Minidsp Flex Review (Audio DSP)

[Ultrasonic]

Do they mean 4 channel input here. I thought there were only stereo analogue inputs. Or do they mean you can record the result of the DSP/crossover/routing. If so that's sort of neat.

I believe miniDSP meant what is written, which is consistent with your suggestion of it being a way to monitor the four output channels. I think this is possible. (There are indeed only two input channels i.e. stereo.)

With my SHD I've used similar functionality as a way to quickly see what the applied Dirac Live filters look like, but this was via REW which only makes use out output channels 1 and 2. I've not tried to see if it is possible to access outputs 3 and 4 as well via USB.

 

[Stephen]

This may be because you've never experienced a subwoofer properly integrated by means of a device like the Flex .

It was using a miniDSP 2x4 HD that made me realise my natural 'audiophile' scepticism over using subwoofers for music was misplaced.

It's maybe psychological: i despise to detect the sub kick in (maybe I inconsciously try to spot it...). And you need 2 subs for stereo... no definetly I prefer large band speakers.

 

[Ultrasonic]

It's maybe psychological: i despise to detect the sub kick in (maybe I inconsciously try to spot it...). And you need 2 subs for stereo... no definetly I prefer large band speakers.

A properly integrated subwoofer behaves no differently to a properly integrated bass-driver in a large speaker. There is no 'sub kick in', whatever that may be, in my system. The functionality of a device like the Flex is key to achieving this though.

Edit: subwoofer integration is the number one reason I use miniDSP products.

Edit 2: and you don't need 2 subs for stereo.

 

[antcollinet]

no that's not how 32bit floating point processing works, no dynamic range is lost when amplifying, whatever you had before within the limits of the Flex's ADC remains till the output stage.

Nothing clips unless something goes above 0dBFS at the very end of the chain.

If you take an input with a noise floor of around -120dB (20 bits DR), inject a 4mv signal into that, and then amplify it by 60dB to get to the 4V equivalent for full range balanced input, then you are also amplifying the noise by 60dB (Nothing to do with clipping and it makes no difference whether you amplify in the analogue or digital domains). So your noise floor goes up also by 60dB. Much higher than the 25dB benefit you get over a very good preamp. This is simply due to the dynamic range of the ADC, of whch, with a 4mV signal compared with a 4V full scale input you are only using a tiny part.

 

[Stephen]

A properly integrated subwoofer behaves no differently to a properly integrated bass-driver in a large speaker. There is no 'sub kick in', whatever that may be, in my system. The functionality of a device like the Flex is key to achieving this though.

Edit: subwoofer integration is the number one reason I use miniDSP products.

The other problem is to set these subwoofers in a living room. I have Revel performa B15 in the home theater: 40cm3, 50kg, to set at some distance from the walls.... So selecting large band speakers are my best option in my living room. And they were difficult to choose and find.

 

[Pdxwayne]

What aspect were you interested in? I doubt getting the crossover shapes right it at all demanding and will just be 'right'?

I've used crossover filters on miniDSPs for subwoofer integration. Given these are applied to in-room speaker and subwoofer responses that are a long way from flat, small variations from perfect filter shape wouldn't really matter. What would matter I suppose is significant 'leakage' of sound beyond the intended fall-off but I have seen no evidence of this being an issue myself.

I appreciate crossover performance would be more critical between higher frequency drivers in an active speaker but I'd still be pretty shocked if the Flex's performance wasn't easily 'good enough'.

In the Onkyo thread in https://www.audiosciencereview.com/...50-review-home-theater-avr.30842/post-1087191

for analog in, if not in pure mode, it losses 6db sinad. I wonder adding crossover will degrade sinad.further for analog in.

 

[Ultrasonic]

In the Onkyo thread in https://www.audiosciencereview.com/...50-review-home-theater-avr.30842/post-1087191

for analog in, if not in pure mode, it's losses 6db sinad. I wonder adding crossover will degrade sinad.further for analog in.

The analogue inputs on the Flex will always be converted to digital. Does the 'pure mode' on the Onkyo prevent this? If so there is no equivalent to test for here.

 

[Pdxwayne]

The analogue inputs on the Flex will always be converted to digital. Does the 'pure mode' on the Onkyo prevent this? If so there is no equivalent to test for here.

Hmm, so are we all expecting that with and without crossover, flex will have the same sinad?

 

[Ultrasonic]

Hmm, so are we all expecting that with and without crossover, flex will have the same sinad?

I'm interested in what others think but that would be my guess. Assuming we're assessing SINAD at a frequency that isn't attenuated by the crossover filter of course.

 

[PeteL]

The analogue inputs on the Flex will always be converted to digital. Does the 'pure mode' on the Onkyo prevent this? If so there is no equivalent to test for here.

There are literally hundreds of analog preamps out there, I am not sure why would they want to put an analog input on this that would bypass all DSP capabilities? The idea is to have DSP on analog source too isn't it?

 

[mdsimon2]

I'm interested in what others think but that would be my guess. Assuming we're assessing SINAD at a frequency that isn't attenuated by the crossover filter of course.

I agree. As mentioned a few times in this thread you can record the digital signal that has been processed by the DSP using the bi-directional USB audio so it is easy to tell what if any degradation has occurred from applying DSP.

Here are some measurements from a 2X4HD. What I have seen is that applying low frequency EQ / x-overs does cause a slight increase in noise but you will still be limited the DAC analog performance not any digital noise added by the DSP. I have not seen the same increase in noise when applying higher frequency EQ / x-overs, in fact some times it goes down because the noise floor is attenuated by the x-over.

2X4HD, no DSP. -127.8 THD+N.

2X4HD, 80 Hz LR4 HPF. -122.4 THD+N.

2X4HD, 5000 Hz LR4 LPF. -130.5 THD+N.

Michael

 

[Ultrasonic]

There are literally hundreds of analog preamps out there, I am not sure why would they want to put an analog input on this that would bypass all DSP capabilities? The idea is to have DSP on analog source too isn't it?

Not quite sure why you're asking me this question? It would indeed largely defeat the point of the Flex to bypass the DSP capabilities. I can see there might be a market for those who only wanted to apply DSP to digital but not analogue sources but adding in a bypass ability to support this would add to the cost of the unit, and would require some other device to control volume as well.

 

[Pdxwayne]

I agree. As mentioned a few times in this thread you can record the digital signal that has been processed by the DSP using the bi-directional USB audio so it is easy to tell what if any degradation has occurred from applying DSP.

Here are some measurements from a 2X4HD. What I have seen is that applying low frequency EQ / x-overs does cause a slight increase in noise but you will still be limited the DAC analog performance not any digital noise added by the DSP. I have not seen the same increase in noise when applying higher frequency EQ / x-overs, in fact some times it goes down because the noise floor is attenuated by the x-over.

2X4HD, no DSP. -127.8 THD+N.
View attachment 186396

2X4HD, 80 Hz LR4 HPF. -122.4 THD+N.
View attachment 186395

2X4HD, 5000 Hz LR4 LPF. -130.5 THD+N.
View attachment 186394

Michael

Hmm, charts look really good. I wonder what could cause the loud hiss when using minidsp 2x4 HD in my chain?

 

[abdo123]

Hmm, charts look really good. I wonder what could cause the loud hiss when using minidsp 2x4 HD in my chain?

The measurements were digital in digital out, not digital in analogue out.

 

[Pdxwayne]

The measurements were digital in digital out, not digital in analogue out.

I see. I was using analog in and out....Thus kind of curious about flex's sinad when doing the same: analog in and out with crossover enabled.

 

[mdsimon2]

Hmm, charts look really good. I wonder what could cause the loud hiss when using minidsp 2x4 HD in my chain?

Read my post, this is digital only before the 2X4HD DAC.

This is why you have hiss, 20.5 uV residual noise measured at the analog output which is a dynamic range of 20 x log (2 x10^6 / 20.5) = 100 dB. If you have higher gain amplifiers and/or sensitive speakers this will hiss. Using analog input will only make this worse.

Michael

 

[Ultrasonic]

Hmm, charts look really good. I wonder what could cause the loud hiss when using minidsp 2x4 HD in my chain?

Gain structure is very important, and note that the noise floor on the 2x4 HD is higher than the Flex. SINAD relates to maximum signal levels - if your amplifier gain / loudspeaker sensitivity combination mean you never listen at close to maximum signal level then this increases the chance of an audible noise floor with no music playing. Or indeed if you do listen at close to maximum signal levels but this is VERY loud!

 

[Pdxwayne]

Gain structure is very important, and note that the noise floor on the 2x4 HD is higher than the Flex. SINAD relates to maximum signal levels - if your amplifier gain / loudspeaker sensitivity combination mean you never listen at close to maximum signal level then this increases the chance of an audible noise floor with no music playing. Or indeed if you do listen at close to maximum signal levels but this is VERY loud!

No, I could never listen at anywhere near max volume of my Gustard x16 when directly connect to my amp (Parasound A21). If I recall correctly, maybe -40db from my Gustard DAC was mostly loud enough.

Does this mean if I lower my amp's gain, increase my DAC's volume, I could place minidsp 2x4HD in my chain and not worry about the hiss? Or should I just get the flex and skip my DAC and hd2x4?

Thanks!

 

[Ultrasonic]

No, I could never listen at anywhere near max volume of my Gustard x16 when directly connect to my amp (Parasound A21). If I recall correctly, maybe -40db from my Gustard DAC was mostly loud enough.

Does this mean if I lower my amp's gain, increase my DAC's volume, I could place minidsp 2x4HD in my chain and not worry about the hiss? Or should I just get the flex and skip my DAC and hd2x4?

Thanks!

Yes, lowering your amp's gain would definitely help. Depending on the range of adjustment you have, an in-line passive attenuator at the amplifier input could help further.

 

[mdsimon2]

No, I could never listen at anywhere near max volume of my Gustard x16 when directly connect to my amp (Parasound A21). If I recall correctly, maybe -40db from my Gustard DAC was mostly loud enough.

Does this mean if I lower my amp's gain, increase my DAC's volume, I could place minidsp 2x4HD in my chain and not worry about the hiss? Or should I just get the flex and skip my DAC and hd2x4?

Thanks!

Parasound A21 has a gain of 29 dB which is going to be rather noisy with the 2X4HD. Lower gain will definitely help. Here are some residual noise calculations ignoring amplifier noise for the 2X4HD and the Flex with amplifiers of various gain, as you can see the Flex is a lot better but you can also do a lot better by using a lower gain amplifier.

Why would you use a DAC upstream of your 2X4HD? That is only going to add noise from the ADC and likely cause you to lose even more SNR due to less than perfect gain staging. If for some reason you did use an upstream DAC you would want it to be such that it was at maximum volume without clipping the 2X4HD input and then do volume control in the 2X4HD.

Michael