How to input in Boxsim minimal phase, SEO and position of sound to the listener

[ygg-it]

Hello I'm using boxsim and it is very powerful, but I have a doubt and I cannot go further.
I need here a kind help from someone who knows how Boxsim works very well.

I'm building a 2-way steppedbox louspeaker where the tweeter's acoustic center and basket frame are not aligned with the acoustic centre and basket frame of the woofer.

Let us say (I'm NOT questioning now where are the acoustic centers) that the frame position is at the mounting holes of the basket in the baffle (Woff= 0cm; Toff= 2cm stepped back from Woff), and the acoustic centers are at coils edges or first magnetic plate (Tac= 1cm from TWoff; Wac= 55mm from Woff)

In principle, if I'm correct and I have understood well:

the "position of sound to the listener of the woofer" = Woff = 0cm (in baffle and position folder)
the "position of sound to the listener of the tweeter" = Toff = +2cm (in baffle and position folder)
the "position of point of sound origin (SEO) of the woofer" = Wac = +5.5 cm (in data driver folder)
the "position of point of sound origin (SEO) of the tweeter" = Tac = +1 cm (in data driver folder)


My problem is that I don't have the measured acoustic phase resp. but only the minimal phase response of each driver which I have extrapolated from the frequency response of the two drivers by using VACS software.
In fact there is a 1:1 relationship between the magnitude response of a system and its minimum phase (Hilbert trans.).

In Boxsim, if you don't input any acoustic phase data, SEO doesn't have any effect.

Now the question is: how would you simulate my system (which input field, what number and which sign) using the minimal phase of the drivers and/or whithout using any acoustic phase at all?

In theory if SEO has to be set null for both drivers, I will probably need to use the delta of the acoustic centers, hence, I think that the "position of sound to the listener of the tweeter" must be set to 0 cm (because that is the driver with acoustic centre closer to the listener in my design) and "position of sound to the listener of the woofer" should set to 55-10-20= 25mm. But I obtain different results.

Thanks. If you can use the numbers in the example will be much easier for everybody to understand.

[walwal]

http://forum.visaton.de/showthread.php?t=30736

"..It's funny, lately we get very similar SEO requests about our consumer address. Here is the explanation:

The statement in BoxSim "Location of the place of the sound source" is unfortunately often misinterpreted, which is probably due to the naming. Briefly explained: This value is used only for the phase characteristic as a reference. Otherwise the acoustic phase can not be assigned.

For an exact explanation I have to go further. For a correct simulation BoxSim needs the following measurements: Frequency Response, Acoustic Phase, Electrical Phase, Impedance History. These measurements must be carried out according to a defined work procedure and under defined conditions (such as the frequency response measurement in anechoic chamber in a standard baffle according to DIN IEC 60268-5) in order to be able to guarantee a real simulation.
The acoustic phase is to be considered extremely critical here, since a few millimeters can already cause problems in the simulation (one can easily explain with the formula "c = lambda * f" -> the wavelength is directly related to the frequency). A short example: If one measures the acoustic phase at a distance of 1 m, this is shifted to a measuring distance of e.g. only 0.7 m.

The true generation of sound on the membrane is frequency-dependent and not so easy to determine because e.g. At low frequencies, the entire membrane oscillates evenly piston-shaped, but at high frequencies only a certain proportion. But also several parts can oscillate simultaneously and in different directions (eg parts of the bead). The SEO is therefore necessary to define the acoustic phase at the measuring point. Quasi a kind of distance assignment of the acoustic phase.

When determining the acoustic phase for BoxSim, we proceed as follows:
The loudspeaker is flush mounted (ie milled) in our standard wall. The microphone is then placed using a gauge to an exact and known distance centered above the speaker chassis. The known distance (in our case: 1 m) is thus between microphone and baffle. Now the sound source is determined by an acoustic distance measurement (via an impulse response) and calculated with the known distance (micro baffle). This allows the acoustic phase to be assigned perfectly. This value usually fits quite well with the distance "baffle - Staubschutzkalotte", but this is just a coincidence. It is important that this value may only be changed in connection with the acoustic phase. Otherwise, the assignment is incorrect and the simulation is incorrect.
Incidentally, since BoxSim 2.0, this measurement is also measured at 18 angles at 5 °, so that the phase etc. is also recorded at an angle in order to simulate the emission characteristic even more realistically..."

[ygg-it]

Thanks for the link. That confims how to input the driver offset.

Now I just need to understand if "minimum acoustic phase" extrapolated from the SPL curve can be used instead of the measured acoustic phase which I don't have, or better to leave the acoustic phase blank.

And (in the case it is used) if/how to set SEO in order to avoid double counting somewhere.

[walwal]

I am sorry, i'm not shure about this. Please ask UweG, he developed Boxim.

http://forum.visaton.de/member.php?u=4930

[ygg-it]

I am sorry, i'm not shure about this. Please ask UweG, he developed Boxim.

http://forum.visaton.de/member.php?u=4930

Thanks, I written to him a private message but I don't know if right as I don't know your language... Request is this, if Mr. Uwe is still around here:
Dear Mr.Uwe, are you so kind to reply to this my message relating to my quotes in the attached design or in general:

if "minimum phase" extrapolated from the driver SPL curve can be used instead of the missing acoustic phase, or it is better to leave the acoustic phase blank and let Boxsim to calculate it.

And (in the case minimum phase is used) if/how to set SEO in order to avoid double counting with the distance to the listener (which should consider only the multibox baffle tilting/shifting and not the offset of the acoustic centres)"

Many thanks

[UweG]

Now I just need to understand if "minimum acoustic phase" extrapolated from the SPL curve can be used instead of the measured acoustic phase which I don't have, or better to leave the acoustic phase blank.

Leave it blank. In this case Boxsim is using the TSP model to simulate the phase. This is much better than using strange data. A speaker is not a minimum phase system in this context.
If you are following this advice, please make sure that you entered correct values for the diameter of the diaphragm and the shape of the disphragm (both on page "baffle and position" in the chassis editor.

[ygg-it]

Leave it blank. In this case Boxsim is using the TSP model to simulate the phase. This is much better than using strange data. A speaker is not a minimum phase system in this context.
If you are following this advice, please make sure that you entered correct values for the diameter of the diaphragm and the shape of the disphragm (both on page "baffle and position" in the chassis editor.

Thanks! I will leave it blank!

Last question then (as I cannot input the SEO): should I put the different distance to the listener calculating from the baffle holes or from the acoustic centers? (I'm using tilted shifted boxes..)
Thanks..

[UweG]

If you enter the TSPs, the diameter of the sound output and the shape of the diaphragm, Boxsim will estimate the phase referring to the baffle (if you did not enter a measured phase response).
So the only thing you have to enter to let Boxsim know the step in your baffle is the diffrent position of the sound output in absolute (global) coordinates. The bass drive of your box is shifted 2cm towards the listener (or the tweeter 2cm away from the listener = -2cm towards the listener).

[ygg-it]

If you enter the TSPs, the diameter of the sound output and the shape of the diaphragm, Boxsim will estimate the phase referring to the baffle (if you did not enter a measured phase response).
So the only thing you have to enter to let Boxsim know the step in your baffle is the diffrent position of the sound output in absolute (global) coordinates. The bass drive of your box is shifted 2cm towards the listener (or the tweeter 2cm away from the listener = -2cm towards the listener).

Many thanks.
Good news: your software works very well!
Bad news: I definitely had to reverse the polarity of the tweeter and review the tweeter offset position. (I’m using a fist order filter). Fortunately, I hadn't glued yet the tweeter box.
Also, I did a mistake in your program as I put +2cm (a positive offset) for the most distant speaker!
Now not only it sounds much pleasant with the reversed tweeter, but the optimum 3D image I found by ear matches the most linear curve I run in your simulator by input different offsets!
I'm happy to give-up to the acoustic phase: not only it is difficult to measure it correctly, but the "SEO" position is too much critical as the driver's acoustic center position is a big question mark. Around internet I saw too many discrepancies where it should be (coil position, at spider, at top magnetic plate, 1/3 of the cone, top or at base of dust cap..etc. etc..) and even if you exactly know where it should be, it is critical to measure.
Attached the comparison of your software and the basic measurement I tried to make (with a primitive microphone) using pink noise at 1/3 octave band, at 2 meters, and between woofer and tweeter.
I’m waiting for a better microphone for a friend of mine. But I found the curves similar, don’t you?