Why would a normalized graph be more complicated?
Btw, the suggested EQ for that speaker is kinda weird.
EQ'ing dips is rarely a good suggestion, and in this example a couple of those are clearly from diffraction problems.
Clearly visible in the power response.
Which now has to much energy at certain frequencies (it's not that horrible, I have seen far worse).
Looking at the first graph, there seem to be als quite a bit of room reflections (or similar) left in the measurements.
There is no reason that driver does these kind of scribbles by itself.
Upsetting that these days people think everything can be EQ'ed without thinking.
Worst I have seen, is the suggestion to boost well over 6dB to "get rid" of a certain dip. Not knowing that the dip in question was actually a destructive resonance from the port. Meaning this will always be a dip no matter how much you boost that particular frequency (acoustics 101).
So in the end it's nothing more than a artifact of the representation of the measurements why it looks better after EQ'ing.
Really bad that such kinda of terrible misinformation is just being shared left and right, especially by people that clearly have a certain status in the community.
Big thumbs down. 👎
No, the other way.
The non-normalised graph was purported to be:
And that normalised was easier to view.
Yes it IS easier to view, because it’s all related to the reference axis. Assuming a flat on-axis was a good idea to start with. And sometimes it’s not. Eg. EQ for flat!
Exactly.
EQ it all and it’s all better, according to ASR…
Final time, to be clear, EQ for flat on axis or listening window is NOT always a good idea. And that is why a non-normalised graph can be useful.
If you don’t like looking at non-normalised graphs, move along, nothing to look at or learn here.
If this doesn’t help you see the difference in vertical response of the KEF Blade 2 Meta and R11 Meta (both with 4 6.5” woofers) one which is side facing and the other which is one the front panel, please look at spinorama.org
The non-normalised graph was purported to be:
And that normalised was easier to view.
Yes it IS easier to view, because it’s all related to the reference axis. Assuming a flat on-axis was a good idea to start with. And sometimes it’s not. Eg. EQ for flat!
Exactly.
EQ it all and it’s all better, according to ASR…
Final time, to be clear, EQ for flat on axis or listening window is NOT always a good idea. And that is why a non-normalised graph can be useful.
If you don’t like looking at non-normalised graphs, move along, nothing to look at or learn here.
If this doesn’t help you see the difference in vertical response of the KEF Blade 2 Meta and R11 Meta (both with 4 6.5” woofers) one which is side facing and the other which is one the front panel, please look at spinorama.org
Last edited:
It's not just easier, there are things that are simply not visible on a non-normalized graph.
(And vice versa obviously)
So much for the science part in that name. 👎
The irony....
I’ve been preaching this for decades coming from the engineering side of things……..I’ve bought, built and sold too many ultra flat monitors to count and I wouldn’t listen to ANY OF THEM if I were listening for pleasure. Just like the NS10 phenomena, I’ve come to know and understand the Q150’s off axis response over the past 2 years so they’re staying on my bridge……I get to make critical decisions on mid side mixes without cringing while the point source reveals the phase information and preserves the detail often lost on conventional mid tweeter two ways.
Listen with your ears……not your eyes……and…..we ARE NOT all the same as listeners which has nothing to do with experience over time and critical listeners……music moves the soul. Remove the soul and all that’s left is a boring clinical mess of extended and chaotic test tones. We’re not wired like that……win-stay-lose-shift still applies despite the ruling classes morbid attempts to extinguish it and homogenize everything……critical thinkers take note…….it’s not what you think, it’s how.
The difference between visual perceptions of test tones and complex music are exponential to say the least.
You basically mean, don't listen with ignorance?
Because my eyes see a lot of things that just would never sound well.
But in the end I do agree. The proof is in eating the pudding.
Some people just enjoy a certain sound or tone, and there is also nothing wrong with that either.
That's right, the answer lies between the two.
Here is how I do it, with Dr Geddes in agreement - https://www.diyaudio.com/community/threads/geddes-on-waveguides.103872/post-3389842
SUMMARY:
Here are HF devices that exhibit smooth directivity
1. Compression driver in well optimized large horn design by @mabat
2. The same compression driver with an off-the-shelf horn- notice how it is slightly less smooth between 3-10K Hz
Dome drivers in waveguides:
3. Hard dome tweeter with DXT® lens, with large round over starting at edge of tweeter, crossed to a 5" midrange @2K Hz:
3. soft dome in a Sonoma Sonus waveguide designed by @augerpro
4. Hard dome in a waveguide
5. Same hard dome in a WG optimized for a 9.5" wide cabinet, and for use down to 1K Hz:
Dome tweeter loaded by cone drivers ie. coaxial drivers
7. A 1" dome tweeter at the centre of a 5" midrange:
8. Same driver as above, in a 21cm cabinet with 1" round-over: WIDER
Here are HF devices that exhibit smooth directivity
Compression drivers and horns
1. Compression driver in well optimized large horn design by @mabat
2. The same compression driver with an off-the-shelf horn- notice how it is slightly less smooth between 3-10K Hz
Dome drivers in waveguides:
3. soft dome in a Sonoma Sonus waveguide designed by @augerpro
NB. Taken on quasi-IB; data only available up to 60 degrees. On a real cabinet there will be some differences.
4. Hard dome in a waveguide
NB. Taken on quasi-IB; data only available up to 60 degrees. On a real cabinet there will be some differences.
5. Same hard dome in a WG optimized for a 9.5" wide cabinet, and for use down to 1K Hz:
Dome tweeter loaded by cone drivers ie. coaxial drivers
8. Same driver as above, in a 21cm cabinet with 1" round-over: WIDER
Reference:
1: Tymphany DFM-2535R00-08 in ATH A460
link: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7804444
2. Tymphany DFM-2535R00-08 in SB Audience H225
link: https://www.diyaudio.com/community/...0x60-degree-constant-directivity-horn.405050/
3. Kii Three, as measured by @bikinpunk
Measurement from: https://www.erinsaudiocorner.com/loudspeakers/kii_three/
4 .SB Acoustics SB26ADC/CDC on 8" WG on quasi-IEC baffle by augerpro
Link: https://www.somasonus.net/sb-acoustics-sb26
5. SB Acoustics SB29SDAC on 8" WG on quasi-IEC baffle by augerpro
Link: https://www.somasonus.net/sb-acoustics-sb29sdac
6. SB Acoustics SB26ADC in 8.5" WG on 240cm cabinet designed by @fluid
7. KEF R3, measured by @bikinpunk
Measurement from https://www.erinsaudiocorner.com/loudspeakers/kef_r3/
8. KEF R3's coaxial driver inserted in 21cm wide cabinet with 1" round over
Reference: Speaker 2
MORE to COME...
1: Tymphany DFM-2535R00-08 in ATH A460
link: https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-7804444
2. Tymphany DFM-2535R00-08 in SB Audience H225
link: https://www.diyaudio.com/community/...0x60-degree-constant-directivity-horn.405050/
3. Kii Three, as measured by @bikinpunk
Measurement from: https://www.erinsaudiocorner.com/loudspeakers/kii_three/
4 .SB Acoustics SB26ADC/CDC on 8" WG on quasi-IEC baffle by augerpro
Link: https://www.somasonus.net/sb-acoustics-sb26
5. SB Acoustics SB29SDAC on 8" WG on quasi-IEC baffle by augerpro
Link: https://www.somasonus.net/sb-acoustics-sb29sdac
6. SB Acoustics SB26ADC in 8.5" WG on 240cm cabinet designed by @fluid
7. KEF R3, measured by @bikinpunk
Measurement from https://www.erinsaudiocorner.com/loudspeakers/kef_r3/
8. KEF R3's coaxial driver inserted in 21cm wide cabinet with 1" round over
Reference: Speaker 2
MORE to COME...
Last edited:
9. 33mm hard dome tweeter with optimized dome & waveguide for smooth and wide +/- 70° (-6dB) beam-width all the way to past 20KHz:
Now for the ribbon tweeters-
A. Fountek NeoCD3.5H (a horn-loaded NeoCD3) in a cabinet 24×35cm with 18mm round-over, not flush-mount, center-line with driver unit center 8cm from the top of the baffle, with acoustic LR4 3KHz filter applied:
B. Fountek NeoCD3.0 in 11" wide cabinet, with filter for 3rd order Butterworth crossover at 3700Hz:
Reference: Statement II measurements, with filter simulation applied
Reference: traced and XO simulation applied, courtesy @dcibel
Could this be why Zaph liked this ribbon speaker, all those years ago?
D. RAAL ribbon in Ascent Acoustics Luna V2 speaker
compare with the 26mm SEAS alloy dome tweeter version:
Does this wide, constant directivity frequency response go some way into explaining why some enjoy/prefer ribbon tweeters?
In the past people have often complained about the slightly higher harmonic distortion performance. Or stated that they complained about the high vertical directivity (standing up/laying down allegedly affects the tonal balance). This author wonders whether the wide constant directly gives a "soft, spacious sound" in spite of the reduced floor/ceiling reflection
Teleological thinking?
Or perhaps directivity measurements helping us better understand how something sounds?
A. Fountek NeoCD3.5H (a horn-loaded NeoCD3) in a cabinet 24×35cm with 18mm round-over, not flush-mount, center-line with driver unit center 8cm from the top of the baffle, with acoustic LR4 3KHz filter applied:
Reference: Gornir's measurement
B. Fountek NeoCD3.0 in 11" wide cabinet, with filter for 3rd order Butterworth crossover at 3700Hz:
Reference: Statement II measurements, with filter simulation applied
C. Fountek NeoCD3.0 in a slimmer rectangular cabinet 8.75" wide (x 16 5/16" H; round-over 10mm, vertically centered, 3.75" from top edge) with acoustic LR2 filter @ 2.9KHz.Reference: traced and XO simulation applied, courtesy @dcibel
Could this be why Zaph liked this ribbon speaker, all those years ago?
D. RAAL ribbon in Ascent Acoustics Luna V2 speaker
Reference: Spinorama.org
compare with the 26mm SEAS alloy dome tweeter version:
Reference: Spinorama.org
Does this wide, constant directivity frequency response go some way into explaining why some enjoy/prefer ribbon tweeters?
In the past people have often complained about the slightly higher harmonic distortion performance. Or stated that they complained about the high vertical directivity (standing up/laying down allegedly affects the tonal balance). This author wonders whether the wide constant directly gives a "soft, spacious sound" in spite of the reduced floor/ceiling reflection
Teleological thinking?
Or perhaps directivity measurements helping us better understand how something sounds?
On the topic of "Do different cone materials sound different?"
Many people will exclaim- "well of course!"
Some will say "You can hear it, but you can't measure it"
Let me show you one counter-example.
One cabinet- 2 different drivers with common magnet system but different cone
1. Fiber cone-
2. Aluminum cone:
Here you will say... well, of course they will sound different.
But what it we carefully apply a crossover so that their response is the same? Let's suppose we have a high tech crossover that allows us to match the low pass filter such that both have a response that perfectly follows an acoustic LR4 with a cut off of 2.5KHz:
The fibre cone:
The aluminum cone:
At this point, you might be inclined to say- that looks very similar.
POLAR MAPS:
Fibre cone:
Aluminum cone:
As we can see, how one cone radiates sound in three dimensions is different another cone, irrespective of our perfectly tuned crossover.
Many people will exclaim- "well of course!"
Some will say "You can hear it, but you can't measure it"
Let me show you one counter-example.
One cabinet- 2 different drivers with common magnet system but different cone
1. Fiber cone-
2. Aluminum cone:
Here you will say... well, of course they will sound different.
But what it we carefully apply a crossover so that their response is the same? Let's suppose we have a high tech crossover that allows us to match the low pass filter such that both have a response that perfectly follows an acoustic LR4 with a cut off of 2.5KHz:
The fibre cone:
The aluminum cone:
At this point, you might be inclined to say- that looks very similar.
POLAR MAPS:
Fibre cone:
Aluminum cone:
As we can see, how one cone radiates sound in three dimensions is different another cone, irrespective of our perfectly tuned crossover.
So at the end of the day -as always- it is all SPL related, albeit off axis rules.
Yet such differences are highly system/appplication dependant, and thus i.m.o. not totally dependant on one (out of three e.g.) driver behaviour in a multiway system.
Furthermore the differences shown are relatively small; these are not the day and night differences claimed here in a number of threads, when advocating driver Brand X over driver Brand Y.
That being said: thanx a million for showing us all this and collecting all these data.
Yet such differences are highly system/appplication dependant, and thus i.m.o. not totally dependant on one (out of three e.g.) driver behaviour in a multiway system.
Furthermore the differences shown are relatively small; these are not the day and night differences claimed here in a number of threads, when advocating driver Brand X over driver Brand Y.
That being said: thanx a million for showing us all this and collecting all these data.
Got the blade 2 myself never considered the change in the directivity with eq actually I got 2 presets and the one equalized till 500hz sound better , I may need even lower because the speakers are already linear
I take strong exception to this position. Normalization removes a lot of flaws in a measurement. It makes sense to normalize a display before one has EQ'd the final system, but for a completed system, unnormalized is the only real view into how well a system works.
The correct answer is that we need both.
So saying that unnormalized "is the real view" is just not explaining the entire story and oversimplifying everything.
It most certainly isn't the correct answer either.
We indeed need unnormalized graphs to see all the "flaws" you're talking about and we need normalized graphs to understand the proper relative difference.
There's no way that you otherwise can estimate how well the directivity behaves with a unnormalized graph.
This has to do with the fact that directivity is a relative parameter (compared to some kind of choose reference axis), not an absolute parameter.
Btw, in the quoted text I was very clearly talking about directivity, NOT any other issues, "flaws" or other problems.
So taking "strong exception" is kinda out of context here, because I never said unnormalized graphs are useless.
It's just apples vs pears, because we use both graphs on very different ways for very different purposes.
In my opinion, thinking that one view or another is "better" is quite a dumb idea to begin with. (even more so to discuss, nitpick and be pedantic about it)
All views and datasets have their advantages and disadvantages.
By definition you'll loose any irregularities from absolute data in a graph that represents it on a relative way.
I think that's pretty obvious.
Claiming (therefor) that one cone material here radiates differently is oversimplifying everything again.
There are so many more parameters to a driver than just only the cone material.
Since the materials have different inherent properties, they therefor very often are physically different.
I have been involved in quite some custom drivers, and from experience I can tell that it's rare that you'll get the same cone shape/size/other physically properties when changing just the material.
In most cases even the spider (compliance) and surround shape and contact points already change, since different materials have different masses and stiffness's.
For identical T/S paramters, other properties therefor have to change as well.
In your particular example there is a VERY clear difference in the dustcap (convex vs concave), which has a major contribution in directivity and off-axis behavior!
So in my opinion the claim and conclusion that therefor different materials measure different, cannot be concluded from this experiment.
Maybe it's true, but we can't say from these particular measurements.
If I just tap things on my desk or listen to sounds in the course of life it's obvious that they do not hurt my ears the same way a flat on-axis response does. When I do an actual comparison it's clear nothing sounds like that at all, even if it is technically flat. Most microphones seem to have a peaking response, most producers seem to prefer a shouty response, and hearing loss is the giant wrinkly pink elephant in the room.
I always seem to prefer a somewhat downward sloping response, flat but sloped. Otherwise I will gladly live in silence. Others don't find this to be an issue when they listen to the speakers I've tuned. Drivers with exceptionally clean sound get away with the shouty sound better though. Whether that is related to directivity, distortion or something else I do not know.
I always seem to prefer a somewhat downward sloping response, flat but sloped. Otherwise I will gladly live in silence. Others don't find this to be an issue when they listen to the speakers I've tuned. Drivers with exceptionally clean sound get away with the shouty sound better though. Whether that is related to directivity, distortion or something else I do not know.
This just adds to the original point rather than taking away from it.
People like claiming how certain materials sound when clearly there's more to it than just the cone material.
Obviously the dust cap differences will affect the off axis and, as per the measurements above, could, very well, account for why someone prefers the paper to the metal. But they would then attribute this preference to the cone material when in fact it's something completely different.