Correct calibration of my active two-way speakers with REW

[Horst]

As I understand it, my HTPC plays the sweep and not the Tascam, which I only use for my mic to measure with REW. No output activated on the Tascam. I somehow get strange impulse responses from my speaker. I am not sure if everything is ok with my measurement chain, maybe test a UMIK 2.
I use the latest driver

Ok thanks, yes i will try it and show you my measures after.

 

[Horst]

Ok i try the Java and now looks the Impuls very better - but look
i think with the Asio 2.15 is something wrong
the first two measures are with java the other with ASIO4ALL
Dont look the frequency response, that is my full active speaker without EQ.
what you mean ?
The REW Beta has Problems with ASIO or is it my old Tascam?

 

[thothsong]

If the Java driver works, there's really no reason not to use it. But with ASIO4ALL, did you try increasing the ASIO Buffer Size in the ASIO Control Panel in REW Soundcard Preferences?

 

[Horst]

No i dont

 

[Horst]

You mean that?

 

[ddude003]

I wonder if this could be a multi-client vs exclusive mode thing??? How does Java vs aiso, aisoforall vs flexaiso manage this? I have no clue...

Any way to describe the weird tones? Pitch or speed changes?

 

[thothsong]

You mean that?

Yes, try increasing from 512 to 2048.

 

[thothsong]

I wonder if this could be a multi-client vs exclusive mode thing??? How does Java vs aiso, aisoforall vs flexaiso manage this?

With the new Java driver, if you select the variant of the device with an EXCL: prefix, it will use exclusive mode, and also give you multichannel support for output.

 

[Horst]

Interference, strange slight crackling, irregular noise but with Java it is now good.
You can see the fidgety curve with the impulse, just like with the SPL.
Also the waterfall changes drastically with ASIO when changing the length (256 to 2M). With Java, the curves look more credible. However, I have only observed this since the latest beta.

And what is the recommendation with the new REW Beta version and the new JAVA version in relation to my "old" hardware (Tascam)? This question is important for me before I remeasure everything.
I also have no problem ordering a UMIK 1 (should be perfectly sufficient) to adapt to the new software.
Maybe John can say something about this?

 

[John Mulcahy]

The ASIO issues sound like dropouts so a larger buffer should tackle those, but you are better off using the Java EXCL option. No problem with using the Tascam.

 

[Horst]

Thank you then i will measure with Java the Excl. Mode.
Do I have to pay special attention to the Windows settings since Java takes over the Windows settings? Or is that different now with the excl. mode?

 

[John Mulcahy]

The EXCL device entries should apply the chosen sample rate to the device directly.

 

[Horst]

OK thanks

 

[Horst]

I have now had time to measure, but REW makes different measurements with Java or ASIO4All, especially in the level.

I calibrated the level with Java_default to 77.5db and with ASIO4All the measurement is different but especially different in level - 10db diff.

Left LS large measured at listening position

blue curve : Java_default

green curve : Java Excl. CX-A5000

red curve : Asio4All

What is REW doing there? There are 3 different measurement graphs, especially in the level. I only changed the output in the preferences, nothing else.

Which measurement can I trust now?

 

[John Mulcahy]

Please attach the mdat file.

 

[Horst]

Here the mdat

 

[John Mulcahy]

ASIO4All has a different SPL calibration, by 2.14 dB. Adjusting that to be the same as the other measurements means the ASIO and WASAPI Exclusive results are very close:

Since "left m EQ Java default" matches the WASAPI exclusive result below 80 Hz and above 16 kHz I would say that has some EQ applied to it, perhaps by some Windows process which gets bypassed by the exclusive driver.

 

[Horst]

Thank you, does ASIO4ALL have advantages over Java exclusive?
Which drivers do you use?

 

[John Mulcahy]

I can't think of any advantages of using an ASIO wrapper instead of WASAPI exclusive if WASAPI exclusive works. Some ASIO wrappers (e.g. FlexASIO) allow different backend drivers to be used if there are specific circumstances that would require that. I use WASAPI Exclusive.

 

[Horst]

Ok thank you very much for your tips. It often happens that REW loses the driver (Java excl. CX-A5000) and I have to select it again in the preferences. This is not a problem for me, but once i forgot to reselect the driver when it gave me the message and I continued to measure with the Java_default driver and equalised my fronts.
That's why I noticed the different measurements, so I will use the Java_exclusive drivers in the future.
Thank you!

 

[John Mulcahy]

It often happens that REW loses the driver

If the interface is disconnected while REW is running or not connected when REW starts that would happen, otherwise it should be remembered.

 

[Horst]

Unfortunately, this happens to me during a measurement session. Today I even had to restart the PC because the driver (CX-A5000) no longer appeared. Everything is running normally, everything is switched on and connected, but REW still loses the setting frequently.

 

[John Mulcahy]

REW offers the interfaces the OS (via JavaSound) says are available. For an interface to disappear the OS has to have lost sight of it somehow.

 

[Horst]

ok thank you very much i will look what happens but it is not really a problem for me.

 

[Horst]

My DOUBLE BASS ARRAY

The delay of the first measurement of the left loudspeaker with the Java excl. I entered the driver at delay (-0.1044) at "timing / use acoustic timing reference". I made all of the following measurements with this timing reference. I would like to simulate my DBA based on the measurements with the "alignment tool". But this tool is so extensive and I would like to know how to use it to simulate the DoubleBassArray or a transition from the subwoofers to the loudspeaker - I would also like to understand it. Here are the front and rear array measurements.


With the classic DBA you invert the rear array, enter the room length as the delay and lower the gain of the rear array by 1 to 2db.
With my DBA I get a great frequency response and waterfall without inversion and without delay?




The classic DBA with inversion etc. also works, but the waterfall is not that nice and I need more DSP filters. How do I optimize the phase with the simulation or what tools does the simulation (alignment tool) provide me with to adjust or simulate the DBA in the best possible way?

Can anyone help me with the simulation?