TOPIC:

Hi - first of all thanks to both "pos" for rephase and to minidsp making diy really fun

I am buiding a two-way, crossing the drivers at 350 hz using 24db LR filters. These 24 db IIR filters are easy to just implement on the 2x4 HD. The next step is to linearize these filters. I found also this to be easy in rephase under the "Filters Linearization" tab, 1000 taps was ok. Now comes the questions:

1) Do I import the resulting correction file to all four channels on the 2X4HD?
2) Does this result in total 4X1000 taps = 4000 taps?

By the way, rephase works flawlessly on my Mac through the Wine application

Best regards Alf from Norway:-)

Hello Alf,

Thanks for the kind words, and glad to hear you manage to run rephase on Mac without too much hassle.

I don't know the 2x4HD enough to answer your question, it depends if the FIR is applied on the input (ie before the crossover) or on the output.

Regarding the correction itself, please be aware that a LR 24 dB/oct acoustical target crossover will typically call for different and asymetrcial electric slopes (ie filters in the 2x4HD).
You cannot simply apply 24dB/oct filters, you need to get the existing response of the driver (without filters) into account.
You will typically need a softer slope (eg 12dB/oct) on the high pass part because of the natural rolloff of the driver/enclosure. You will also most likely need some EQ.

In the end, once you get an acoustical LR 24dB/oct crossover (ie perfect LR 24dB/oct acoustical filters on both sides, with good phase tracking and proper time alignment) you can indeed simply apply the phase correction on that target.

Getting the minimum-phase crossover right is much more difficult than the phase correction part, but it is also the most important and rewarding part.

Hope this makes sense

Thanks for your clarifying answer:

I can confirm that the FIR block in 2x4HD are placed after the crossover on each output channel. Based on your suggestions I will proceed as follows:

1) Selecting a crossover point where the drivers in the enlosure have an even frequency response on both side of the crossover point, 350 hz is fine for my two SEAS 6.5" drivers (bass + fullrange)
2) Use the EQ blocks on the 2x4 HD (PEQ) to linearize the drivers in the crossover area
3) Adjust gain on the output channels, so they have similar output in crossover area
4) Add high and lowpass crossover on each channel, I use 24db LR in the 2X4 HD
5) Timealign the two drivers using the delay function on each output channel on the 2x4HD
6) Import the correction file from RePhase into the FIR block, specifically made for 24 db LR, 350hz

I use 1000 taps for all four output channels, so no more time-delays should be necesarry. I can confirm that Wine runs RePhase without any problems on my Mac, exporting correct file formats. Importing them into the 2x4HD is done correctly

Best regards
Alf

Sounds like a plan!

Linearizing the response with IIR requires lot's of EQ though, especially if you want to get the phase right as this implies going far down the stop band.
An easier solution is to get it directly to where you want it to be, adding just what is needed.
To do so you can use rephase as a simulation tool, only using EQ and filters that can be replicated in the minidsp IIR engine (proportional Q EQs, etc.). You simply upload a measurement into rephase and set your target filter (here LR 24dB/oct @ 350Hz) in "compensate" mode in the minimum-phase filter tab. You then search for the most effective/simple way to get a flat magnitude and phase trace with EQ and filters you can reproduce in IIR.

You need a measurement with a good low noise floor to be able to dig deep into the stop band, as any noise floor will result in a 24dB/oct artifact once in compensate mode, and you will of course not want to correct anything in that area.

Thanks, good advice:

Yes I can see how using a measurement of each of drivers response in the enclosure is an elegant approach. The challenge is perhaps as you mention to take reliable and good measurements. Perhaps outside the house is the best way here.

Best regards
-Alf

You can probably go away with close mic measurements at these frequencies, with also some additional averaged measurement around the listening position for the magnitude alone (eg MMM, moving mic measurement).
You cannot (and should not) separate the speaker from the room at these frequencies anyway.
Also make sure the baffle size is the same for the woofer and midrange, as you don't want to have to deal with different baffle steps throughout the crossover.

alfegutt wrote:

Thanks for your clarifying answer:

I can confirm that the FIR block in 2x4HD are placed after the crossover on each output channel. Based on your suggestions I will proceed as follows:

1) Selecting a crossover point where the drivers in the enlosure have an even frequency response on both side of the crossover point, 350 hz is fine for my two SEAS 6.5" drivers (bass + fullrange)
2) Use the EQ blocks on the 2x4 HD (PEQ) to linearize the drivers in the crossover area
3) Adjust gain on the output channels, so they have similar output in crossover area
4) Add high and lowpass crossover on each channel, I use 24db LR in the 2X4 HD
5) Timealign the two drivers using the delay function on each output channel on the 2x4HD
6) Import the correction file from RePhase into the FIR block, specifically made for 24 db LR, 350hz

I use 1000 taps for all four output channels, so no more time-delays should be necesarry. I can confirm that Wine runs RePhase without any problems on my Mac, exporting correct file formats. Importing them into the 2x4HD is done correctly

Best regards
Alf

Alf ,

Why not just use the 1024 taps on all channels ?

Why 1000?

Just curious..... maybe a reason, just asking

Hi

Nothing wrong by using all 1024 taps, it will produce the same result. Its just that 1000 is a "convenient" number:-)

Best regards
-Alf

Okay lol