Minidsp Flex Review (Audio DSP)

[Glasvegas]

MiniDSP's wording is confusing, hopefully they clean it up. Amir measured a 15dB degradation from 105dB to 90dB when using an unbalanced source to the balanced INPUT of the flex.

It is confusing. Amir measured a 15dB degradation on INPUT. They’ve noticed a 15dB degradation on OUTPUT using this cable. However, they only recommend it’s use on OUTPUT.

So, a 15dB degradation is acceptable on OUTPUT - but not INPUT.
….or I have totally misread it?

 

[Holmz]

@waynel your quote says “(Amit noticed…) 15dB on input.”

And the quote from the manual said, “use it on the output.”
???

 

[rodtor]

It is confusing. Amir measured a 15dB degradation on INPUT. They’ve noticed a 15dB degradation on OUTPUT using this cable. However, they only recommend it’s use on OUTPUT.

So, a 15dB degradation is acceptable on OUTPUT - but not INPUT.
….or I have totally misread it?

I don't think you have misunderstood the text they've written. But as @waynel said, and you too, the passage is unclear. Just above it in the same posting, moreover, they propose that users employ a TRS to unbalanced cable for an unbalanced subwoofer (the kind many people likely have, even if they have a balanced amp). So they may really have meant 'OUTPUT' when they said 'INPUT', and should clarify.

Edit: I"ve just sent a message to MiniDSP's support portal, asking for a clarification. I'll report back if I get an answer.

 

[Ultrasonic]

Many preamps have gain in addition to attenuation, so if the preamp had 10dB of gain, then it seems like it all gets to be at a good level?

Indeed - which is why I wrote this above:

Where signal level is an issue for the main speakers I suppose it would be possible to use a fixed gain pre-amp (or one made so by just not adjusting the volume knob) between the Flex output and the power amp input. I can't see this ever being an issue for an active subwoofer due to their own gain controls.

However this won't help in the slightest if the amp is already clipping. The solution to that needs to be one or a combination of a more powerful amp or a loudspeaker that is easier to drive or is more efficient. I believe you're right hiss is more likely when chasing higher output levels though.

We're getting away from the Flex here though and my key point was that the concerns raised will be far from a universal deal-breaker.

 

[Ultrasonic]

I don't think you have misunderstood the text they've written. But as @waynel said, and you too, the passage is unclear. Just above it in the same posting, moreover, they propose that users employ a TRS to unbalanced cable for an unbalanced subwoofer (the kind many people likely have, even if they have a balanced amp). So they may really have meant 'OUTPUT' when they said 'INPUT', and should clarify.

Edit: I"ve just sent a message to MiniDSP's support portal, asking for a clarification. I'll report back if I get an answer.

Like you I think they meant to say Input rather than Output but it will be good to hopefully get this cleared up.

 

[Draki]

A measurement of the Flex on the miniDSP site. It looks a little different. Apparently that was a second unit (a replacement).

Balanced Flex- Inconsistent performance from each output channel.

Hi Devteam, Just testing a balanced Flex here and 1 channel has 35dB more distortion. Any thoughts on what could cause this? I swapped cables, rebooted,...

www.minidsp.com

 

[mdsimon2]

A measurement of the Flex on the miniDSP site. It looks a little different. Apparently that was a second unit (a replacement).

Balanced Flex- Inconsistent performance from each output channel.

Hi Devteam, Just testing a balanced Flex here and 1 channel has 35dB more distortion. Any thoughts on what could cause this? I swapped cables, rebooted,...

www.minidsp.com

If it was anyone other than Mike doing the measurements I might be more concerned, still warrants investigation though.

@abdo123 you have a decent ADC these days, correct? Can you make some quick measurements and see if you see any difference between the channels?

Michael

 

[abdo123]

If it was anyone other than Mike doing the measurements I might be more concerned, still warrants investigation though.

@abdo123 you have a decent ADC these days, correct? Can you make some quick measurements and see if you see any difference between the channels?

Michael

I’m quite distortion limited but all channels measured the same for me.

 

[Ultrasonic]

Amir has shared data for all four channels of his unit of course.

The harmonic distortion levels shown on all 4 channels of Mivera1's latest unit are rather higher than Amir's data shows, but it sound like you think there's reason to doubt the accuracy of the data @mdsimon2 ?

 

[mdsimon2]

Amir has shared data for all four channels of his unit of course.

The harmonic distortion levels shown on all 4 channels of Mivera1's latest unit are rather higher than Amir's data shows, but it sound like you think there's reason to doubt the accuracy of the data @mdsimon2 ?

I don't have anything quantitative to go off of but if you google Mivera Audio you'll find old posts of Mike's and he is not very technically competent. His comments so far definitely show he doesn't have much experience with his analyzer. But all of this is my own bias. Mike is currently claiming that Amir's unit was a cherry picked golden sample.

I'm definitely interested to what miniDSP find on his first unit as that was exceptionally out of spec.

Michael

 

[rdenney]

Just out of curiosity, what sources outside of a DAC, would go above 90DB? DVD audio maybe? I noticed the best tested Phono preamp (cambridge audio duo) scored 90DB and that's messing with the gain settings to achieve the best outcome. So, it seems the loss in Sinad on the MiniFlex using an RCA to TRS adapter on the input side should have little to no consequence if using these older sources (CD, Phono, Tape, ect).

I was anticipating that the analog input would be my current analog preamp. I currently have a range of digital sources--at least three and sometimes four--in addition to my various analog sources. I was anticipating, therefore, that I would use the miniDSP between my analog preamp and my power amplifier and (potential) subs.

Two of my best CD players only have analog outputs, and test to over 90 dB S/N even in the analog domain.

My main issue, though, is that the miniDSP is not a preamp in the sense of actually, you know, amplifying anything. It's rated at unity gain, like most DACs with a volume control. That means it probably won't drive my amplifier to full power, at least using analog outputs (which max at 2 VRMS at 0 dBFS digital), and if I didn't want that powerful an amp, I wouldn't have bought it.

So, if I put the miniDSP between my analog preamp and my power amp, I really must use the balanced version to be able to drive the amp to full output. But then I have to feed an unbalanced output from my analog preamp into the balanced input of the miniDSP, which they recommended against. I'm just taking them at their word.

The SHD suffers the same fault in my setup--unity gain and 2V in and out. The difference seems to be that it allows direct unbalanced inputs and balanced outputs, with some form of active amplification in between. But it's $1200.

Maybe my setup isn't rational, but I do use the analog sources I have. I probably use the tuner at least as much as I use CD players.

Rick "whose analog preamp has more than enough preamplification headroom to drive the power amp to clipping" Denney

 

[rdenney]

In that case it honestly sounds like your speakers aren't sensitive enough for what you're trying to achieve. Running an amp right up to and hitting its limits really isn't going to be optimal, as demonstrated by the measured distortion data for your particular amp. High-pass filtering the signal sent to your main speakers may increase the headroom you have though, since I'd bet it's bass notes that mostly cause you to see clipping.

As I've said, you've raised an important aspect of the Flex performance for people to be aware of but your situation is unusual and I'm confident that the vast majority of users won't have the issues that you are concerned about.

With all due respect, that aspect of my system is no mystery to me. The amp is new and I was testing the system at its limits, with a recording of a drum solo that is extremely dynamic. Yes, I was able to flash the clipping indicators, but I heard nothing amiss, and I have not yet put my scope on the outputs to see what waveform it takes to cause those indicators to flash.

The amp is quite powerful and the speakers are reasonably efficient (91 dB) Revel F12 towers. One of my use cases is playing orchestral music at reference levels so that I can play along with my tuba and not have to hold back. Yes, I want the recording to be as loud as sitting in an orchestra, despite that I know the mix will not be the same. I don't do it often, but when I do it I want it to do what I selected and paid for it to do.

The vast majority of users may not have the issue, but I wasn't representing them. I was simply thinking out loud to determine if this device could replace my current Yamaha YDP2006, which is a commercial-sound DSP-based parametric equalizer. That unit has line-level inputs and outputs, and I have it installed in the processor loop of my analog preamplifier. It works well, but my goal would be to do what it does, plus provide the opportunity for managing subs with REW integration.

Rick "betting he's not the only one doing this arithmetic" Denney

 

[rdenney]

When they are hissing like a Pentecostal revival, they are either too sensitive, or the amps have too much noise.


If they hiss without any signal, then that would be noise from the amp.
We can talk about THD+N, but the noise part can be distressing, and the HD part may not be too distressing on the ears.



It is also possible, and likely, that a DAC could be plugged into a preamp.
Many preamps have gain in addition to attenuation, so if the preamp had 10dB of gain, then it seems like it all gets to be at a good level?

What other “line level” things are at the 20v that Denny described needing?

If someone wanted to use this to drive a Benchmark amp, they would not be able to use it in low gain, even if they still wanted to use their nice DAC going through this device. The AHB2 in low gain has a sensitivity of 9VRMS as I recall.

Rick "thinking active analog preamps typically produce around 10 VRMS--10dB of gain when standard line level was 1VRMS unbalanced" Denney

 

[Ultrasonic]

...That unit has line-level inputs and outputs, and I have it installed in the processor loop of my analog preamplifier.

Could you potentially do the same with a Flex, leave your pre-amp gain fixed at whatever you need to reach the peak levels you want, but control the volume on the Flex? This would allow you to send a third output to an active sub as well, on which note my point about about this probably allowing you to reach higher peak SPL from your main speakers once they're high-pass filtered.

 

[rdenney]

Indeed - which is why I wrote this above:



However this won't help in the slightest if the amp is already clipping. The solution to that needs to be one or a combination of a more powerful amp or a loudspeaker that is easier to drive or is more efficient. I believe you're right hiss is more likely when chasing higher output levels though.

We're getting away from the Flex here though and my key point was that the concerns raised will be far from a universal deal-breaker.

Believe me, I think this is an amazing product. If I didn't, I'd have just moved on. I want someone to make me believe that it will do what I want it to do by correcting my understanding.

Rick "but not by challenging my use cases" Denney

 

[Ultrasonic]

Believe me, I think this is an amazing product. If I didn't, I'd have just moved on. I want someone to make me believe that it will do what I want it to do by correcting my understanding.

As I've said, your understanding isn't wrong.

 

[cyruz]

You can digitally capture the processed output signal on a computer. You cannot capture it before processing and it works for any input.

Michael

Thanks Michael, so it's post processing. I'm wondering if you can apply Dirac correction on PC with a simple stereo license.

 

[antcollinet]

I was anticipating that the analog input would be my current analog preamp. I currently have a range of digital sources--at least three and sometimes four--in addition to my various analog sources. I was anticipating, therefore, that I would use the miniDSP between my analog preamp and my power amplifier and (potential) subs.

Two of my best CD players only have analog outputs, and test to over 90 dB S/N even in the analog domain.

My main issue, though, is that the miniDSP is not a preamp in the sense of actually, you know, amplifying anything. It's rated at unity gain, like most DACs with a volume control. That means it probably won't drive my amplifier to full power, at least using analog outputs (which max at 2 VRMS at 0 dBFS digital), and if I didn't want that powerful an amp, I wouldn't have bought it.

So, if I put the miniDSP between my analog preamp and my power amp, I really must use the balanced version to be able to drive the amp to full output. But then I have to feed an unbalanced output from my analog preamp into the balanced input of the miniDSP, which they recommended against. I'm just taking them at their word.

The SHD suffers the same fault in my setup--unity gain and 2V in and out. The difference seems to be that it allows direct unbalanced inputs and balanced outputs, with some form of active amplification in between. But it's $1200.

Maybe my setup isn't rational, but I do use the analog sources I have. I probably use the tuner at least as much as I use CD players.

Rick "whose analog preamp has more than enough preamplification headroom to drive the power amp to clipping" Denney

You could use a decent unbalanced to balanced converter between the preamp and the balanced Flex, rather than a bodged cable. I believe you do a little DIY - so something like this with a simple linear PSU, as a low cost solution - or there are fully finished devices that do the same.

KMTech 2-Channel DRV134 based Differential Balanced Line Driver converter | eBay

Manufacturer: KMTech Design (UK). A high performance, stereo, low noise unbalanced-to-balanced adapter by KMTech Design. However if your priority is sound quality and value for money, then look no further!

www.ebay.co.uk

You then shouln't lose 15dB of SINAD - though it would have a 6dB gain so probably lose 6dB.

(Edited to paste in a link to the correct device)

 

[rdenney]

Could you potentially do the same with a Flex, leave your pre-amp gain fixed at whatever you need to reach the peak levels you want, but control the volume on the Flex? This would allow you to send a third output to an active sub as well, on which note my point about about this probably allowing you to reach higher peak SPL from your main speakers once their high-pass filtered.

Oh, now that is an interesting thought.

I had thought that I would leave the analog preamp in passive mode with the volume turned all the way up, and use the miniDSP as the volume control for the system. That would also eliminate the Johnson noise in the Noble pot used in my preamp, not that such has even been a issue for me. That way, the miniDSP could be downstream from the preamp where it should be, and would never have more than 2VRMS coming in (well, maybe 2.1 with some devices that cheat the standard). I'd run my digital sources straight into the miniDSP, which would include a CD transport (Coax), a computer (USB). But that approach depends on the ability of the miniDSP to drive my amp without clipping.

One approach would be to install balanced outputs on my analog preamp. The pisser is that the preamp I use (B&K MC-101) theoretically provided balanced outputs as an option, but I've never seen one that exercised the option and I can't get a schematic for the unit to figure out how they did it. I might be able to use one of these TI-chip differential amps to produce that, but all I've seen are kits of unverified source for sale on ebay. However B&K did it, it managed the entire amplification range of the line-stage amp.

But no matter balanced or not, I'd have to be very careful with the preamp not to drive the inputs to the miniDSP to clipping--in active mode it provides too much voltage gain for the minimal line-level inputs.

The option you are suggesting is to split out the sub outputs in the processor loop rather than downstream of the preamp. That would mean using the preamp as a fixed line amp--and it would have to be active mode to do that--and adjusted to provide just enough gain to be able to amplify the output of the miniDSP enough to drive the amp to full power. I would definitely want as little gain as possible there, to avoid amplifying noise. As I'm seeing it, that would take only a couple of dB of gain to bridge the gap between the miniDSP and the power amp. That would leave a lot of resistance in the path, but would retain access to the other preamp features, including tone controls. I would need to mark very clearly exactly where that knob has to be.

Fiddling with the gain control on the analog preamp would then change the balance between the mains and the subs, which is why it would have to be put in the right place and then not moved. The miniDSP would become the volume control.

All the line-level switching for the record bus and for input switching is upstream from the processor loop. I think even the tape loop is upstream from the processor loop, which I need--I'm already running a dbx 400 to manage a couple of tape machines.

Yeah, that might work. I'll need to think it through some more.

Rick "and it would work with any vintage active preamp that has a processor loop" Denney

 

[rdenney]

You could use a decent unbalanced to balanced converter between the preamp and the balanced Flex, rather than a bodged cable. I believe you do a little DIY - so something like this with a simple linear PSU, as a low cost solution - or there are fully finished devices that do the same.

KMTech 2-Channel DRV134 based Differential Balanced Line Driver converter | eBay

Manufacturer: KMTech Design (UK). A high performance, stereo, low noise unbalanced-to-balanced adapter by KMTech Design. However if your priority is sound quality and value for money, then look no further!

www.ebay.co.uk

You then shouln't lose 15dB of SINAD - though it would have a 6dB gain so probably lose 6dB.

(Edited to paste in a link to the correct device)

I was looking for something like this, but didn't find anything I felt I could trust on U.S. ebay. Any chance that this device is available from a U.S. source or dealer? I've had "interesting" results shipping stuff from the other side of the pond lately, and even this is promising shipping of a month and more.

I have not seen any finished devices that offer the specifications this one is claiming. I want it to be at least as good as my analog preamp. I'd probably install it inside my preamp and get power from its internal supply, which I think provides +/- 12V rails. But small 12VDC supplies are not exactly challenging.

Rick "definitely put it on the watch list" Denney