• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

step response is a important part to show speed of speaker that is good enough for ITD. See measures

Status
Not open for further replies.
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
And the wavelet:
View attachment 175699
With 1/24 frequency smoothing, because I'm no 'smooth criminal'.

with your resolution upto 100 ms can not good see mid speed. on hearing position( around 50 cm away from speaker) with 1/24 frequency i get only worse, is room influenece.

1_24 smooth.jpg
 

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
Sorry, check my big thread over on diyaudio for many many more IR/STEP results, but the answer you seek isn't found there.
The IR/STEP is just a presentation of a way more complex bundle of information, the real pulse.
That's why there are so many tools available to be able to see what we want to see. The waterfall, a frequency plot, the corresponding phase plot, the wavelets, the ability to dissect the IR on a (filtered) frequency level etc.

Every bump and dip has it's effect on how the STEP and IR looks, just look at that early waterfall of yours to know it will present trouble for the STEP as well. The first milliseconds of it mainly consists of the high frequency range anyway.

So use the tools, that's where they are there for. That's where you can find the answers with regards to timing, fast decay etc.
 

kyle_neuron

Active Member
Joined
Jun 18, 2021
Messages
149
Likes
254
with your resolution upto 100 ms can not good see mid speed. on hearing position( around 50 cm away from speaker) with 1/24 frequency i get only worse, is room influenece.

View attachment 175747
I already showed the first 10 ms of response for the same speaker via the waterfall plot, in the very same post.

Your plot clearly shows poor time alignment, and behaviour that is typical for measurement in the near field of the speaker. All of this is well documented in literature, if you'd care to read some of the very good book or paper recommendations that people have graciously given to you.

A comprehensive literature review is well over 50% of a good research project...

The room response will dominate any measurement below the Schroeder frequency of the space.

The ‘speed’ of the system step response is entirely due to the resonant frequency of the cutoff for the passband.
 
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
I already showed the first 10 ms of response for the same speaker via the waterfall plot, in the very same post.

The waterfall plot too goes upto 100 ms. then not good speed can see and of course it look nicer.

Your plot clearly shows poor time alignment, and behaviour that is typical for measurement in the near field of the speaker. All of this is well documented in literature, if you'd care to read some of the very good book or paper recommendations that people have graciously given to you.

It was at hearing position. for me it is important what happen on hearing position. I play sine tones from 100 hz upto and hear if they sound strange or are sound out of mid. so i choose 50 cm. but still not good. when i hear music at very low volume around 60 db then it sound muddy and i get a strange feeling on my left ear. on headphones i have this mudd in the freq area around 100-300 hz never hear on low volume. sure this arte room modes. but with more precise mid speakers the brain can correct better. with the kali i need volume 90 db db(C) that it sound good for me. which is not good when hear longer. with the JBL i am happy with 75 db(C) with the sound, but a little mud or better named as less transients in compare to headphones i can still hear in the 100-300 hz range. higher freq sound on speaker and headphone same clarity i think. both speakers are measure and correctet so its not FR. The kali halso have larger directivity as my other speakers. this confirm that larger directivity does not help. only rule i can see. faster raise and decay time in step response give better stereo width and depth and less strange feeling in ear on left side on low volumes. leftn are more room modes i measure too.


The ‘speed’ of the system step response is entirely due to the resonant frequency of the cutoff for the passband.

The rise time of the step response is from cutoff. but it is usefull to see this because steep filter sound worse for ITD. the single-cycle interaural phase differences (IPDs) thats need that the brain can create width and depth get wrong because the phase change alot by the steep lowpass. on stereo signal never play on both sides same cycle period time.

you can compare stereo singnals with video on a slow TFT monitor. not a very good compare but maybe explain better whats happen. when the video have slow or no movement(the movement is diffrence between left and right speaker), all look good. but when have movement edges look blur. and the blur in video is smaller stereo width in audio. you can phase correct that it look on measure good. but this give in real life a simular effect as the overdrive on TFT display. it is better but not perfect. perfect it can get only with fast mid range speakers
 
Last edited:

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
index.php

Benny, I thought we were getting somewhere. This is Kyle's waterfall. See the time scale on the upper left? 0 to 10 ms.

Now let's see yours:
index.php

Same time scale. Again, I'll ask the question: which one is faster?

Maybe you need some help reading a waterfall plot. Do you see the big bump at ~1.8 KHz in your plot? The one that looks cut off? That's because it truly is cut off, because we activated that "Use CSD mode" radiobutton. It is ringing there and will keep on ringing due to the uneven response your results show at that frequency. Kyle's result show nothing like that, they are at least 20 dB down within a 5 ms time frame. Pretty fast in my book. Faster than your current results. In other words: there's room for improvement on your side.
Why not focus on getting that improved result instead of making up stories about STEP results? It would be time better spend. If what Don said is right, and you've spend the better part of half a year on this: making up false claims about STEP responses, you could have spend all of that time trying to improve your own results. I promise you everything will look better if you first make sure that frequency plot we see in this very early time frame is way more flat. That means looking critically at how your speakers are placed within your room, your microphone as well (away from boundaries) and making sure you measure at a distance your speakers are meant to use.

Don't make up the rules... they are fixed. So everything you come up with as an explanation that suits your "case" should follow those rules.
That goes for us all.... Except for the media guys for some big Audiophile companies.... they can make up anything they want.
You don't want to belong to that crowd, do you?
So just take to heart the rules of the game, stick to those rules and improve your result using them...
That will get you ahead in this game. ;)

By the way, Happy New Year to all....

P.S. Everything I've shown here was recorded at my listening spot. At 2.7 meter from the array, or when it's mentioned, from
both arrays in the exact sweet spot. See a plan view of my room here:
Room-a-small.jpg

As Kyle mentioned to you: perhaps you're measuring too close to really see what's going on. The speaker(s) often need a certain distance to make sure all signals arrive with the proper timing. That would depend on how they were designed to be used.

My speaker's result would look horrible at 50 cm, even though you'd see less of the room in the results. They were meant to be used at 2.7 meter and show good results from about 2.5 meter up to 4 meter. That's what I designed them to do.

If your speaker needs more room to properly integrate all the drivers at the measurement spot, move them out into the middle of the room and position the microphone as far from all walls as well. That may mean you need to elevate the speaker and microphone to get the best results while making sure all of the room's surfaces are as far from the speaker as possible.
Even better: move them outside and high up in the air with the mic up too, but that's not possible for all of us I guess. I know it can't be done around my house. Too many city noises that would interfere with the results. And possibly upsetting my neighbors :D.
 
Last edited:
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
index.php

Benny, I thought we were getting somewhere. This is Kyle's waterfall. See the time scale on the upper left? 0 to 10 ms.

Now let's see yours:
index.php

Same time scale. Again, I'll ask the question: which one is faster?

Maybe you need some help reading a waterfall plot. Do you see the big bump at ~1.8 KHz in your plot? The one that looks cut off? That's because it truly is cut off, because we activated that "Use CSD mode" radiobutton. It is ringing there and will keep on ringing due to the uneven response your results show at that frequency. Kyle's result show nothing like that, they are at least 20 dB down within a 5 ms time frame. Pretty fast in my book. Faster than your current results. In other words: there's room for improvement on your side.
Why not focus on getting that improved result instead of making up stories about STEP results? It would be time better spend. If what Don said is right, and you've spend the better part of half a year on this: making up false claims about STEP responses, you could have spend all of that time trying to improve your own results. I promise you everything will look better if you first make sure that frequency plot we see in this very early time frame is way more flat. That means looking critically at how your speakers are placed within your room, your microphone as well (away from boundaries) and making sure you measure at a distance your speakers are meant to use.

Don't make up the rules... they are fixed. So everything you come up with as an explanation that suits your "case" should follow those rules.
That goes for us all.... Except for the media guys for some big Audiophile companies.... they can make up anything they want.
You don't want to belong to that crowd, do you?
So just take to heart the rules of the game, stick to those rules and improve your result using them...
That will get you ahead in this game. ;)

By the way, Happy New Year to all....

P.S. Everything I've shown here was recorded at my listening spot. At 2.7 meter from the array, or when it's mentioned, from
both arrays in the exact sweet spot. See a plan view of my room here:
View attachment 175914
As Kyle mentioned to you: perhaps you're measuring too close to really see what's going on. The speaker(s) often need a certain distance to make sure all signals arrive with the proper timing. That would depend on how they were designed to be used.

My speaker's result would look horrible at 50 cm, even though you'd see less of the room in the results. They were meant to be used at 2.7 meter and show good results from about 2.5 meter up to 4 meter. That's what I designed them to do.

If your speaker needs more room to properly integrate all the drivers at the measurement spot, move them out into the middle of the room and position the microphone as far from all walls as well. That may mean you need to elevate the speaker and microphone to get the best results while making sure all of the room's surfaces are as far from the speaker as possible.
Even better: move them outside and high up in the air with the mic up too, but that's not possible for all of us I guess. I know it can't be done around my house. Too many city noises that would interfere with the results. And possibly upsetting my neighbors :D.

Happy new year .My measure is very near from the JBL. it is a coaxial speaker but when measure very near it give strange FR and waterfall. so 8 db or more diffrence. in FR. but the step response rise and fall time stay near same. that the FR with the coaxial speaker change so much but the step response stay near same is another fact that show that frequency influence of step response is very small. i can also do measures on diffrent positions distance with room modes and much diffrent bass. step response stay near same.

the picture was very small the 10.0 i read as 100(I have a 4k monitor and screenshots of full hd look very small). wy you change to waterfall and do not use wavelets spectrum ?. this show speed better.
seem he have set the windowing to 4 ms and he test the speaker with large distance to next wall. because after 4 ms see nothing more on his plot. I have my speakers stand on desktop 40-50 cm to table plate. but spectrum wavelets have this problem not and also near no influence of step response rise and fall time because this times are in the first 800 µsec if speaker is not slow. so need no windowing, my measures you need windowing with 800 µsec or so that it give a more realistic looking waterfall.
 
Last edited:

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
Happy new year .My measure is very near from the JBL. it is a coaxial speaker but when measure very near it give strange FR and waterfall. so 8 db or more diffrence. in FR. but the step response rise and fall time stay near same. that the FR with the coaxial speaker change so much but the step response stay near same is another fact that show that frequency influence of step response is very small. i can also do measures on diffrent positions distance with room modes and much diffrent bass. step response stay near same.
True... it is the (very) high frequency information that dominates the STEP. If you want to see more, like the all important midrange, you'll have to dissect that information locked inside that impulse.
the picture was very small the 10.0 i read as 100(I have a 4k monitor and screenshots of full hd look very small). wy you change to waterfall and do not use wavelets spectrum ?. this show speed better.
Quite frankly, because that wavelet doesn't show nearly as much detail of that first wave front.
One can make that wavelet look very pretty and make it look like the performance is quite good, just by changing a few settings.
Actually, the way I showed your measurements in a wavelet was with quite mild (smoothing) settings. I was being "kind" to your results.
With the same settings Kyle used, which were like the settings used by me with that wavelet of my speakers with both arrays firing, are way more detailed already. But not quite as detailed as the waterfall shown after that.
seem he have set the windowing to 4 ms and he test the speaker with large distance to next wall. because after 4 ms see nothing more on his plot. I have my speakers stand on desktop 40-50 cm to table plate. but spectrum wavelets have this problem not and also near no influence of step response rise and fall time because this times are in the first 800 µsec if speaker is not slow. so need no windowing, my measures you need windowing with 800 µsec or so that it give a more realistic looking waterfall.
He used the settings that I had shown, it is cut short by using the "Use CSD Mode". The rate at which the energy drops (speed) varies with frequency. Look at the ideal waterfall I showed earlier, you just can't be any faster than that. Impossible.
idealwaterfall.jpg


That above is the "perfect ideal". At these settings, mind you... we can alter the settings to make everything pretty, or even fast.
But will that make our results any better? Will it sound better? I simply use settings that I have found that relate to what I can perceive.
I can dig even deeper, create even more detailed results. I can make the graphs look better but if it doesn't relate to what i hear, what good will it do me?

So I stick with what I know and find the clues of what works and what doesn't. One conclusion: the room is one of the biggest influences one can find as a reason as to why things sound like they do. One of my missions: make that room work with me, not against me.

Look at Kyle's waterfall. Quite controlled looking graph and pretty fast and clean (remember the decay within a few ms compared to your results?). Why and how? Waveguides and horns help to lessen the influence of the room.

My results are "helped" with DSP, but not just to make the graphs pretty. But to make it sound good even when I move up, down or left and right at that listening spot. That itself is a puzzle to work out within a room, but with dedication it can be done. Treatment of the room, DSP to work out the final details and it isn't done for show. It is done because it does sound better! Change one thing at a time and one thing only. Evaluate and compare. Live with it for a while, is it really better or just "new to you". "New" often gets mistaken for better, but can be proven not to work long term.
I know I've made a lot of judgement errors myself. Hopefully I've learned enough from them and over time evolve and be able, over a long time, to learn from (all?) those mistakes and most important: keep an open mind to learn even more. Can I be wrong? Yes I can. Do I know it all? No I don't.

Just saying that with proper dedication and an open mind, we can grow and learn and achieve ever improving results. Not without making mistakes, mind you. I've learned way more from my mistakes than bragging about results achieved. :D Have I done the latter? Sure, I'd be just as guilty of that as the next guy. But the thing that interest me most is what we can learn from one another. That is the true reason for me to take the time and type these messages. You might learn something, but so will I...
 
Last edited:
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
True... it is the (very) high frequency information that dominates the STEP. If you want to see more, like the all important midrange, you'll have to dissect that information locked inside that impulse.

I dont know how see this in the impulse. In step response the step rise is around 500 µsec of the woofer/mid of the slow kali lp6(see my intro post in this thread.) 1 half period of 1 khz is 500 µsec. so this mean more than 1 khz bring trouble to the speaker because it can not go faster but it should because crossover is 1.5 khz. so the 1.5 khz filter do a rise time of 333 µsec. the additonal 167 µsec i assume is worse for sound.

and in fact look at the distortion of Kali LP6 in this test. https://www.audiosciencereview.com/forum/index.php?threads/kali-lp-6-review-studio-monitor.17978/ It increase above 1 khz alot upto 2%

a high line array have diffrent distance to ears. I not hear any how it sound in stereowidth and depth. but 1 feet or 0.3 meter is a delay of 1 ms. thats alot. maybe best for a non effect sound is in line array use only 1 mid /high freq system and use LP for the others below 100 hz

maybe you can upload a wav of your impulse record and i and other can hear in headphone how it sound with a convolution response VST. I guess a high line array have a own sound that can be sound full and nice. because every speaker distance to ear diffrence from the nearest is simular as a delay.

He used the settings that I had shown, it is cut short by using the "Use CSD Mode". The rate at which the energy drops (speed) varies with frequency. Look at the ideal waterfall I showed earlier, you just can't be any faster than that. Impossible.
View attachment 175955

It seem this speaker is measure freefield in the garden or so that no room reflections happen. a waterfall with headphone look also not so good because of reflections or so, without a special headphone measure head. see headphone waterfall. step response show rise times too. waterfall show only decy time and this make trouble with room reflections. so step response look at rise time does not make problems with room reflections when they only occur after 800 µsec


I also measure not loud. only around 70 db or so. maybe when measure at 90 db look better. but i dont want make measure tones so loud.


headphone waterfall.jpg
 
Last edited:

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
So you're still jumping to the same conclusions you had when you started this thread, based on speed and STEP responses? That's a bummer.
I highly suggest to start playing with some virtual speaker simulations and tools like RePhase and/or VituixCAD, make crossovers, see what the STEP looks like etc. With a frequency response like your measurement showed, I wouldn't dare jump to any conclusion about anything when seeing the STEP response. Heck, by itself it's only marginally interesting to look at when it's made in an anechoic room. Granted one would know more info about crossover topology used etc.
If you keep jumping back to the same conclusions, there's only one conclusion left for me. Exit this thread trough the back door.

maybe you can upload a wav of your impulse record and i and other can hear in headphone how it sound with a convolution response VST. I guess a high line array have a own sound that can be sound full and nice. because every speaker distance to ear difference from the nearest is simular as a delay.

That's not a line array (lol). Did I say it on this thread? After a few years of running full range line arrays, which worked excellent, I changed them to a frequency shaded array. So not like a shaded array with weighting, but the lower one goes in frequency, the more drivers are involved.

In a Harman spin it would look something like this:
index.php
 

markus

Addicted to Fun and Learning
Joined
Sep 8, 2019
Messages
646
Likes
653
After a few years of running full range line arrays, which worked excellent, I changed them to a frequency shaded array. So not like a shaded array with weighting, but the lower one goes in frequency, the more drivers are involved.
Interesting. Why did you change it? Just for experimentation?
 
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
So you're still jumping to the same conclusions you had when you started this thread, based on speed and STEP responses? That's a bummer.
I highly suggest to start playing with some virtual speaker simulations and tools like RePhase and/or VituixCAD, make crossovers, see what the STEP looks like etc.

I know this. A speaker is a electro mechanical Bandpass Filter and of course Low pass+speaker give the overall step response rise speed and decay speed which depend on the speed of the filter and the speaker. you can correct phase and when tweeter and woofer are good aligned then can not see in 2 way systems the woofer/mid step.

rephase or phase linear filter for low frequency is not usable for me. because i make music and play real time keyboard. a speaker should not have more than 5 ms latency. DAW 8 ms + speaker 5 ms is 13 ms. thats what i not notice as a lag when play keyboard. but more as 16 ms is bad. .

But wy so many speakers have a woofer/mid delay of about 500 µsec ?. is this realistic sound. I have a ribbon tweeter set on top of klai and when i choose less delay as 300 µsec it sound better or go lower in frequency. https://www.monacor.de/produkte/components/lautsprechertechnik/hi-fi-hochtoener-/rbt-35sr/ t can go very low but this ribbon have some small distortion peaks that can hear much when go lower as 2 khz even when not hear loud.

With a frequency response like your measurement showed, I wouldn't dare jump to any conclusion about anything when seeing the STEP response. Heck, by itself it's only marginally interesting to look at when it's made in an anechoic room. Granted one would know more info about crossover topology used etc.
If you keep jumping back to the same conclusions, there's only one conclusion left for me. Exit this thread trough the back door.



That's not a line array (lol). Did I say it on this thread?

you give link to this thread. https://www.rsr-concepts.com/vandermill/ DIY-Line Arrays


After a few years of running full range line arrays, which worked excellent, I changed them to a frequency shaded array. So not like a shaded array with weighting, but the lower one goes in frequency, the more drivers are involved.

but you still can upload impules. If it is no neutral sound doesnt matter and it is nice to hear. so all can hear the sound too.
 

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
@markus Yes, purely out of curiosity. As a learning experience. As I would have to park my speakers in our garage, for a living room renovation, I saw it as an opportunity to try something I had been simulating. The discussion and simulations that tweaked my curiosity started quite a while ago in this thread.
 

markus

Addicted to Fun and Learning
Joined
Sep 8, 2019
Messages
646
Likes
653
@markus Yes, purely out of curiosity. As a learning experience. As I would have to park my speakers in our garage, for a living room renovation, I saw it as an opportunity to try something I had been simulating. The discussion and simulations that tweaked my curiosity started quite a while ago in this thread.
How would you summarize the perceptual differences between line array and the shaded implementation above? Looks like it's largely "just" a difference in DI?
 

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
rephase or phase linear filter for low frequency is not usable for me. because i make music and play real time keyboard. a speaker should not have more than 5 ms latency. DAW 8 ms + speaker 5 ms is 13 ms. thats what i not notice as a lag when play keyboard. but more as 16 ms is bad. .
I only suggested RePhase as a tool to make IR's with crossovers so one can build ideal speaker models. There's not even a need to make anything linear phase. It's just a handy tool to make an IR or two to combine within REW as a virtual example. VuituixCAD can do way more and still show you the IR. You're simply reading a lot into a STEP responses that isn't there or true. Playing with ideal IR's and crossovers might make you see that. I'm convinced that I can't make you see it. :)

you give link to this thread. https://www.rsr-concepts.com/vandermill/ DIY-Line Arrays
Yep, my speaker definitely is an array. But what you described, a single top end driver with the rest coming in below 100 Hz is not an array.
There are plenty of reasons for me to choose an array instead of, for example, a single full range with helper woofer.
One of them being the way it reacts to the room. Another would be dynamic capability.

Remember me stating that I want to work with the room? Not against it?

Here's the influence of the floor and ceiling of 'one driver' compared to 'an array of 25 drivers' (the last mentioned being frequency shaded). The orange trace is the 'in-room' prediction:
index.php

The loudest of the two is the array, but I hope that part is obvious. What's probably way less obvious to think about is the large difference the floor and ceiling have on the results once we bring these two type of speakers into our room. Reflectiveness of floor and ceiling are the same in both sims. A hint of this result could already be seen in my in-room waterfall plot.
but you still can upload impules. If it is no neutral sound doesnt matter and it is nice to hear. so all can hear the sound too.
I could, in my huge thread over on DIYaudio I've uploaded some soundclips already, That's enough for me.
There's also some links to reviews from a few friendly visitors / fellow DIY builders in the very first post.
 
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
I highly suggest to start playing with some virtual speaker simulations and tools like RePhase and/or VituixCAD, make crossovers, see what the STEP looks like etc.

mechanical equivalent circuit loudspeaker is here. so the speaker can simulete together with the filters.


on a whole speaker LP filter+ driver give the result. and of course when you shift the input of the measure signal with rephase or EQ phase shift then you can fake a good result. but thats same as diesel fake meaasure then. so when want measure LP + driver then input should not change with EQ ir phase shifters.

I see in this post https://www.diyaudio.com/community/...-corner-placement.337956/page-25#post-6199645 that you have calc the distance to ear of positions. it is mm right ?. now when you add this distances in ms in a multi tap delay in each tap then you can simulate the sound too. A step response of course look strange then. the top system have 216 mm longer way. this mean 0.62 ms delay of sound. how it sound i did not know it can sound better. it is simular as a early reflection reverb. to simulate early refelctions there are also multi tap delay used.

suchg phase shift tools do only work with continues cylces periodic signals. but music is no periodic signal and so the phase shift tool can not wok good. how much cycles it need to detect the frequency and choose the delay for that frequency
 
Last edited:

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
How would you summarize the perceptual differences between line array and the shaded implementation above? Looks like it's largely "just" a difference in DI?
I cannot say too much definitive yet. Yes, I hear differences, as in the floor/ceiling influence is a bit less, the top end suffers way less from comb filtering, making it practically a non issue. Can you hear that? Euhhh, that's already a bit harder to tell. The drop in room influence is noticeable.
Overall they still sound quite similar, with the frequency shaded arrays being a bit more clear and more consistent over a larger area. I've used them for Home Theater quite often, there I definitely prefer the frequency shaded array. Very good intelligibility and low coloration. Again, more consistent over a large area. Worth the tweak? I wasn't missing anything before. Had I not known the difference ....
Fun as a learning experience? Always!

Vertical directivity plot of a full range array compared to a frequency shaded array:
shaded-unshaded.gif


I haven't spend enough time on it yet due to not having as much free time alone with my speakers as I used to have due to Covid. My girl works from home, giving me less and less opportunity to play extensively with sweeps etc.
I've had some driver issues lately, they had spend a winter in my garage and some of them broke down due to corrosion of the lead wires of the drivers, where they connect to the voice coil. (open circuit) At that time they were about 10 years old.
That made me line up for yet another curiosity driven experiment. All of the Vifa TC9 FD18-08 drivers will be replaced with the Scan Speak 10F 8414G10.
Once that has taken place I will (have to) redo all DSP processing. Do I expect huge differences? Not really, but I must confess I've been surprised before (with a fun test of 5 different types of amplifiers).

IR and STEP are even more clear than the full range array, that part is obvious. The IR/STEP I've shown here was from the full range array, as was the Wavelet. The presented Waterfall plot is from a more recent frequency shaded array measurement.
 
Last edited:

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
I see in this post https://www.diyaudio.com/community/...-corner-placement.337956/page-25#post-6199645 that you have calc the distance to ear of positions. it is mm right ?. now when you add this distances in ms in a multi tap delay in each tap then you can simulate the sound too. A step response of course look strange then. the top system have 216 mm longer way. this mean 0.62 ms delay of sound. how it sound i did not know it can sound better. it is simular as a early reflection reverb. to simulate early refelctions there are also multi tap delay used.
Did you miss the part where I said: frequency shaded array? Anyway, if you really want to know how an array works as opposed to a single driver, look at that plot I just posted and look at the theory of infinite arrays right here.

If you avoid the part of using an array in a room, you basically do not get why anyone would use an array (in a room). You're already making up how it sounds. Yet the plot I've posted is conveniently skipped.

Another way of thinking about it: as soon as you use a single driver in a room with a floor and ceiling, you've created a virtual array already.
Just look at your own waterfall plot, taken inside a room. And after that look at that predicted orange plot of 'one driver in a room' again, does it look familiar? Now measure it from half a meter further away, did it change?

Now where's that back door (lol).
 
Last edited:
OP
B

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
691
Likes
124
Location
germany
Did you miss the part where I said: frequency shaded array? Anyway, if you really want to know how an array works as opposed to a single driver, look at that plot I just posted and look at the theory of infinite arrays right here.

I ask you which speaker you have and you answer in this post nothing about frequency shaded array, also on your linked pages i have not read about frequency shaded https://www.audiosciencereview.com/...r-itd-see-measures.28585/page-12#post-1030534 what have i miss ?

If you avoid the part of using an array in a room, you basically do not get why anyone would use an array (in a room). You're already making up how it sounds. Yet the plot I've posted is conveniently skipped.

Another way of thinking about it: as soon as you use a single driver in a room with a floor and ceiling, you've created a virtual array already.
Just look at your own waterfall plot, taken inside a room. And after that look at that predicted orange plot of 'one driver in a room' again, does it look familiar? Now measure it from half a meter further away, did it change?

Now where's that back door (lol).

sure room make too reflections but an array with much distance make even more room reflections and give more dense so maybe can sound depend matter of taste better. It have own sound. dont understand me wrong. when not mix and make music can use what like and sound better. when make music there is a neutral speaker usefull .if like line array sound then use a convoltion impulse for that or record a vocal or so over line array. so it can sound on headphone and speakers more simular. make and mix music on a line array then sound on a neutral speaker maybe not so good as on a line array. And when mixing music maybe use a line array impulse let the sound of vocal or instruments sound better. there are many guitar cabinet impulses and mixed cabinet impulses wav out but no impulse of a line array i find so you maybe can upload one.

I find a frequency shaded array but the specs are not show clear after which frequency only 1 speaker is used. https://www.av-iq.eu/avcat/ctl18527/index.cfm?manufacturer=innovox-audio&product=sla-micro . seem all speakers play upto 650 hz because they write

Maintains 25° vertical pattern control to 650Hz

I think there need limit high freq when want no effect sound. in your case the 0.6 ms is around 1.5 khz. maybe can do swichable if want a neutral or effect sound. with neutral to avoid comb filter effects the highest frequency of the many speakers need at least below 1 khz and only 1 speaker play more than 1 khz in your case i think.
 

Zvu

Addicted to Fun and Learning
Joined
May 1, 2020
Messages
830
Likes
1,416
Location
Serbia
....I've had some driver issues lately, they had spend a winter in my garage and some of them broke down due to corrosion of the lead wires of the drivers, where they connect to the voice coil. (open circuit) At that time they were about 10 years old....

How could this be ? Your cabinets are closed and cones of TC9 are treated. Are you sure that it happened over one winter period ?

...That made me line up for yet another curiosity driven experiment. All of the Vifa TC9 FD18-08 drivers will be replaced with the Scan Speak 10F 8414G10........

Wuuut ? :)
 

Wesayso

Active Member
Joined
Dec 30, 2021
Messages
123
Likes
291
Location
The Netherlands
I ask you which speaker you have and you answer in this post nothing about frequency shaded array, also on your linked pages i have not read about frequency shaded https://www.audiosciencereview.com/...r-itd-see-measures.28585/page-12#post-1030534 what have i miss ?
Well, your jumping pretty fast. I've ran my arrays as full range array from the end of 2004 until september 2020. Over the winter of 2020-2021 I have converted my speakers, out of curiosity, to frequency shaded arrays. That doesn't change anything I said, you just had a misunderstanding of my post where I tried to explain that your idea of a single driver above 100 Hz wasn't an array in any form, way or shape.

sure room make too reflections but an array with much distance make even more room reflections and give more dense so maybe can sound depend matter of taste better. It have own sound. dont understand me wrong. when not mix and make music can use what like and sound better. when make music there is a neutral speaker usefull .if like line array sound then use a convoltion impulse for that or record a vocal or so over line array. so it can sound on headphone and speakers more simular. make and mix music on a line array then sound on a neutral speaker maybe not so good as on a line array. And when mixing music maybe use a line array impulse let the sound of vocal or instruments sound better. there are many guitar cabinet impulses and mixed cabinet impulses wav out but no impulse of a line array i find so you maybe can upload one.
The picture I've shown you demonstrates that after DSP the frequency curve of an array shows less bumps and dips due to room effects than a single driver speaker would exhibit. True for both the full range array and the frequency shaded example by the way. Meaning it varies less in room than a single driver speaker would. That doesn't mean it will make it sound more dense. Quite the opposite.
Actually, if it weren't for the physical senses you get you'd be fine to compare it to giant headphones, but in a good way. Better than headphones because of the imaging/stage and the physical senses. It does sound dryer in-room than the single driver with DSP applied and absorbing the early reflections. Strange huh? How a limited vertical pattern can clean up the results. I hope you didn't miss the tight vertical pattern an array exhibits. It covers seated and standing positions but not much more. It "seems" like it misses the floor and ceiling, while in reality they are virtual extensions of line length. Can you imagine that?

So no! You can't mimic an array with your speakers. In reality I stand a better chance of mimicking your speakers with my arrays. So You'd be better of finding those sound clips if you're really curious.

I actually use more than the arrays alone, another novelty in this thread. As I rob the room of its (obvious) early reflections with my damping panels, I bring back that energy as ambient sound trough a pair of ambient speakers. I could explain it all here, but you'd jump to various new conclusions that are not valid.

If I had an ideal space available, I would do what I do with the ambient speakers in a passive way. Meaning I would have enough room to divert all acoustic energy around in the room using diffractive panels and well placed absorption. Don't just throw away the energy but rather redirect it. My place is too small for that and it is used for multiple purposes (living room) so I do the next best thing I can come up with.

I find a frequency shaded array but the specs are not show clear after which frequency only 1 speaker is used. https://www.av-iq.eu/avcat/ctl18527/index.cfm?manufacturer=innovox-audio&product=sla-micro . seem all speakers play upto 650 hz because they write "Maintains 25° vertical pattern control to 650Hz"
Another wrong assumption. It is the "array nature" that lets it have the restricted vertical beam like nature down to 650 Hz. That means that more than one drivers are playing and are causing the beam-width to be narrower.

Lets see, the vertical pattern of a single 3.5" driver:
1x TC9 FR Directivity (ver).png

Vertical control down to about... 7 KHz, but in a very limited way I'd say...

Now let's see the array's vertical pattern, 25x the same driver as the above.

25x TC9 FR Shaded 19.0 as build-notches-ABEC-minphase-20dB Directivity (ver).png

Vertical beam width control down to about 200 Hz. How is that happening? With all those drivers? You know what? It is because all those drivers that it has the limited beam-width vertically. Like a big 15"driver starts to beam way earlier than a 3.5" driver.

Are there any drawbacks? Sure, you see those horn shaped curves in the plot? They are the "dreaded" side lobes of an array consisting of multiple drivers. That is because it isn't a continuous array, but rather a series of 3.5"drivers. Use a smaller driver and those horn like lobes would go up in frequency. One drawback of doing just that, smaller drivers wouldn't be able to play the bottom end, so you'd need a support system down low.

But look at how much energy is hitting the floor and ceiling with that single driver. Depending on the room the ceiling reflection will be at about ~41 degree to 57 degree with the array, and at about ~51 degree in the single driver case. (in my room with a 3 meter high ceiling at my listening distance)
That means the single full range driver would see strong reflections from ~6 KHz on down, while the array would have (a few of) them around 5 KHz - 8 KHz and not much more.

Hey, didn't we see that in the graph I posted earlier?
index.php

I think there need limit high freq when want no effect sound. in your case the 0.6 ms is around 1.5 khz. maybe can do swichable if want a neutral or effect sound. with neutral to avoid comb filter effects the highest frequency of the many speakers need at least below 1 khz and only 1 speaker play more than 1 khz in your case i think.
Again with the assumptions... not much of a reality check though... Your thinking is off, as it does not factor in the distance from the array. Comb filtering with the filtered array is a non issue. With the unfiltered array it was a minor problem at frequencies above about 8 KHz. Reflections due to lobes were there too, but no comb problems and the same avoidance of floor and ceiling reflections. Avoidance isn't the right word, as the floor and ceiling even help to make the array seem longer than it is at low frequencies, making the array act like an almost infinite array. Actually making it worth using an array in a room.
Unshaded Power+DI.png

I will state that things might not be intuitive right away, but maybe, just maybe some of this will sink in. Many people make too many assumptions about arrays anyway. The theory might be more dense than it's sound (lol).
At least I am happy enough to state that reality lines up pretty good with the above theories. Totally unpredictable, the direct sound vs indirect sound of a line array has a more favorable balance than that of a single driver. Almost horn like in a way, as it controls directivity.

The fun part is: the horizontal pattern is quite comparable to that of a single driver, making the 3.5" driver a pretty good choice. Especially in a rounded, egg shaped cabinet.

I hope this clears up some of your misconceptions. But I rather doubt it. Because even the mind is more dense than my sound is :D.
 

Attachments

  • Unshaded Power+DI.png
    Unshaded Power+DI.png
    40.8 KB · Views: 48
Last edited:
Status
Not open for further replies.
Top Bottom