesn't fully answer your question of the maximum SPL with real music content.
Some testers perform successive sine sweeps, from 76dB, then 86 dB, then 96 dB, then 102dB etc. This is one method. The downside is that each sweep takes 3 secs; which means a total of 12 seconds of sine wave stress testing. It may exhibit reduced output at the final test, but is representative of compression of real music?
It’s a quandry!
References:
*
https://www.aes.org/standards/AES75/
2. It is an assumption to say that "better power response and directivity from lower crossover point and better high frequency dispersion. That's already a lot of areas of improvement over other designs"
In the
original patent granted to Dr. Sean Olive 20 years ago the power response is used only in relation to determining the low frequency extension.
What was found to be important (with a correlation of 0.86 with 13 loudspeakers in mono, in their listening room, and 0.7 when applied to the 70 loudspeakers) was
the lack of resonances in the on-axis response, lack of resonances in the predicted in-room response, the bass extension and the smoothness of the predicted in room response.
(bold: my emphasis)
More specifically:
Preference Rating=12.69 - 2.49*NBD_ON - 2.99*NBD_PIR - 4.31*LFX - 2.32*SM_PIR
where-
NBD: Average Narrow Band Deviation in each 1/2-octave band from 100 Hz-12 kHz
ON: On-axis
PIR: Predicted in-room response
LFX: Low frequency extension (Hz) based on -6dB frequency point transformed to log; applied to sound power relative to mean sensitivity in listening window from 300 Hz - 10 kHz
SM: Smoothness; (r) in amplitude response based on a linear regression line through 100 Hz-16kHz
In fact, having a lower crossover point often has the opposite effect of enabling a smooth predicted in-room response, but that is another topic for another day.
For a dome tweeter, the directivity depends on the interaction with nearby surroundings, such as a phase shield, structures abutting the dome, adjacent surfaces which may act as a waveguide or horn, as well as further a field like the cabinet dimensions and shape.
Now let us look at the Directiva ASR, a great community project headed by
@RickS with the Purifi 6.5” midwoofer, and SEAS DXT tweeter.
Hats off to
@amirm for showing these graphs for all us to view:
When
normalized to the on-axis response, as shown above, helps clarify
how well the speaker will respond to equalization.
(bold: my emphasis)
Another view is shown below, using the exact data offered by ASR, except not-normalized. Thank you
@pierre
(Despite the legend on the right Y axis starting from 0 and going down to -30 dB this only used internally to compare to OTHER speakers)
This shows speaker's actual directivity, in the horizontal plane, anechoically, without EQ.
Directivity is complex. If I had to put this into words, I would say it displays omni-polar characteristics from 100Hz to 200Hz; from about 400Hz to 700Hz about a -6dB beam-width of +/- 110°, and then from 800Hz to 5KHz +/-60°, narrowing to +/-50° around 7KHz and then +/- 40° at 18KHz.
If the reader is lost at this point, the computed average for the -6dB beamwidth is
+/-60° from 100Hz to 10KHz.
Reference:
https://www.spinorama.org/speakers/ASR Directiva Open Source Speaker/ASR/index_asr.html
For completeness, I will quickly show the distortion of the Directiva.
Looking at the tweeter- At 86dB, H2 from the tweeter are barely visible at 86dB.
At 96dB, H2 from the tweeter is around 40dB. H3 is skirting around 20dB.
Reference:
https://www.audiosciencereview.com/...r-directiva-open-source-speaker-review.27094/
We must remember that in Floyd’s ground breaking research, it was clear that how a speaker appears affected a listener’s assessment. In short, "
in sighted listening evaluations, what we "hear" is related to more than just the sounds arriving at the ear"
To control for these biases, these tests were double blinded. In addition, listening tests were done in mono, as listeners were found to be more discriminatory in mono listening tests, Vs stereo or multichannel.
Sean acknowledged some limitations of the study, which was mainly conducted in one good room. And although double blinded testing of 3 very good speakers in 3 differently shaped/sized rooms showed that they were performed similarly on listening tests, it must be noted that all 3 rooms had reverberation times of about 0.4s (+/- 0.1 sec) (an IEC room has RT 0.3sec)
He hypothesized that, for instance, in a very damped room, the direct sound may be more important to listener preferences. In a more reverberant room the predicted in-room response may have a higher weighting in indicating listener preferences
References:
The challenge for reviewers is very real-
1. It is impractical (?impossible) for reviewers to do unsighted listening.
2. How is their listening room different to your listening room? What is the reverberation time of their room, and what is the the reverberation time of your room?
3. If they are listening in mono (and thus more discriminatory), but when you
buy a pair in stereo, would this mean you are less discriminatory in your listening (ie. good speaker -> great speaker, or worse “I can't hear much difference to what I already have ). The evidence suggests that we, as human, may very well limited in our assessment in non-mono.
So to answer the question- is a speaker with a tweeter with a
wide and constant directivity of ~ +/- 70° degrees horizontally a benefit, (or hindrance) in a sighted, mono test?
Whatever benefit (or hindrance) of a stereo pair of speakers, that has ultra low distortion, and a wide constant -6dB beam-width of +/- 70° horizontally, has not yet been ascertained.
We may be in uncharted territory here.
Appendix/further reading-
More challengers for listeners and reviewers here:
