Development of open baffle 3-way plus subs

[pompon]

Acourate not generate filter 64 bits resolution ?

I don't understand why you talk about resolution lost.

[UliBru]

Acourate calculates with 64bit float precision, also the filters show up 64bit float.
(Indeed there are some special functions in Acourate which even use 80 bit float numbers!)

But we are not talking about this. We are talking about the numbers sent to the soundcard or DAC. Here we have a switch to typical 24 bits integer numbers. And so it does not help if the filter uses xxx float numbers. If a filter has to attenuate by -6 dB this will cause loosing a bit of 24 bit PCM. No way around.

[pompon]

My dac (Exasound E38mk2) work in 32 bits.
"ESS Technologies ES9028PRO Reference DAC Chip

exaSound DACs are based on the ESS Technologies ES9028PRO Sabre32 reference DAC chip, the newly released successor of the legendary Sabre ES9018. ES9028PRO delivers 32-bit precision capable of resolving even the smallest signal details. "

So, I am not impacted with resolution lost because my hardware work at 32 bits ?

[UliBru]

What is the SNR value of your DAC?
I do not believe in the 32 bit resolution which corresponds to 192 dB.

[pompon]

Channel Separation, typical
133dB @ 1kHz for adjacent channels
123dB @ 1kHz between channel groups

[UliBru]

The DAC datasheet specifies a dynamic performance of
DNR = 129 dBA (A-weighted)
for the 8-Ch differential current mode (= using more than two channels).
This corresponds to a resolution of 21.4 bits which is far away from 32 bits.
Please note that this is already a best cast scenario.

[Lewinskih01]

I have been making varous experiments and been short of time, but finally arrived at trying these U-baffles for the 18" midbasses, flat OB for the twin 8" mids, open back Beyma TPL. XO; 60, 300, 1300Hz NT2 for now.

20220703_105000.jpg (515.31 KiB) Viewed 1464 times

Next I want to linearize the drivers (to be able to time align). Where should I measure the midrange? Mic pointing to the driver on axis horizontally, very near the driver (like 5cm). From which driver, or should I measure at the tweeter height? Or disconnect one 8" and measure on axis with the Active 8"?

[UliBru]

IMO with open baffles a close-up measurement only makes sense for small corrections of the driver itself.
But if e.g. the two midrange drivers are wired together it is necessary to measure in a greater distance. Even 40 cm may be too close.
And at (the quite usual) 1m distance you have to keep room influences in mind. Then it is better to measure the speaker far away from boundaries, e.g. in the center of the room. Place the mic between the drivers.

[Lewinskih01]

I'm looking into linearizing as part of the path towards time-alignment, per Mitch's ebook. Maybe worth asking the bigger picture question: for OB, does it make sense to do this, or am I better off going straight into time-alignment?

[UliBru]

In a first step ypu can directly go with time-alignment. So you reach a first result without too much effort (including possible traps).
Then in further steps you are of course free to dive deeper. You will also learn a lot
IMO this sequence is way better than to start immediately with a complex setup and possible bad experiences.