Minidsp Flex Review (Audio DSP)

[mike70]

I activated my license today. And performed a measily single measurement and exported an automatic DIRAC filter.

I should have started experimenting with room correction before. My preliminary opinion (based on a "bad" single measurement filter) is that a lot of phase magic happens. I have always had relatively small listening rooms, and not enough room to take the speakers more than 20cms off the wall. DIRAC seems to make a big difference, mostly a lot of phase magic, the soundstage is more to my liking. e.g. Until now I've not always liked that the music was very cleary coming from my 2 speakers, I was able to localise them easily, now though, the sound is more detached from the speakers. And yes, some frequency response sounds altered, I don't really mind, doesn't sound better or worse for me, just different.
I don't like it on all songs, but this might also be due to my "bad" filter. I really appreciate the DIRAC button on the remote so you can switch instantly without interruption.

That sound stage though
I should have started experimenting with room correction before.

People needs to understand that they don't need to try to "equalize / process" sound buying more expensive amplifiers / cables ... they need to work with acoustics / real DSP.

 

[tvih]

People needs to understand that they don't need to try to "equalize / process" sound buying more expensive amplifiers / cables ... they need to work with acoustics / real DSP.

But how can they feel smugly superior about their needlessly expensive gear that way?

 

[mike70]

But how can they feel smugly superior about their needlessly expensive gear that way?

well, yes, for some audio is only other way to show status ... there's no arguments / reasons for them, fancy and expensive is the way to go.

Money is the best way to get bad kins and friends

 

[phoenixdogfan]

People needs to understand that they don't need to try to "equalize / process" sound buying more expensive amplifiers / cables ... they need to work with acoustics / real DSP.

Using components like "synergistic" amps with your speakers instead of EQ and DSP is like eating a steak with a screwdriver.

 

[Keened]

Using components like "synergistic" amps with your speakers instead of EQ and DSP is like eating a steak with a screwdriver.

I think it's kind of like the romanticism with complicated wood joinery, if done correctly you can achieve a beautiful 'pure' piece of art, but it take a very high level of skill to get results better than using metal hardware and it can't pack flat and be reused in new configurations.

In some ways I think it is a good mastery test for a full stack audio production engineer:

Can you understand room interactions and calculate optimal positioning purely from theory?
Can you create a speaker and case with desired properties tailored to that placement?
Can you design an amplifier with explicitly synergistic properties that take into account the room and speaker design?

You'd have to really grasp the underlying physics in multiple sub-domains to make it work right and know the explicit parameters you'd be designing for.

 

[TreveK]

I activated my license today. And performed a measily single measurement and exported an automatic DIRAC filter.

I should have started experimenting with room correction before. My preliminary opinion (based on a "bad" single measurement filter) is that a lot of phase magic happens. I have always had relatively small listening rooms, and not enough room to take the speakers more than 20cms off the wall. DIRAC seems to make a big difference, mostly a lot of phase magic, the soundstage is more to my liking. e.g. Until now I've not always liked that the music was very cleary coming from my 2 speakers, I was able to localise them easily, now though, the sound is more detached from the speakers. And yes, some frequency response sounds altered, I don't really mind, doesn't sound better or worse for me, just different.
I don't like it on all songs, but this might also be due to my "bad" filter. I really appreciate the DIRAC button on the remote so you can switch instantly without interruption.

That sound stage though
I should have started experimenting with room correction before.

Agreed - soundstage is the biggest improvement for me in my nearfield desktop setup.

 

[dougi]

I just did my first DIRAC calibrations for my Flex Digital (with ADI-2 pro). The first one wasn't good, with 7.8ms delay applied to one speaker! The second was better (0.67ms, tightly focussed setup). Yes the soundstage change was quite noticable. I did reduce the bass correction range a bit as it was pushing the speakers down low a lot (+10dB at 25 Hz so lifted correction to 35 Hz). It certainly changed the impulse response and phase. Not sure it looks an improvement on paper.

 

[dougi]

Why not?

How so?

Yes, it alleviates the artefacts produced by the reconstruction filters in DACs such as ringing and time smearing which I've already mentioned. Have you ever compared decent hi-res to CD quality on a decent DAC? You might want to try it!

Correct me if I'm wrong but I believe that wasn't using Dirac, so was 96KHz not 48KHz.

MQA is definitely an advantage for 44.1KHz content as far as I can hear, and although I'm still on the fence about its advantage in hi-res I feel it has a slight advantage there too. Judging by your comment it seems you've never listened to it on a good DAC that supports it

I have just got a Flex Digital and am running it into a RME ADI-2 PRO. With no SRC and sample rate source set to auto in the RME, mine is showing 96kHz even with DIRAC on. Surely the Flex is not doing DRC on both input and output? I suppose I could sweep the Flex using the ADI-2 pro to confirm the bandwidth supported under DIRAC.

 

[Harryharryharry]

I was thinking of a hack to get 6 channels out of the flex and wondered if there was a reason not to...

Planning on bi amping my Kef R3s with two Topping PA5s, the question would be how to keep the sub involved.

Simple Y splitter on the bass left and right outputs could let me get a signal to my sub, which has its own DSP so I could do some basic integration.

Would it work or is it greedy!?

 

[tonycollinet]

It will (probably) work.

Of course, you are not getting 6 channels, simply splitting two of the channels to two speakers each. You will have to match the LPF for the subs to the natural roll-off of the bass units in the R3's.

 

[MarcosCh]

This is interesting. Do the connections for hf and lf of the kef r3 defeat the speaker xover?

 

[Chagall]

It will (probably) work.

Of course, you are not getting 6 channels, simply splitting two of the channels to two speakers each. You will have to match the LPF for the subs to the natural roll-off of the bass units in the R3's.

This kind of defeats the purpose of flex, IMO. You want to send high passed freq to both amps (if bi-amping) and low freq to sub.
If anything, splitter should be on the flex outputs 1 and 2 that goes to amps and to the main speakes - I don't know if this will work.

This is interesting. Do the connections for hf and lf of the kef r3 defeat the speaker xover?

No, speaker passive xover works in that case also.

 

[Absolute]

Quick question, can you do crossover and speaker correction AND dirac afterwards?
Any limitations, like only output EQ or is it possible to do both input and output EQ plus Dirac afterwards?

 

[Chagall]

Quick question, can you do crossover and speaker correction AND dirac afterwards?
Any limitations, like only output EQ or is it possible to do both input and output EQ plus Dirac afterwards?

I didn't do separate speaker correction - basically Dirac does that for you depending on house curve you load.
Crossover and sub time aligment was done to a clean (noEQ signal and no room correction) in REW and ran dirac after.

But you can do EQ before dirac, not sure what would be the benefit though.

Bare in mind I have Dirac on PC and use minidsp 2x4 HD, but same logic would apply to flex with dirac.

 

[Tangband]

Hi,
before I pull the trigger, could really use some advice. Have read the whole 50 pages of this, but still...

My music system is Node to Topping e50 to Emotiva a300, powering modified Klipsch rp600m (biwired, yay!) and having 2 BK subs connected to it with high level ( stereo subs, like REL). Speakers are shared between this and avr through Beresford amp switch, subs lfe from avr, high level from emotiva.

Although my room is well treated (waterfall graph is pretty good as I understand it, only that nasty peak close to 400 ms and 30 hz and below more than that) , still having some modes, mains have a big dip from 55 to 80, like 20 db down, while their 50-55 are 15ish db up. Audyssey works really well down low, so I'd like to achieve similar or better results, but without avr in the chain (hence music only system).
Tried placement, phase, xover etc, so finally grew tired of it and would go for something like flex.
Really digging the xover and time alignment part Flex analogue offers, as the dac looks good and I think I don't lose audible quality vs my Topping. Feel free to chime in.

Is this the right way to do this:

Subs have two rca in - at the moment avr using both inputs with 1 to 2 rca cables. High level cable from emotiva to sub, which would become obsolete with Mini dsp.
Can I use my subs' rca inputs like this:
-one cable from avr to only one input on sub?
-one cable from flex to the second input on sub?
-try not to have both avr and Flex playing at the same time ?
-Subs don't care one input is dead?
-phase to 0, gain as for avr, xover max?

So, Node to Flex over coax, then pair my left channel with one output (number 1?) to amp's left input and one to the left sub (number 3?) ? Sub becomes extension to mains, not like mono lfe to both subs? Same for right channel. Set crossovers and then get into the rabbit hole of tweaking?
Oh, I'd use Dirac live, seems simpler.

Also - I'm from Croatia, EU, Audiophonics asks for 700 euros without dirac, can someone confirm how much would it cost me / or no customs applied, if I order it straight from USA? Flex + dirac is 694 dollars on their website, convert to euros + shipping, and then VAT 25%, + customs...or what is the formula? I saw one member from Greece here, ...

I'm at work now, so probably left some things out, so I'll have to ask later, sorry :/.

And thanks for all input, in advance,

Best


Edit: how would flex digital work in my case, considering the subs? Additional dac, and then how do I control volume? Let's say I get some cheaper topping. One topping rc to rule them all?

To still have passive crossovers in your main loudspeakers, just using this minidsp flex to make a digitalcrossover for the subs and main loudspeaker is really a big waste of money. For real, big improvements you really should go active in every register, putting the passive crossover away gonna make the sound much better. But of course you need measurement tools to do this.

 

[Absolute]

I didn't do separate speaker correction - basically Dirac does that for you depending on house curve you load.
Crossover and sub time aligment was done to a clean (noEQ signal and no room correction) in REW and ran dirac after.

But you can do EQ before dirac, not sure what would be the benefit though.

Bare in mind I have Dirac on PC and use minidsp 2x4 HD, but same logic would apply to flex with dirac.

Ideally I would want to create linear phase crossovers and anechoic speaker tuning first so that I could experiment effectively whether or not Dirac degrades the sound if used fullrange. This is my experience with all room corrections above the lower mids, but it's difficult to A-B effectively when there's a 10 second delay between turning on and off the correction by using my pc for it.
I doubt there's enough processing power in the Flex to do both FIR crossovers and then Dirac on top of it, so I would do the next best thing - regular crossover with speaker tuning before Dirac.

Knowing how many EQ points is available is helpful. 10 per channel on both input and output plus separate crossover filters?

 

[Harryharryharry]

This kind of defeats the purpose of flex, IMO. You want to send high passed freq to both amps (if bi-amping) and low freq to sub.
If anything, splitter should be on the flex outputs 1 and 2 that goes to amps and to the main speakes - I don't know if this will work.



No, speaker passive xover works in that case also.

Hmm good points.

I think my plan fails because whilst I could high pass the sub on its own DSP that would mean I couldn't be able to low pass the kefs (though maybe the bi amp benefit would outweigh this cost?).

My understanding is bi-amping is only worth while if you crossover before the amps, but I don't have much knowledge at all on this topic.

 

[Wikedlok]

To still have passive crossovers in your main loudspeakers, just using this minidsp flex to make a digitalcrossover for the subs and main loudspeaker is really a big waste of money. For real, big improvements you really should go active in every register, putting the passive crossover away gonna make the sound much better. But of course you need measurement tools to do this.

I do agree, this is not cheap, but I do have the means at the moment. Went for flex digi, still in delivery, while Topping e30 arrived. I intend to use xovers for high pass to mains and low pass for subs, and mini dsp peq for starters, if not satisfied, I might go for Dirac.
Could you explain what do you mean, removing crossover from speakers, and what then? Dac for tweeter and dac for woofer and sub?

 

[Tangband]

I do agree, this is not cheap, but I do have the means at the moment. Went for flex digi, still in delivery, while Topping e30 arrived. I intend to use xovers for high pass to mains and low pass for subs, and mini dsp peq for starters, if not satisfied, I might go for Dirac.
Could you explain what do you mean, removing crossover from speakers, and what then? Dac for tweeter and dac for woofer and sub?

Read more about making a passive loudspeaker active - before you start .
You must know how to do it. You can destroy your tweeter if youre not careful.
If flex digi only have digital outputs then yes- you need a dac for every driver .

 

[Wikedlok]

Read more about making a passive loudspeaker active - before you start .
You must know how to do it. You can destroy your tweeter if youre not careful.
If flex digi only have digital outputs then yes- you need a dac for every driver .

There are also active alternatives much better than passive Klipsch you can buy .
Adam a7x and Genelec 8030c or 8330, 8040 is four good examples.

Maybe I misunderstand, but my way I split the signal at 80 Hz to mains and subs. The way you're suggesting I would have the woofers play full range up to let's say 2k where tweeter takes over? Meaning woofer and sub are getting eq-ed as a unit and woofer playing to 40ish.
This is not my intention nor need, as placement dictates mains have a hole in punch area, while subs do not, so I do want a crossover at 80ish. This is actually the main reason I went to dsp, as I like everything else coming from speakers. Btw, they are modded like gr research does, but with better parts, FR is great now.