MiniDSP Flex, 2 speakers and 2 subs. Recommended test settings?

[Boog1000]

Due to room modes in a small room, with a huge null in the listening position, I use 2 subwoofers (rear L and rear R) in addition to the front L and R monitors. This does a good job of evening out the low end.

Because the MiniDSP Flex only allows for 2 inputs, AFAIK it is not possible to separately measure the subwoofers, only the L and R monitors.

However, if I disengage the subwoofers, the microphone will measure a very large null in the low end, which will then affect the correction.

I'm assuming that if I run a measurement sweep with the subwoofers engaged, it will throw the timing off, and generally interfere with the quality of the measurement.

Do people have any thoughts on this?

My first measurement today, just to see what happened, I set it up like this.

As a part 2, I was also wondering about the best I/O setup for using a UMIK 1 microphone. I have an XLR based one but it's not calibrated.

I've read elsewhere on here that it's recommended to get the UMIK directly connected to AL via the advanced menu.

Would you recommend then having the input set directly to UMIK and output direct to ASIO (with ASIO4ALL pointed at the MINIDSP interface?)

 

[kevinzoe]

Boog
You are realizing the constraint of having multiple sound sources but only two channels to do measurements and corrections via Audiolense. You can keep the subs on with the monitors and measure each channel separately but you'd have to do the sub timing outside of Audiolense and have it ready to go before taking a AL measurement. This is a sub optimal approach in my opinion unless you have REW or some other tool to help get you the delay times you require.

I think you will find that most of us (I'm guessing here) use multi-channel DACs with either software like JRiver or ROON etc (or multi-channel miniDSP models) that can also handle multiple channels independently for freq and time domain corrections. In my case I run a HP Laptop as my music server that also has on it both JRIver and Audiolense. I map the JRiver channels to the AL ones, turn off JRiver when taking AL measurements and then use JRiver to invoke the AL created FIR filter within the JRiver DSP area of the software - dirt easy. Then a USB connection takes 6 parallel channels of info (e.g. music signal+freq response corrections+time domain corrections) to the DAC where its processed and then spits out 6 channels to my biamped mains and pair of subs

 

[Boog1000]

Thanks a lot Kevin for the reply. I am planning to use REW to optimise the subwoofer timing and XO (using the MiniDSP internal software) then do a series of 2 channel sweeps on Audiolense, initially to get the gain matched between subs and mains.

My listening setup will be Audirvana -> Hang Loose Convolver -> MiniDSP Flex -> 2 monitors, 2 subs.

Last week I did this measurement in REW which gives an idea of the current frequency response:

The nulls around 300hz and 8k are a bit of a mystery. I have not really started investigating them - the last year has been about the low end (previous REW measurements I wouldn't post here, they are serious null gore).

I have a hope that the big one around 8K is relatively easy to solve with some acoustic type solutions, e.g. moving things in the room a bit.

Despite posting about subwoofer timing, at this point mostly I'm hoping to focus my AL attention on classical music, or the issues from around 200Hz upwards.

I'm aware that I'll need to be careful designing any AL filters because some of these nulls can throw the whole thing off.

 

[kevinzoe]

Boog,
Thanks for sharing your REW freq response chart. Can you please confirm that the chart is without Audiolense FIR filter correction?
Some thoughts . . .
> your high freq roll off may be a result of speaker toe-in. Try changing the toe-in (or is it tow-in?) amount . . .
> Sub timing based on distance alone is likely not accurate due to group delay caused by the sub's electronics. Is the sub closer to you than the mains or are the mains closer in which case they may need delaying . . . This is where Audiolense shines in using impulse response to time align everything!
> Null at 300Hz - this may not be as ominous as it looks. An actual instrument note of D and D sharp and corresponding frequencies of 293 and 311Hz are affected by this null BUT they appear to be -4dB down from what looks like an 80dB average level. That's only 1dB more than the industry standard of +-3dB range.
> Null 70-150Hz - this is an awfully wide null. I'd be inclined to experiment with delay settings for the sub assuming that it's contributing sound within this range.

 

[Boog1000]

Thanks Kevin! I'm about to go to Seoul for a month, I'll pick this up when I'm back.

 

[juicehifi]

Unless/until you decide to invest in a multichannel dac that gives Audiolense direct access to the subs, you should configure your system with MiniDSP with the best possible sub integration, particularly with regards to avoiding cancellation between mains and subs. And then you should simply measure & correct the whole assembly as a regular pair of stereo with Audiolense. This usually works really well. I expect a subwoofer integration that is perceived to be seamless with this approach.

 

[trekker444]

Do you have Dirac installed? I also have the Minidsp Flex w/Dirac in my small office listening room (10’x11’ w/vaulted ceiling’s 8’-10’). This is definitely Not an ideal room, but I’m able to get pretty good results.

I initially was planning on using 2 subwoofers (SVS 3000 micros), but since I only have “1” chair I was able to just use (1) subwoofer. I just moved the sub to the area with the best response and let Dirac do it’s thing. Then I eq to my specific taste.

In a larger room, 2 subs definitely help even out the low frequency, but in a smaller room with 1 listening position 1 sub worked. You may want to try just 1 sub if you only have 1 listening position also.

Arnold

 

[deercreekaudio]

Due to room modes in a small room, with a huge null in the listening position, I use 2 subwoofers (rear L and rear R) in addition to the front L and R monitors. This does a good job of evening out the low end.

Because the MiniDSP Flex only allows for 2 inputs, AFAIK it is not possible to separately measure the subwoofers, only the L and R monitors.

However, if I disengage the subwoofers, the microphone will measure a very large null in the low end, which will then affect the correction.

I'm assuming that if I run a measurement sweep with the subwoofers engaged, it will throw the timing off, and generally interfere with the quality of the measurement.

Do people have any thoughts on this?

My first measurement today, just to see what happened, I set it up like this.

View attachment 55113

As a part 2, I was also wondering about the best I/O setup for using a UMIK 1 microphone. I have an XLR based one but it's not calibrated.

I've read elsewhere on here that it's recommended to get the UMIK directly connected to AL via the advanced menu.

Would you recommend then having the input set directly to UMIK and output direct to ASIO (with ASIO4ALL pointed at the MINIDSP interface?)

1. Try setting the two subwoofers to mono, see image.
2. If REW has UMIK-1 selected, the other default settings should work.
Here is a link to our tech blog which may help you out: https://deercreekaudio.com/tech-blog/f/stereo-21-22-subwoofer-optimization
Deer Creek Audio is and Authorized miniDSP dealer: https://deercreekaudio.com/about

 

[Ofer]

Do you have Dirac installed? I also have the Minidsp Flex w/Dirac in my small office listening room (10’x11’ w/vaulted ceiling’s 8’-10’). This is definitely Not an ideal room, but I’m able to get pretty good results.

I initially was planning on using 2 subwoofers (SVS 3000 micros), but since I only have “1” chair I was able to just use (1) subwoofer. I just moved the sub to the area with the best response and let Dirac do it’s thing. Then I eq to my specific taste.

In a larger room, 2 subs definitely help even out the low frequency, but in a smaller room with 1 listening position 1 sub worked. You may want to try just 1 sub if you only have 1 listening position also.

Arnold

View attachment 55388View attachment 55389

Can you produce the step response before and after Dirac? preferably speakers and subs combined.

 

[trekker444]

Can you produce the step response before and after Dirac? preferably speakers and subs combined.

I‘m pretty new to using REW, miniDSP Flex and Dirac so I hope the attached impulse graphs what you are looking for.

Arnold

 

[Ofer]

If this is the impulse tab in rew then change dBFS to % and in the control (gear on the top right) tick "plot response normalized" then you need to zoom in on the peak. You can see Mitch audiolense walk through to see how it should look like.

 

[juicehifi]

This looks like an IR with dB view. It is not fit to show what Dirac does.