Minidsp Flex Review (Audio DSP)

[AdamG]

Thread Notice: Friendly Reminder that this is an Official Product Review Thread. Please keep your comments and posts directly related to the Product and/or the test bench results/data. Thank you kindly for your understanding and assistance.

 

[anotherhobby]

Curious about the (quite good!) analog input performance: does anybody know what ADC & DAC chips are used? I believe the Sharc ADSP-21489 DSP is pure digital, correct?

This was posted earlier:

ADC chip: TI PCM4202 https://www.ti.com/lit/ds/symlink/pcm4202.pdf

DAC chip(s): TI PCM1795 https://www.ti.com/lit/ds/symlink/pcm1795.pdf

 

[rdenney]

So, I'm trying to figure out how this device would work, given my unnecessarily but pleasingly complicated setup with a lot of analog stuff in it.

(Adam, I think this is on topic--it's about how this device gets integrated.)

Here's my thinking:

--Thorens table > B&K preamp phono input > Benchmark ADC1 USB (via Rec Out bus) > Vinyl Studio in laptop (via USB).
--Thorens table > B&K preamp phono input > (line amp bypassed, a feature of the B&K) > MiniDSP device via analog stereo input > Buckeye + (future) Sub plate amps > Speakers
--Thorens table > B&K preamp phono input > JDS Atom headphone amp (via Rec Out bus) > headphones
--Analog line source (tuner, vintage CD player, tape decks routed through dbx 400 line router, whatever else I'm playing with at the moment) > B&K line inputs > (line amp bypassed) > MiniDSP device via analog stereo input > Buckeye + Sub plate amps > speakers
--Computer playback > MiniDSP device via USB input > amp/speakers
--Digital sources (mainly, CD transport) > MiniDSP device via Coax input > amp/speakers

The only route I can't figure out is how to get digital sources into the headphone amp. Currently, the digital sources go at line level into the preamp as line-level sources, and then into the head amp via the Rec Out bus.

The input to the MiniDSP device would be the main outputs of my B&K preamp. With the line amp bypassed, I would move the volume on that device to full (which, by the way, would eliminate the Johnson noise from that pot), and the voltages at the outputs of the preamp would be the same as at the inputs from the sources. The MiniDSP could handle that fine, and add enough preamplification to drive my Buckeye amp to full power.

Come to think of it, if I do that, then I can run the second set of main outputs from the B&K preamp to the Atom headphone amp. That would present line level to the head amp, and all is well. But I would lose the ability to use the MiniDSP to provide corrective EQ for the headphones--which I don't now do anyway. (My headphones include the AKG K371 and the Sennheiser/Drop HD6XX, neither of which are unacceptable without EQ, and my Sennheiser RD185 wireless headphones, which work from an optical digital input anyway.

One advantage to this setup is that for others using the system, I can constrain the volume on the MiniDSP and let people use the big knob on the B&K. That would keep them from blowing things up--though I think my Revels can take whatever the Buckeye can give it if not overdriven.

I have too many line-level analog sources and too much interest in old analog stuff to eliminate the B&K preamp altogether. Of course, I could put in the processor loop, but then I would lose the sub integration and still be limited to the line amp in the B&K preamp, which is good but not as good. That may be my starting point, however.

Rick "thinking out loud" Denney

 

[phoenixdogfan]

That's what BluTak is for, no?

Or velcro.

 

[anotherhobby]

Regarding using unbalanced input adapters on the balanced Flex, I noticed miniDSP recommends specifically not doing this due to a 15 dB SINAD drop, so be aware that while it'll work, they do call out a performance issue when doing so:

WARNING: This connectivity work (TRS to RCA) workaround is ONLY to be used for outputs. We've noticed a performance drop on outputs (15dB SINAD drop) for that reason will not recommend customers to use an unbalanced to balanced cable on the "input" side of the Flex. If you want to have unbalanced in/out for both inputs and outputs, it's strongly recommended that you just use the miniDSP Flex unbalanced version. The miniDSP Flex was engineered with best performance in mind and it's therefore our intention to stay consistent on making sure best audio quality is kept for this platform!

 

[Sam Ash]

I just use a cheap laptop--some HP plastic cheapie that I bought on sale at Best Buy. I was about to abandon it, because the fan started getting noisy. But I found a replacement fan on Amazon for $8 and it only took ten minutes to replace it, and blow out all the cat hair that was making it noisy in the first place. It's silent again. I think I'll buy several of those fans for the future--by the time I use them up, I'll have to replace the thing anyway. It connects to my home network via WiFi, and the Music Bee player software I use can access the library on my home server. I suppose it could also access the online streaming service.

The good news is that I can also run Vinyl Studio on it for making needledrops, and it works fine running REW, Audacity, and Presonus DAW software, at least for the simple things I do.

With a device like this, I can still use my Benchmark ADC (which has superb front-panel indications and controls) for making needledrops, and use this as the input for other devices.

Rick "cheaper than many streamers, even with the price of the DAC" Denney

Thank you for that detailed response. What do you think about Audirvana?

 

[TheBatsEar]

So Darko was right? Does it improve the tone though?

I have no doubt it does for some.

 

[DHT 845]

Nice device giving many possibilities. I could build onwall line arrays plus woofers in my bedroom with it ...

 

[rvsixer]

Regarding using unbalanced input adapters on the balanced Flex, I noticed miniDSP recommends specifically not doing this due to a 15 dB SINAD drop, so be aware that while it'll work, they do call out a performance issue when doing so:

WARNING: This connectivity work (TRS to RCA) workaround is ONLY to be used for outputs. We've noticed a performance drop on outputs (15dB SINAD drop) for that reason will not recommend customers to use an unbalanced to balanced cable on the "input" side of the Flex. If you want to have unbalanced in/out for both inputs and outputs, it's strongly recommended that you just use the miniDSP Flex unbalanced version. The miniDSP Flex was engineered with best performance in mind and it's therefore our intention to stay consistent on making sure best audio quality is kept for this platform!

Good post, thanks. I wonder if this is due to comparing 4V balanced vs 2V unbalanced input levels?

My own use case for Flex device(s) would be for my music/HT setup (colocated but separate 2-ch and AVR systems, up to the shared 2.x preouts to shared amps/speakers/subs). If I use a 4V clean unbalanced out AVR (i.e. one of many Denon's, etc.) that already are in sub-100dB SINAD range, the Flex unbalanced SINAD drop even if either 2V or 4V level should be of no audible consequence?

 

[eardiggler]

It tends to be confusing...in some (most?) cases it would be better to buy the unbalanced version if you are using unbalanced inputs, then connect the unbalanced output to the balanced input on the amp/monitors via properly terminated cables. This is frequently misunderstood (most often by me)...there is no real downside feeding the unbalanced output to the balanced amp input.

Doing the conversion on the input side can be more problematic (though of course, not always).

MiniDSP recommends purchasing the unbalanced model for mixed use cases like this.

---
View attachment 185919

Thanks for posting this!
I just cancelled my order and purchased the unbalanced version. The appeal to me is being able to plug in my old analog stuff (with a switcher) and use this as an analog preamp and dsp/dac all in one.

 

[phoenixdogfan]

Good post, thanks. I wonder if this is due to comparing 4V balanced vs 2V unbalanced input levels?

My own use case for Flex device(s) would be for my music/HT setup (colocated but separate 2-ch and AVR systems, up to the shared 2.x preouts to shared amps/speakers/subs). If I use a 4V clean unbalanced out AVR (i.e. one of many Denon's, etc.) that already are in sub-100dB SINAD range, the Flex unbalanced SINAD drop even if either 2V or 4V level should be of no audible consequence?

If I made use of this, I would need balanced outputs to my main speaker amp (Pufifi) and unbalanced to my subs (SVS), so I would probably use something like a Benchmark Balanced to RCA with a floating ground for the sub channels. I would need a balanced TRS to Balanced XLR cable to connect the XLR end of the Benchmark cable.

 

[The Jniac]

Hot damn, after that Emotiva thing, it is nice to see a company putting real effort into developing a product that performs well. A splendid show that is all the more welcome given the price point.

 

[eardiggler]

My thoughts on this are similar to the response you got from Feynman.
Full scale input and output on the RCA version is 2.0 Vrms
Full scale input and output on the Balanced version is 4.0 Vrms.

Consider the output levels of your legacy RCA equipment and to what extent of full scale they are able to drive the analog inputs of the Flex.
For example, the Zone 2 output of my Yamaha AVR drives at 200 mV. Hooking them up directly, I'm going to be throwing away a lot of SNR / ADC bits and won't be getting a lot of volume out of my amp. I haven't tried it yet, but suspect I will need to find a gain stage to put in between. It would be better to attenuate balanced outputs into RCA inputs in my opinion.

Thank you for explaining this. Still learning a lot here. So, unbalanced is really the way to go if one wants to plug their old analog gear in. The SNR will get chopped down on the TRS version when using single ended (RCA) inputs.

 

[abdo123]

For example, the Zone 2 output of my Yamaha AVR drives at 200 mV.

you will lose roughly 18dB of S/N. but since i already run my master volume at -20 dB with no hiss whatsoever i don't think it's an issue at all.

 

[Ultrasonic]

Yes. MiniDSP has free PC software that connects to the Flex via USB where you can set LP and HP crossovers for all 4 channels as well as time delay, etc. It's great for bass management with up to two subs. Dirac works like a next step, additional software functionality that generates a target curve (that you can modify) based on measurements of the room.

I'd add that a Flex without Dirac can still be used to adjust the in-room response to match a target curve, by using REW to help calculate appropriate PEQ filters that are then uploaded to the Flex.

 

[Ultrasonic]

OK does this mean that I can shove the FIR tap into the unit?

You have a choice of either having Dirac Live functionality or sticking to the cheaper Flex without Dirac Live and having manually adjustable FIR filter capability. You can't have both.

I have always thought of Dirac more like speaker correction and the PEQ stuff as room correction.

I'd suggest more the opposite personally. I've used the PEQ filters on other miniDSPs to apply speaker correction based on quasi-anechoic measurements, whilst I'd view Dirac Live as principally software to adjust the in-room response. Yes it does its impulse correction thing as well, although I find it impossible to judge its value since you can't choose to turn this aspect on/off in isolation.

Personally I like the results of using Dirac Live just in the low-bass region rather than over the full frequency range. Others like the latter but I would encourage anyone to experiment with both to see what they enjoy most.

 

[dziemian]

Looking forward to 8 channel (or more) version

 

[rvsixer]

If I made use of this, I would need balanced outputs to my main speaker amp (Pufifi) and unbalanced to my subs (SVS), so I would probably use something like a Benchmark Balanced to RCA with a floating ground for the sub channels. I would need a balanced TRS to Balanced XLR cable to connect the XLR end of the Benchmark cable.

Thanks, but I am not seeing any unbalanced signal to Flex balanced input connection (and the associated Flex SINAD drop question) in this scenario at all.

 

[Holmz]

Thanks @Ultrasonic a couple more questions/comments .

You have a choice of either having Dirac Live functionality or sticking to the cheaper Flex without Dirac Live and having manually adjustable FIR filter capability. You can't have both.

Ok, so if I want to use something like rePhase or Eclipse’s-FIR Creator or FIR-Designer (or something else), and I can show the FIR taps in with only the cheaper option?

And with Dirac I cannot get the taps/weights out? And can also get the weights pushed in?
or is it… why would I want Dirac if I only want to push the taps/weights in?

I'd suggest more the opposite personally. I've used the PEQ filters on other miniDSPs to apply speaker correction based on quasi-anechoic measurements, whilst I'd view Dirac Live an principally software to adjust the in-room response. Yes it does its impulse correction thing as well, although I find it impossible to judge its value since you can't choose to turn this aspect on/off in isolation.

Personally I like the results of using Dirac Live just in the low-bass region rather than over the full frequency range. Others like the latter but I would encourage anyone to experiment with both to see what they enjoy most.

I would like to use it for a 2-way XO, and also the sub’s XO.
But… I want to ONLY do the 2-way speaker’s frequency response to get it to what it would be in an anechoc sense, as well as to get the phase aligned.
For the sub, I would like to get any group-delay removed, should it need to be.

I also have an AVR with a room correction, which I ‘could’ use to do the low frequency room correction.

I would not really want to do room correction beyond, say, 300-400 Hz, but I would like the step function response and impulse response to good all the way up, without really rolling the room into the higher freqs.


Lastly if the FIR is applied to the whole channel first… and then XO bandpass, and other PEQ is applied later as PEQs… then it sounds like I would use two units in a horizontal fashion? (One Flex for the Pari of 2-ways. And another Flex for the pair of subs.)

 

[Ultrasonic]

Ok, so if I want to use something like rePhase or Eclipse’s-FIR Creator or FIR-Designer (or something else), and I can show the FIR taps in with only the cheaper option?

Yes, although be aware the number of taps available is somewhat limited. (I forget the numbers but I think I remember someone posting this above.)

And with Dirac I cannot get the taps/weights out? And can also get the weights pushed in?
or is it… why would I want Dirac if I only want to push the taps/weights in?

You cannot view or edit the Dirac Live filters directly. Changes are made via adjustments to the target curve and the only way to 'see' the actual filters is to measure them.

I would like to use it for a 2-way XO, and also the sub’s XO.

That would take 6 independent output channels rather than the 4 that the Flex models have. Some have daisy-chained a couple of 2-in 4-out miniDSPs to give the required 6 channels but this is not something I've tried myself to comment on directly.