thank you for your quick responce.
i just finnished this arta box ,so i just took a qiuck reading.
when i scaled the responce curve its looks more reasonable,so i measured another driver and it looks normal.
i gues its working as it should
so what is the "best"way to measure ts parameters?
turn down the drive level of stepped sine,to the point that the curve is about the same as pn?
or just use pn?
tnx for helping
i just finnished this arta box ,so i just took a qiuck reading.
when i scaled the responce curve its looks more reasonable,so i measured another driver and it looks normal.
i gues its working as it should
so what is the "best"way to measure ts parameters?
turn down the drive level of stepped sine,to the point that the curve is about the same as pn?
or just use pn?
tnx for helping
ARTA & iMat
Hi Guys,
just for Information: Since a few days Ivo Mateljan is joining this forum as iMat.
If you want to adress questions, bugs, wishes or whatever please do it here.
I am sure Ivo will check this thread from time to time.
But please do not await an answer in every case.
Regards
Heinrich
Hi Guys,
just for Information: Since a few days Ivo Mateljan is joining this forum as iMat.
If you want to adress questions, bugs, wishes or whatever please do it here.
I am sure Ivo will check this thread from time to time.
But please do not await an answer in every case.
Regards
Heinrich
Hi,
yes I've decided to be the member of DiyAudio Forum. Lot of Forum members are using ARTA.
Up to now, I've been exclusively using email for the communication with ARTA users. Now I think that some discussions on Forum are interesting for large number of users and word of author will surely be welcomed.
Best,
Ivo
ARTA wish list
As iMat is listening I would take this opportunity to mention the following :
Measurement of Impedance using swept sine wave would be a welcome addition.
It would also be enlightening to know the procedure to calibrate a usb soundcard like Focusrite Scarlet 2i2 where one input is used for a calibrated mic (No separate mic pre used and so gain is unknown) and the other one for reference.
Thanks and regards
As iMat is listening I would take this opportunity to mention the following :
Measurement of Impedance using swept sine wave would be a welcome addition.
It would also be enlightening to know the procedure to calibrate a usb soundcard like Focusrite Scarlet 2i2 where one input is used for a calibrated mic (No separate mic pre used and so gain is unknown) and the other one for reference.
Thanks and regards
In my opinion half-sine wave measurements are a good method of dynamic measurements. I think this option should be in the Arta.
http://www.myro.de/piw-en.pdf
http://www.myro.de/piw-en.pdf
Hi soundrex,
I did not implement impedance measurement with swept sine intentionally.
The swept sine is excellent signal for loudspeaker response measurement, it give highest S/N without averaging, but makes problems if there is impulsive type of noise and noise at very low frequencies. When measuring impedance the loudspeaker behaves as microphone. We rarely have quiet environment, and impulsive noise is common - it gives resonance like spikes in measured impedance.
ARTA uses pure sine and random noise signals.
Sine signal gives hundred times better measurement S/N, but measurement is slow.
Random noise excitation gives fast measurement, but noise influence is larger than in measurement with swept sine. The good thing is that effect of noise is equally distributed over all frequencies, and there are much less chance that we detect false resonance.
Focusrite Scarlet 2i2 is relatively cheap, it has stable driver and it is very good for acoustical measurements. It has inputs/outputs typical for all professional sound cards:
a) Balanced outputs
b) Unbalanced headphone outputs
c) Balanced microphone inputs
d) Unbalanced line/instrumentation inputs
The trick in calibration is that you would calibrate front balanced mic. input connected to rear balanced output, and reference line input connected to headphone output. In balanced XLR connector signal is measured between pin 2 and 3.
For impedance measurements you can use headphone output and two line inputs with reference resistor 47-100 ohms (using circuit that is shown in figure 3.1 in Limp user manual).
Best,
Ivo
Your document clearly shows that half sine gives a special insight into dynamic behaviour of loudspeakers. The problem is that it is not always possible to know what behaviour is better. I just need more informations.
Best,
Ivo
Interesting!
This reminds me of this WT2 vs WT3/DATS debate (points vs sweep) :
WT3/DATS author: http://www.diyaudio.com/forums/multi-way/142440-wt2-wt3-2.html#post2618826
WT2 author: http://www.diyaudio.com/forums/multi-way/142440-wt2-wt3-3.html#post2619173
Last edited:
Hi,
yes, we can compare the systems that are close to ideal system.
But, as soon as we have non-minimum phase systems or systems with reflection the testing with signal is useless.
Why I say useless?
Because there are lot of excellent loudspeaker,that are non-minimum phase system. Even those few that are close to definition of ideal system becomes non-minimum phase in real rooms.
I do not doubt that if someone makes measurements on hundreds of loudspeakers he can adopt some response pattern recognition and make some conclusion, but it is a hard .. hard way ...
Best,
Ivo
My approach, which I have implemented in ARTA before ten years, is that measurement with both signal is useful.
Why?
Loudspeaker is highly nonlinear device, especially its suspensions, whose compliance changes with level of excitation. It is well known that change in excitation level shifts loudspeaker resonance frequency.
Measurement with wideband noise signals is equivalent to small signal excitation, as only a fraction of signal energy is located in the frequency region near resonance.
Measurement with sine signal is equivalent to large signal excitation. Additionally, ARTA uses heterodyne principle in acquisition of sine response that additionally increase measurement S/N (explained in user manual).
Literally, when measuring with sine signal you may sing, while during measurement with noise you must keep you mouth shut.
Best,
Ivo
The front-box integrated with the software, would that be a good idea? I doubt it... It would ARTA tend towards the more professional packages with integrated hardware.
Don't worry, ARTA will always be usable with standard and pro audio sound cards, as it is ARTA main purpose - to be virtual instrument for acoustical measurement using PC sound card.
If in future releases ARTA maybe will get interface for driving some dedicated automation hardware but it will be just luxury addition.
The good thing is that quality of acoustical measurements does not depend on such hardware as by using pro sound card with phantom powering, high quality microphone, sound calibrator and few cables, user gets true first class acoustical instrumentation.
Best,
Ivo
Hi Ivo,
Firstly, welcome to the forum!
I've been a registered user of ARTA for a number of years now and wanted to thank you for such a brilliant piece of software. (which I first heard of through this forum)
It's really hard to find any other software that can do so many types of measurements so accurately and relatively easily, certainly within the same price range. After trying a number of programs over the years I've settled exclusively on ARTA for measurement duties.
I've not been active on this forum nor used ARTA much over the last year as personal circumstances have precluded any speaker work but that will soon change after a house move and I will be dusting off my copy of ARTA and putting it back to work on some new designs.
I've posted elsewhere on the forum discussing the difference in results with sine and noise for impedance measurement of drivers, for example this one:
http://www.diyaudio.com/forums/multi-way/182688-help-understanding-physics-vas.html#post2461984
So thanks for confirming what I had thought - that the noise test signal simulates a much lower level signal around the resonance frequency than sine and thus results in a different resonant frequency due to suspension non-linearity...
Many people don't realise just how much Fs can vary with drive level, I actually plotted Fs vs drive level for two different drivers (measured using ARTA but plotted in excel) here:
http://www.diyaudio.com/forums/multi-way/182688-help-understanding-physics-vas.html#post2462768
On some drivers the shift is dramatic! Because T/S parameters change so much with drive level I model the bass response of the speaker for both the small signal and large signal values to see how much the response will vary with drive level...
Firstly, welcome to the forum!
I've been a registered user of ARTA for a number of years now and wanted to thank you for such a brilliant piece of software.
(which I first heard of through this forum)It's really hard to find any other software that can do so many types of measurements so accurately and relatively easily, certainly within the same price range. After trying a number of programs over the years I've settled exclusively on ARTA for measurement duties.
I've not been active on this forum nor used ARTA much over the last year as personal circumstances have precluded any speaker work but that will soon change after a house move and I will be dusting off my copy of ARTA and putting it back to work on some new designs.
I've posted elsewhere on the forum discussing the difference in results with sine and noise for impedance measurement of drivers, for example this one:
http://www.diyaudio.com/forums/multi-way/182688-help-understanding-physics-vas.html#post2461984
So thanks for confirming what I had thought - that the noise test signal simulates a much lower level signal around the resonance frequency than sine and thus results in a different resonant frequency due to suspension non-linearity...
Many people don't realise just how much Fs can vary with drive level, I actually plotted Fs vs drive level for two different drivers (measured using ARTA but plotted in excel) here:
http://www.diyaudio.com/forums/multi-way/182688-help-understanding-physics-vas.html#post2462768
On some drivers the shift is dramatic! Because T/S parameters change so much with drive level I model the bass response of the speaker for both the small signal and large signal values to see how much the response will vary with drive level...