• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

miniDSP Flex

radix

Major Contributor
Joined
Aug 1, 2021
Messages
1,359
Likes
1,291
It's better than no correction, but I was not impressed at all. I can get better response and sound quality with DSP alone, and I think the problem here is that Dirac runs BEFORE the routing, so it don't split the subs from the mains and has no idea I have subs.

minidsp has several application notes on sub tuning, like this one. You should look at those, not just the flex docs. As was mentioned above, you want to set crossovers first.

This is from the Flex manual

Screen Shot 2022-01-29 at 7.31.26 PM.png
 

anotherhobby

Addicted to Fun and Learning
Forum Donor
Joined
Dec 17, 2021
Messages
623
Likes
1,319
minidsp has several application notes on sub tuning, like this one. You should look at those, not just the flex docs. As was mentioned above, you want to set crossovers first.

This is from the Flex manual

View attachment 183077
Yes, that's the diagram that shows Dirac does not have any control over the subwoofers in the Flex. Dirac only sees a right speaker and a left one. That led me to think it's probably better to EQ the subs independently before running Dirac.
 

radix

Major Contributor
Joined
Aug 1, 2021
Messages
1,359
Likes
1,291
Does anyone have the balanced flex and drive it with unbalanced analog? I assume that should work just fine, but at -6 dB?

I'd like to have usb -> balanced analog for a streamer, but also hookup a turn table preamp.

I currently have the SHD, but pretty much dislike the built-in streamer. The 30s or so startup time from standby just really annoys me.

Thanks,
Marc
 

anotherhobby

Addicted to Fun and Learning
Forum Donor
Joined
Dec 17, 2021
Messages
623
Likes
1,319
Does anyone have the balanced flex and drive it with unbalanced analog? I assume that should work just fine, but at -6 dB?

I'd like to have usb -> balanced analog for a streamer, but also hookup a turn table preamp.

I currently have the SHD, but pretty much dislike the built-in streamer. The 30s or so startup time from standby just really annoys me.

Thanks,
Marc
I just grabbed a pair of TS>RCA adapters I have and used them to connect my Topping D10s analog out to my balanced Flex to test for you. Works fine and is -6 dB compared to digital source (generating noise in REW, comparing dBFs in miniDSP plugin app). I don't have an analog balanced source to compare.
 

radix

Major Contributor
Joined
Aug 1, 2021
Messages
1,359
Likes
1,291
I just grabbed a pair of TS>RCA adapters I have and used them to connect my Topping D10s analog out to my balanced Flex to test for you. Works fine and is -6 dB compared to digital source (generating noise in REW, comparing dBFs in miniDSP plugin app). I don't have an analog balanced source to compare.
Thank you very much.
 

Ultrasonic

Addicted to Fun and Learning
Joined
Dec 30, 2018
Messages
742
Likes
593
Location
UK
When I ran it the first time with no crossovers, it was because I was following the Dirac manual which stated this for subwoofer settings:

If using a subwoofer in a multichannel system, we recommend that its low pass filter be disabled if possible, or set to its highest frequency if it cannot be disabled. Any EQ on the subwoofer should be disabled or set “flat.” High pass filters used to protect the driver from over-excursion should be left in place. If the subwoofer’s low pass filter is an important part of the overall bass management in the system, it can be re-enabled after completing Dirac Live calibration and loading correction filters into the miniDSP processor.

However, after after running it, seeing my results, and thinking about it a bit more, I realized that my 4 channels (2 speakers and 2 subs) is probably not what they mean by multichannel, and I'm assuming they meant 5.1 or whatever. Also, thinking a bit further, since Dirac is applied before the routing matrix, it has no idea about even having subs, or how many, and has to more broadly apply the filters to both the subs and the mains together. Granted with a crossover it should mostly land where you expect, but it clearly can NOT apply any logic to the subwoofers individually, especially since the subs are mono. I'm honestly a bit disappointed that Dirac is being done to before the routing and is completely ignorant of subs, but maybe I'm missing something.

What you have with the Flex is the stereo version of Dirac Live not the multi-channel version. The latter works with AV receivers and can include Dirac Live Bass Management which does handle subwoofers separately to aid their integration.

What Dirac Live 'sees/hears' in your system is two full range speakers - left main plus sub and right main plus sub. The optimisation that it calculates is to each of these. What miniDSP products gives you that other stereo Dirac Live products generally don't is the means to properly integrate the subwoofers with the mains independently of Dirac Live. This integration is though something that you need to do yourself with the aid of REW rather than it being automated. It is also not something that will not be covered by any Dirac Live manual.

Following the logic that it can't separate subs from mains in the Flex, I then EQ'd the subs individually first,

In my experience it's not really worth applying manual EQ prior to running Dirac Live as the final result will end up being essentially the same, but it takes more work. Here's a demonstration of this that I posted on the miniDSP forum (don't focus on what the overall response is like - I was just testing the concept):


So rather than PEQ, it is specifically the crossover filters and time delays that I'd be spending time on getting right in the miniDSP plugin. The other thing to have have adjusted prior to running Dirac Live is the subwoofer gain, which I'd suggest having slightly on the high side so that Dirac will be less likely to apply large boosts (the max. it will do is 10 dB by the way).

I'm really happy there are 4 preset slots so that I can try some more ways of doing things and maintain reasonable a/b testing. What I do know thus far is that the process I went through sounds really good. Better than EQ on my own, and way better than doing nothing.

Yes the ability to quickly swap between different filtering options is really handy, and it's great you're enjoying the results so far :). I'm just commenting to hopefully help you and others get even better results.

As for the crossover slope, I tried some other slopes between my bookshelf sized Revel's and the subs, and the slower roll off sounded more natural and smooth/blended to me, but it was a pretty short test and not scientific. I'm open to any tips and would like to hear your thoughts on other slopes and what advantages there might be. This is a very new setup of mine (never had a miniDSP or Dirac) and I haven't messed with it that much yet.

I would be genuinely surprised if a 12 dB/octave slope when applied as both a high-pass and low-pass filter pairing sounded best. The reason is that you'll get significant high-frequency sound sent to your subwoofer which is more likely to make it localisable / mess with imaging. You could play some music with just your sub on but not the amp for your main speakers to see what I mean.

Have you measured the responses of each main speaker and your subwoofer playing on their own? This is what I'd be looking at to decide on crossover settings. Where there could be an advantage of a slower roll-off I suppose is if there is a big dip in the sub response somewhere below the crossover frequency, and the main speakers then help to fill this in a bit. A different choice of crossover frequency would be another thing to consider if this is the case.

Unrelated to all of the above I'll just mention two key points with Dirac Live that sometimes get missed:
  • The default target curve is a starting point for adjusting to taste rather than some sort of universal 'best' that shouldn't be changed.
  • The range of measurement positions used will affect the results.
 

Ultrasonic

Addicted to Fun and Learning
Joined
Dec 30, 2018
Messages
742
Likes
593
Location
UK
I'll just add one other important general point regarding Dirac Live on miniDSPs that can get missed, which is that there is no attenuation automatically applied to prevent digital clipping from filters applied. Or at least there isn't on any of their previous models and I'd be vey surprised if the situation was any different on the Flex.

Generally people who use the Flex as their primary volume control won't need to worry, as it would only be if the master volume was above -10 dB that there might be an issue (assuming no PEQ gain in addition to Dirac Live). Typically peak boosts will be well below 10 dB provided a multi-point measurement is used to generate the DL filters too, so commonly issues won't occur until above more like about -5 dB. Anyone using the Flex at fixed gain would be best advised to manually set attenuation either in the plugin or via the master volume to provide the required digital headroom though. Or to not do so but be aware that if you get distortion this is the likely cause and solution.
 

MCH

Major Contributor
Joined
Apr 10, 2021
Messages
2,581
Likes
2,197
Very informative, ultrasonic. I know it is a queation with a different answer for each user, but what would you say, is dirac live worth it for the i think typical use of the flex (stereo w two subs?).
 

antcollinet

Master Contributor
Joined
Sep 4, 2021
Messages
7,408
Likes
12,291
Location
UK/Cheshire
Very informative, ultrasonic. I know it is a queation with a different answer for each user, but what would you say, is dirac live worth it for the i think typical use of the flex (stereo w two subs?).
Good question - I'm interested also.

In particular, can DIRAC do more than the non direc version plus REW and a bit of effort? Or is the effort a lot, and DIREC just saves time?
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Very informative, ultrasonic. I know it is a queation with a different answer for each user, but what would you say, is dirac live worth it for the i think typical use of the flex (stereo w two subs?).
It’s worth it if you have little experience with manual room correction and you’re not bothered spending months learning how things work.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
With the Dirac Flex you need to set the crossovers and delays yourself.

There are many many philosophies on how to determine the optimal crossover. Some aim for anechoicly flat response and some aim for in-room flat response.

There are several arguments too about the crossover frequency and slope and most people take many months of trial and error to get there.

For example for me my Speaker had a Butterworth 12dB/oct tuning @100Hz.

I used two (PEQ) filters to change that behavior to Butterworth 24dB /oct @ 100Hz and then i stacked another BW 24 dB/oct crossover from the crossover function in miniDSP.

This way resulted in a Linkwitz riley crossover of 48 dB/oct @ 100Hz which was absolutely perfect for me.

This whole procedure took me around 6 months to figure out how to do perfectly, but now i’m 100% confident with my integration.

If nothing I mentioned made sense to you, you’re better off buying an AVR with Dirac live bass control.
 

MCH

Major Contributor
Joined
Apr 10, 2021
Messages
2,581
Likes
2,197
Let me rephrase my question: would the minidsp flex with dirac live help us poor dsp illiterates to set up a stereo + 2 subs setup in a matter of hours or due to its limitations we would still need to do, as with the non dirac version, a several months long research and become very knowledgeable to achieve it?
 

Ultrasonic

Addicted to Fun and Learning
Joined
Dec 30, 2018
Messages
742
Likes
593
Location
UK
Very informative, ultrasonic. I know it is a queation with a different answer for each user, but what would you say, is dirac live worth it for the i think typical use of the flex (stereo w two subs?).

The question of whether Dirac Live is worth it is basically the same whether someone is using subwoofers or not, and my short answer would be yes for the vast majority because it makes everything quicker and easier.

Before my SHD I was using a 2x4 HD, which is essentially a Flex without the volume knob, display and Bluetooth connectivity. I bought the SHD principally because I wanted the display (makes a huge difference when used as a pre!) and better DAC and DSP performance. I also didn't think I really wanted the streamer functionality but in hindsight this is great to have too. If I could have bought a cheaper version of the SHD without Dirac I would have, taking the view I'd managed doing everything manually just fine on the 2x4 HD. Having now used Dirac I like it because it makes everything quicker and easier, and for people new to this I'd add it requires less knowledge/skill. There are times when I'd prefer the greater transparency of exactly what is going on that I get with REW but on balance it's Dirac Live that I choose to use.

When I was doing things manually with my 2x4 HD and REW I always just looked at the response at a single point. It's technically possible to do exactly the same with Dirac but it's very much designed to be used with data from a range of measurement positions, and I do think this makes more sense (and is vital for Dirac above low-bass frequencies). For starters nobody has both ears at the same location! It's possible to make measurements at multiple locations with REW and take an average but that would take more effort than using Dirac Live to do the same.

One thing that Dirac Live does than basic PEQ filters in REW don't is the phase correction aspect associated with improving the impulse response. In practice I find it impossible to judge whether this is beneficial or not as it's impossible to apply just the amplitude adjustment part separately to compare the two. It is also technically possible to use the FIR filter capability on the 2x4 HD/ Flex (without Dirac) to make phase adjustments too, although this is a level of complexity that few actually experiment with. (I did but only very briefly.)

I'll add one other point while I'm waffling on...

Sometimes you'll get people posting that the predicted response in Dirac is some sort of approximation because it doesn't match what someone measures at the MLP. What is being missed is that Dirac is working to achieve a particular average response over a volume, and the before and after plots in the software represent this. If someone does a Dirac correction based on a single measurement point and checks it at this point then everything will match up similarly to doing the equivalent in REW. The fact that Dirac is targeting a response over a volume also means that the MLP response will typically look less flat/smooth than the equivalent done in REW using a single measurement point. This means it doesn't look as impressive if posting a response plot on a forum, but it sounds better.
 

anotherhobby

Addicted to Fun and Learning
Forum Donor
Joined
Dec 17, 2021
Messages
623
Likes
1,319
What you have with the Flex is the stereo version of Dirac Live not the multi-channel version. The latter works with AV receivers and can include Dirac Live Bass Management which does handle subwoofers separately to aid their integration.

What Dirac Live 'sees/hears' in your system is two full range speakers - left main plus sub and right main plus sub. The optimisation that it calculates is to each of these. What miniDSP products gives you that other stereo Dirac Live products generally don't is the means to properly integrate the subwoofers with the mains independently of Dirac Live. This integration is though something that you need to do yourself with the aid of REW rather than it being automated. It is also not something that will not be covered by any Dirac Live manual.

Yep, this is additional confirmation to what I observed/realized and commented on earlier.

In my experience it's not really worth applying manual EQ prior to running Dirac Live as the final result will end up being essentially the same, but it takes more work. Here's a demonstration of this that I posted on the miniDSP forum (don't focus on what the overall response is like - I was just testing the concept):


So rather than PEQ, it is specifically the crossover filters and time delays that I'd be spending time on getting right in the miniDSP plugin. The other thing to have have adjusted prior to running Dirac Live is the subwoofer gain, which I'd suggest having slightly on the high side so that Dirac will be less likely to apply large boosts (the max. it will do is 10 dB by the way).

That's interesting. In my first go around, Dirac did much better after I did EQ, but I did not have crossovers set when I ran without EQ, which pretty much invalidates any conclusions, so I need to run that again and see what happens.

Yes the ability to quickly swap between different filtering options is really handy, and it's great you're enjoying the results so far :). I'm just commenting to hopefully help you and others get even better results.

I appreciate the feedback!

I would be genuinely surprised if a 12 dB/octave slope when applied as both a high-pass and low-pass filter pairing sounded best. The reason is that you'll get significant high-frequency sound sent to your subwoofer which is more likely to make it localisable / mess with imaging. You could play some music with just your sub on but not the amp for your main speakers to see what I mean.

This setup is in my office and my subs are almost directly under my speakers. Everything that comes from the subs sounds exactly like it's coming from the speakers, so it's nice that localization is not really a consideration here. When I had one sub on one side, I could totally localize, but not anymore.

IMG_9371.jpeg


Have you measured the responses of each main speaker and your subwoofer playing on their own? This is what I'd be looking at to decide on crossover settings. Where there could be an advantage of a slower roll-off I suppose is if there is a big dip in the sub response somewhere below the crossover frequency, and the main speakers then help to fill this in a bit. A different choice of crossover frequency would be another thing to consider if this is the case.
I didn't spend a lot of time comparing slopes, and the comparisons were before EQ/Dirac. This morning I made an identical config but changed the crossovers to a steeper slope and I might like it better. I'm definitely going to have to play around more and run some measurements. Thanks for the tips!
 

Ataraxia

Active Member
Joined
Sep 22, 2019
Messages
136
Likes
76
With the Dirac Flex you need to set the crossovers and delays yourself.

There are many many philosophies on how to determine the optimal crossover. Some aim for anechoicly flat response and some aim for in-room flat response.

There are several arguments too about the crossover frequency and slope and most people take many months of trial and error to get there.

For example for me my Speaker had a Butterworth 12dB/oct tuning @100Hz.

I used two (PEQ) filters to change that behavior to Butterworth 24dB /oct @ 100Hz and then i stacked another BW 24 dB/oct crossover from the crossover function in miniDSP.

This way resulted in a Linkwitz riley crossover of 48 dB/oct @ 100Hz which was absolutely perfect for me.

This whole procedure took me around 6 months to figure out how to do perfectly, but now i’m 100% confident with my integration.

If nothing I mentioned made sense to you, you’re better off buying an AVR with Dirac live bass control.

Have you ever compared the performance of built in Dirac, AVR, Arcam SA30, etc. to Dirac implemented with separates?
 

phoenixdogfan

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
3,294
Likes
5,070
Location
Nashville
I've tried a few things with setting up my Flex so far, and I thought I'd share my results. I'm also curious to hear what other people are finding works well. My setup is near field with Revel M105s and dual 8" subs in a small treated room (about 115 sq/ft with 8' ceilings).

My first attempt was to follow the manual exactly, and it turned out bad. It said to leave the sub crossovers off. My subs don't have crossovers (that's what the miniDSP is for), so they will try to play whatever you send them. This seemed like a bad idea. It also said to not have any DSP filters active at all. I don't know much about Dirac, so I did what they said and just let it rip. When it was done I set 80 Hz 12dB/oct crossovers and ran a sweep with Dirac on vs Dirac off and here is what I got:

View attachment 182918

It's better than no correction, but I was not impressed at all. I can get better response and sound quality with DSP alone, and I think the problem here is that Dirac runs BEFORE the routing, so it don't split the subs from the mains and has no idea I have subs. For my next run I sat down with REW for a couple of hours first and carefully crafted DSP filters on all of the 4 channels after setting the crossovers. This is where I landed compared to Dirac.

View attachment 182920

And not surprisingly to me, my DSP filters were better than Dirac, hands down. Not just in comparing measurements, but it sounds way better. With the room response better dialed in with DSP and crossovers, I then Dirac directly on top of this without disabling any of it, and here is how that turned out:

View attachment 182922

WAY better! Doing it this way, Dirac offers a noticeable improvement in the measurement. The nulls are handled much better, and the response is smoother overall. In subjective listening, I think it sounds way better with Dirac. I don't know how to describe the sound with flower audiophile review language, but the difference is dramatic and I'm very happy with the result at the moment. That said, I just finished this about 20 minutes ago and have a lot of listening to do.

Here is a final comparison of me setting DSP+Crossovers first then running Dirac vs just running Dirac on it's own with no DSP or crossovers set first. I did not attempt to DSP after running Dirac, but I might get around to trying that too.

View attachment 182929
My question is how much dynamic range did you trade off by using both REW and Dirac for DSP? Looks like you're playing about 78 db, is that system capable of playing 90db with all that dsp?
 

Ataraxia

Active Member
Joined
Sep 22, 2019
Messages
136
Likes
76
Why do we need to use REW with Dirac? I thought the Dirac sweep curves show the in room response just like REW.
 

Ultrasonic

Addicted to Fun and Learning
Joined
Dec 30, 2018
Messages
742
Likes
593
Location
UK
Why do we need to use REW with Dirac? I thought the Dirac sweep curves show the in room response just like REW.

It depends what someone is looking to achieve. With REW you know exactly what processing is being applied and it gives you loads more information. For starters you can't tell the absolute level referred to in the post above yours using Dirac Live.
 

Ataraxia

Active Member
Joined
Sep 22, 2019
Messages
136
Likes
76
It depends what someone is looking to achieve. With REW you know exactly what processing is being applied and it gives you loads more information. For starters you can't tell the absolute level referred to in the post above yours using Dirac Live.
Thanks, so one does not need REW to effectively implement Dirac, correct?

On that note have you compared Dirac performance results with Dirac separates (SHD, etc. + chosen amp) to Dirac integrated with units like Arcam SA30, Arcam AVR's, new Onkyo AVR's, NAD units etc.

Is it easier to calibrate with those units?

And, is the resulting sound performance as good as Dirac separates?
 

Ultrasonic

Addicted to Fun and Learning
Joined
Dec 30, 2018
Messages
742
Likes
593
Location
UK
Thanks, so one does not need REW to effectively implement Dirac, correct?

Correct. Although for someone looking to integrate a subwoofer it would be sub-optimal not to use REW to optimise the integration with the main speakers prior to running Dirac.

On that note have you compared Dirac performance results with Dirac separates (SHD, etc. + chosen amp) to Dirac integrated with units like Arcam SA30, Arcam AVR's, new Onkyo AVR's, NAD units etc.

Is it easier to calibrate with those units?

And, is the resulting sound performance as good as Dirac separates?

My only experience with Dirac Live has been on my SHD.
 
Top Bottom