Hi DT,
wide coverage as novelty for any transducer makes sense on high frequency, where wavelength is smaller than the transducer. On long wavelengths it would be wide coverage naturally. On the other hand narrow coverage would work as novelty for the low bandwidth, because highs beam anyway. If one wants to make a constant directivity over bandwidth where the transducer size is contained then one needs narrowing and widening of the natural response of the transducer.
So, meaningful bandwidth that benefits for addressing this Direct / Reflected sound ratio must be in between, somewhere midrange. Now, where is the meaningful bandwidth? And how high of a D/R one would need for make meaningful change (to perceived sound)? And how perceived sound would differ and what does it mean?
I mean, D/R changes per frequency naturally no matter what the system is, due to wavelength and how much time hearing system has to process the information. Assuming D/R is meaningful at some bandwidth, then DI around this bandwidth means critical distance moves further away when DI is higher. Making "tweeter wide" is feature for very short wavelengths, which are likely outside this meaningful bandwidth. However, making midrange narrower would likely make a difference in this context.
I've tried to bring this stuff up recently when ever there is context, it's what Griesinger has been studying. Pardon using your post as another excuse D/R goes up as listener goes closer to speakers, somewhere close enough there is audible critical distance where perception of (stereo) sound seems to shift so that sound of local room mostly disappears, turns into envelopment. Having wide highs doesn't likely change this critical distance and thus does not have much to do with D/R per se but instead would simply mean toe-in doesn't have to be so exact which means there is bigger area in room where also the highs are audible. It would also affect highs in envelopment and spaciousness, amount of highs in room sound, so sound of highs would be different no matter what the D/R or how far speakers you listen at.
wide coverage as novelty for any transducer makes sense on high frequency, where wavelength is smaller than the transducer. On long wavelengths it would be wide coverage naturally. On the other hand narrow coverage would work as novelty for the low bandwidth, because highs beam anyway. If one wants to make a constant directivity over bandwidth where the transducer size is contained then one needs narrowing and widening of the natural response of the transducer.
Geddes pointed out recently, that at high frequencies wavelength is so short that brain has a lot of time to make sense of the sound before any reflections come in, so it is no problem to hear direct sound at high frequencies no matter of DI. Mind toe-in of course. On the other hand very low frequencies wavelength is so long it is not possible to distinguish direct sound from room sound. It's a sliding concept, point speakers at you and D/R slides from 0 to some higher number on the 20Hz to 20kHz bandwidth and assuming there is a meaningful difference in perceived sound due to D/R it must be somewhere between these extremes.
So, meaningful bandwidth that benefits for addressing this Direct / Reflected sound ratio must be in between, somewhere midrange. Now, where is the meaningful bandwidth? And how high of a D/R one would need for make meaningful change (to perceived sound)? And how perceived sound would differ and what does it mean?
I mean, D/R changes per frequency naturally no matter what the system is, due to wavelength and how much time hearing system has to process the information. Assuming D/R is meaningful at some bandwidth, then DI around this bandwidth means critical distance moves further away when DI is higher. Making "tweeter wide" is feature for very short wavelengths, which are likely outside this meaningful bandwidth. However, making midrange narrower would likely make a difference in this context.
I've tried to bring this stuff up recently when ever there is context, it's what Griesinger has been studying. Pardon using your post as another excuse
D/R goes up as listener goes closer to speakers, somewhere close enough there is audible critical distance where perception of (stereo) sound seems to shift so that sound of local room mostly disappears, turns into envelopment. Having wide highs doesn't likely change this critical distance and thus does not have much to do with D/R per se but instead would simply mean toe-in doesn't have to be so exact which means there is bigger area in room where also the highs are audible. It would also affect highs in envelopment and spaciousness, amount of highs in room sound, so sound of highs would be different no matter what the D/R or how far speakers you listen at.
Last edited:
^hmm just thinking about this stuff, wrote to another thread and must add something here to avoid confusion: Lows and highs would affect perceived sound through envelopment/spaciousness and frequency response as well. While writing the above, I was thinking about shift between kind of two different perceptions of stereo image, which seems to happen at some mid range bandwidth and seems to me to have most audible relation to D/R.
hi DT, it was a design target to enable very wide dispersion in the top octave. the wave guide and ‘coherer’ was designed specifically to the Spk16 box to get constant directivity and smooth off axis response. With the chosen box, the design target lead to the 70deg half beam width. The waveguide limits dispersion down towards the xover at 2k to match the woofer. The most violent directivity change is that all boxed speakers go from omni to 2pi somewhere in the lower midrange. The 70deg half beam is quite usual for the spk16 box size at least up to 5-8kHz. the unique new is that it is kept constant to 20kHz. My experience is that it gives more air and space to the sound. The most important is probably to have a smooth response in all directions so that the tonality of the reflected sound is not altered. Somehow the brain and ear has some clever tricks to separate the effect of the room from the imagined sound stage. All of is probably quite frequency dependent the wide dispersion in the top octave combined with flat on axis does not seem to make the sound bright as some would fear. The stereo image remains very stable when moving around in the room. Listening from a distance near the rear wall gave the best image despite having much reflected sound. Just my experience I don’t do psychoacoustic research 🧐
...The most violent directivity change is that all boxed speakers go from omni to 2pi somewhere in the lower midrange...
Will this tweet be appropriate for eventual cardiod with that crazy biggy PTT8 woofer to extend uniform directivity down to... I don't know... 200-300 Hz?
Just talking with myself...
Will this tweet be appropriate for eventual cardiod with that crazy biggy PTT8 woofer to extend uniform directivity down to... I don't know... 200-300 Hz?
Just talking with myself...
I am largely in agreement, 70% +, there is much more to know.
A smooth dispersion transition is key. I am not so sure about a dead level DI curve being required. Perhaps a a smoothly increasing DI curve above 8kHz is not a irritating thing. Large discontinuities or dog legs in the DI curve may be audible.
Like most, I listen to more than one set of speakers. I use a pair of 6.5 inch Purifi mini monitors on my bench.
The 6.5 inch Purifi mid bass drivers in a ~0.55 cubic foot sealed enclosure with still changing (auditioning) tweeters gets daily use on my bench (when I am home). I listen at direct/near field distance. A 12 inch Scan-Speak XXL sub with PR sits on the shelf under the bench. With the 6 1/2 Purifi mid-bass drivers these speakers have much lower measured distortion than any of the other crop of Mid-Bass speakers that I have measured.
The current tweeters are Fatial Pro HF108R 16 Ohm (13 Ohm resistors in parallel) with STH wave-guides/horns. At the output levels used, measured distortion is in the same league as the mid-bass drivers.
Thanks DT
Just for fun. -6dB coverage angles for different size piston drivers. If I recall this is from Altec Lansing of yore.
My PTT6.5M drivers have such low harmonic distortion that they are below the THD+N noise floor of my Behringer UMC202HD audio interface...
this is a plot @2.83 volts with AP Microphone, APX1701 amp and interface, and APx555 analyzer.
Hi Jim,
Yes agree. To properly characterise ultra low distortion drivers I have to turn off all the HVAC and measure when the environment noise is at its lowest. I also find it help to measure closer than 1m. Eg. 50cm for +6dB and 31.6cm for +10dB improved SNR respectively.
Of course your mic needs to be capable of measuring this SPL without becoming too distorted on its own so be careful of absolute levels.
A clipping microphone/audio input shows high levels of distortion for ALL harmonics at the frequency that it clips.
One then has to wonder; when’s the best time to listen to music?
When you’ve expunged everyone and everything from the household… 😅🤪
Last edited:
A better audio interface is on my shopping list... my mic is 4x the cost of the audio interface, and my recent project used a set of drivers was 25x the cost of the audio interface (L + R). So it is obviously the weak link and obviously out of place.
But the little Behringer has served me well for the essential speaker measurements I need... It was well worth the $100 price.
j.
But the little Behringer has served me well for the essential speaker measurements I need... It was well worth the $100 price.
j.
hifijim,
One thing to do to measure driver THD+N is to put a 0.1 Ohm resistor in series with the driver voice coil and sample across the series resistor. This picks up amplitude modulation do to voice coil excursion, surround modulation Sd modulation and other distortions that originate due to the driver.
This measurement shows up in the PURIFI data sheet, it is labeled current distortion. The noise floor of this measurement will be the noise floor of the test amplifier. Microphone and microphone preamplifier is excluded. Because the resistor value is low the measurement will need to be shown on a relative (dBr) scale, with some relative dB's added to look equivalent to the microphone measurements.
https://www.mouser.com/ProductDetail/Caddock/MP915-0.10-1?qs=YqNA2qefETD/pfz9BBgABg==
Thanks DT
One thing to do to measure driver THD+N is to put a 0.1 Ohm resistor in series with the driver voice coil and sample across the series resistor. This picks up amplitude modulation do to voice coil excursion, surround modulation Sd modulation and other distortions that originate due to the driver.
This measurement shows up in the PURIFI data sheet, it is labeled current distortion. The noise floor of this measurement will be the noise floor of the test amplifier. Microphone and microphone preamplifier is excluded. Because the resistor value is low the measurement will need to be shown on a relative (dBr) scale, with some relative dB's added to look equivalent to the microphone measurements.
https://www.mouser.com/ProductDetail/Caddock/MP915-0.10-1?qs=YqNA2qefETD/pfz9BBgABg==
Thanks DT
Current distortion will NOT pick up surround SD modulation.
It's mostly just the magnet system, some CMS(x) and BL(x) interaction in the lower range, but mostly hysteresis distortion in mid-high frequency. It barely picked up any cone/surround resonances, and those are major contributor to distortion, too.
It's mostly just the magnet system, some CMS(x) and BL(x) interaction in the lower range, but mostly hysteresis distortion in mid-high frequency. It barely picked up any cone/surround resonances, and those are major contributor to distortion, too.
Hello,
You could start a thread here on https://www.diyaudio.com/community/forums/multi-way.6/
Or even post a test plot right here in this thread and if wanted a Moderator could move it somewhere else if they want or need to.
I would like to see your results.
Thanks DT
Looking at the attached photos of the SP16 and SB26ADC-C00-4 tweeters.
The circular disks that cover the center of the tweeters look to be similar.
Do they have similar function? Do they reduce beaming of the tweeters at increasing frequency?
It is my impression that they do.
Thanks DT
In the case of the SB26 tweeter, that is a phase shield. It delays the sound wave from the center of the dome just slightly so that it matches the sound wave from the perimeter. The center is a few mm closer to you than the perimeter, and at 16k - 20k, there can be some cancellation. The phase shield corrects this.
In the case of the Purifi tweeter, I would not presume to guess what is going on. It could be a phase shield, but it could also be a more sophisticated wave guide process...
In the case of the Purifi tweeter, I would not presume to guess what is going on. It could be a phase shield, but it could also be a more sophisticated wave guide process...
@DualTriode
On the SB it’s a phase shield. It functions as @hifijim proposed. Hard domes are pistonic by nature. The second function is to protect against dents, I suppose. A soft dome, if dented, can be massaged out, with little impact on the function of the tweeter’s frequency response. (The breakup in the soft dome was chaotic prior to being dented). The micro-acoustics expert René @ Acculution could explain it better then I ever could.
On the Purifi what you’re seeing, and if we are looking at the same thing, well that’s also a phase shield. But the entire geometry around the dome’s periphery is a waveguide. Its highly optimised geometry + the baffle geometry is responsible for the constant directivity with a beam width of 70 degrees. This overcomes one of the limitations of traditional waveguides- controlled, yet limiting, dispersion
So altogether I think Purifi have called it their Coherer. It’s a nice name and I wonder if it’s patentable.
Is it available for pre-order yet @lrisbo ?
Pretty please?
On the SB it’s a phase shield. It functions as @hifijim proposed. Hard domes are pistonic by nature. The second function is to protect against dents, I suppose. A soft dome, if dented, can be massaged out, with little impact on the function of the tweeter’s frequency response. (The breakup in the soft dome was chaotic prior to being dented). The micro-acoustics expert René @ Acculution could explain it better then I ever could.
On the Purifi what you’re seeing, and if we are looking at the same thing, well that’s also a phase shield. But the entire geometry around the dome’s periphery is a waveguide. Its highly optimised geometry + the baffle geometry is responsible for the constant directivity with a beam width of 70 degrees. This overcomes one of the limitations of traditional waveguides- controlled, yet limiting, dispersion
So altogether I think Purifi have called it their Coherer. It’s a nice name and I wonder if it’s patentable.
Is it available for pre-order yet @lrisbo ?
Pretty please?
Last edited:
It is possible have a 70 degree beamwidth out to 20K without coherer's or phase shield's. Simple is good sometimes.
https://www.diyaudio.com/community/...ifi-woofer-speaker-builds.352063/post-7356561