• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements and miniDSP 2x4 HD DSP and DAC

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
1,591
Likes
636
#21
Thank for article. That’s what may happen when we use the headphone jack of a mobile phono to feed a hifi setup with audio signals.

But I am still somewhat puzzled by people who fear noise (lower SNR) at lower levels. Can we really hear the noise when playback volume is very low?

Am I misunderstanding something? I have always used gear of adequate quality so my only experience with audible noise is with iPhone’s headphone jack.
 

oivavoi

Major Contributor
Joined
Jan 12, 2017
Messages
1,156
Likes
558
#22
Ah, I misunderstood you. If volume is low... I don't know. The main point is that it would not be best practice to connect this minidsp unit to a power amp with a fixed very high gain, and then turn down the volume digitally 20 db until it reached normal to high levels.
 

Purité Audio

Major Contributor
Industry Insider
Barrowmaster
Patreon Donor
Joined
Feb 29, 2016
Messages
2,547
Likes
507
Location
London
#23
Digital attenuation reduces the signal but not the background noise, analogue attenuates both, a good implementation might be coarse control of the analogue output and ‘fine’ control via digital, Weiss, Benchmark both utilise this method.
Keith
 

oivavoi

Major Contributor
Joined
Jan 12, 2017
Messages
1,156
Likes
558
#24
Digital attenuation reduces the signal but not the background noise, analogue attenuates both, a good implementation might be coarse control of the analogue output and ‘fine’ control via digital, Weiss, Benchmark both utilise this method.
Keith
OR: Have active speakers where you easily control the volume directly on the speakers (hope that will be a possibility with the future firmware of Dutch & Dutch!)

OR: Use poweramps with analog attenuation, where the coarse attenuation is done, and then do fine control digitally via the dac (most PA-amps have attenuation, while audiphile poweramps usually don't, for some reason which defies reason).
 

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,103
Likes
1,631
Location
Monument, CO
#26
In general:
  • The amount of hiss depends upon where in the gain chain the volume control (attenuator) is and how it is implemented (e.g. attenuation from reducing the signal level, as is usual, or amplifier gain control, etc.) as well as the noise level of components before and after the attenuator.
  • There will usually be a buffer stage if not a gain stage after the attenuator that will add noise independent of the volume setting. Same is true for the power amp.
  • High gain early in the chain is usually best for overall SNR but worst for distortion (linearity). Analyzing and optimizing the gain chain can be a bit tedious though is usually straight-forward.
  • Consumer (and most pro, though different standards) audio systems have used roughly the same gains at each stage for years so there may not be a lot you can do; the flip side is that the SNR/THD trade is usually pretty good since it is based upon decades of use. Much of what drives the gain chain are fixed (relatively) standardized voltage levels at each point; MC/MM cartridge outputs, line-level source outputs, and power-amp input levels tend to be fairly consistent among brands.
  • Note analog and digital volume control is a bit trickier to define these days as a number of units use digitally-controlled continuous and stepped attenuators. Control is digital, but the actual attenuator is analog.
  • Much ado has been made about the loss of resolution using digital volume control, but remember at lower volume and resolution we hear less low-level detail as well, so arguably it's a wash. Extreme example: Use digital attenuation so you are only using the bottom couple of bits on your 16-bit DAC. Now the SNR is only 12 dB, and without dither the signal looks pretty coarse, but the level is at about -88 dBFS so does it really matter?
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,574
Likes
1,542
Location
Riverview, FL
#27
Extreme example: Use digital attenuation so you are only using the bottom couple of bits on your 16-bit DAC. Now the SNR is only 12 dB, and without dither the signal looks pretty coarse, but the level is at about -88 dBFS so does it really matter?
I can appreciate your point there for 16-bit attenuation... at -88dBfs only 7 "voltage" levels remain... 3, 2, 1, 0, -1, -2, -3

As a practical matter, how many Audio DACs implement 16 bit volume control? (I have no idea)

I'd think 24 bit volume control is the modern norm - leaving 10 bits resolution in your scenario. (the miniDSP2x4HD is in this category)

Or even 32 bit - leaving 18 bits available theoretical resolution at -88dBfs (my Benchmark claims this)
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,574
Likes
1,542
Location
Riverview, FL
#28

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
15,369
Likes
9,358
Location
Seattle Area
#29

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,574
Likes
1,542
Location
Riverview, FL
#30

DonH56

Major Contributor
Technical Expert
Patreon Donor
Joined
Mar 15, 2016
Messages
2,103
Likes
1,631
Location
Monument, CO
#31
I can appreciate your point there for 16-bit attenuation... at -88dBfs only 7 "voltage" levels remain... 3, 2, 1, 0, -1, -2, -3

As a practical matter, how many Audio DACs implement 16 bit volume control? (I have no idea)

I'd think 24 bit volume control is the modern norm - leaving 10 bits resolution in your scenario. (the miniDSP2x4HD is in this category)

Or even 32 bit - leaving 18 bits available theoretical resolution at -88dBfs (my Benchmark claims this)
My example was for a 16-bit DAC so it doesn't really matter if the control before it uses 16, 32, or 128 bits -- the DAC only has 16-bit resolution. You could scale the example to 24 bits. I wasn't going for specific numbers, just the concept that losing resolution when you turn down the volume by reducing the effective resolution (throwing away bits through digital scaling) is not necessarily going to destroy all the "fine detail in the sound" as some reviewers have said.
 
Joined
Apr 7, 2018
Messages
14
Likes
9
#32
Why don't you buy a miniDSP first and then tell me. You can NOT download their software from their website. Here is what you get under "downloads:"

View attachment 12237

See? No software. You need to redeem a stupid code. And then you are greeted with a bigtime warning that you better know what you want as you can't download anything else if you select the wrong thing, etc., etc.

I test dozens of products on my computer. No way I want to download and install software and drivers if the alternative is for it to plug-and-play with some other hardware.

My job as a reviewer is to comment on things that I don't like. That is what a review is.
Amir. His comment is a little difficult to interpret but I'm not entirely convinced he meant it to sound like a personal attack. I read it more of a "things aren't what they used to be" type post. Only he knows for sure though.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
15,369
Likes
9,358
Location
Seattle Area
#33
By measurement section I mean actual Audio Precision graphs like I post. The single number table is not nearly as informative.
 

rebbiputzmaker

Addicted to Fun and Learning
Joined
Jan 28, 2018
Messages
549
Likes
77
#34
Amir. His comment is a little difficult to interpret but I'm not entirely convinced he meant it to sound like a personal attack. I read it more of a "things aren't what they used to be" type post. Only he knows for sure though.
Thank you, it was not anywhere near meant as an attack, but there is a certain hierarchy where at times (some) with different opinions are taken badly.

Seriously most people I know always prefer using manufacture drivers/software. An example even though Win 10 has almost everything covered, and will automatically D/L anything else needed, manufacture drives are still often better. chipset, video drivers may have many more settings or options, or improve stability. Trackpads /mouse /pen type drivers, more settings, features, better stability etc. It goes on.

The Win 98 comment was not totally correct, as Win 2000-Win 7 still needed user input to run properly and get best results.

When you purchase brand name systems, luckily many manufactures makes the process easy and automatic. Just open a page d/l a small app and it scans your system and gives you the updates you need.

It all depends... it all becomes second nature when you have set up hundreds of servers, systems, wans, lans etc.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
15,369
Likes
9,358
Location
Seattle Area
#35
I suspect that is only for the USB input/output, as changing it with analog or toslink as the active input on the hardware doesn't cause any little tick or pop in the analog output.
That's exactly right. As a USB DAC it exposes all of those sampling rates. But as an analog-in/analog-out, you can't control the sampling rate.

I wonder what the internal DSP rate is. It is possible even when it accepts higher input sample rates over USB, it internally resamples to low rates for processing.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
5,574
Likes
1,542
Location
Riverview, FL
#36
My understanding is that the internal rate is 96khz.

Example: If writing your own biquads, you write them for 96kHz. Same for FIR, though I can't locate the reference right now.

upload_2018-4-19_18-15-26.png
 
Joined
Aug 25, 2017
Messages
13
Likes
4
Location
Long Island
#37
Thanks for measuring this unit Amir. I have a minidsp DDRC-22A and after using it for a few weeks I found that while it did make a difference, the mids were no where near as clear with the unit in the path as they were with it out. Even turning the DSP off for frequencies above 500hz, the minidsp DAC chip was still being used and noticeably degraded the sound. Looking at the measurements is nice confirmation that it is indeed that the DAC itself is really not up to the job.
 

oivavoi

Major Contributor
Joined
Jan 12, 2017
Messages
1,156
Likes
558
#38
Thanks for measuring this unit Amir. I have a minidsp DDRC-22A and after using it for a few weeks I found that while it did make a difference, the mids were no where near as clear with the unit in the path as they were with it out. Even turning the DSP off for frequencies above 500hz, the minidsp DAC chip was still being used and noticeably degraded the sound. Looking at the measurements is nice confirmation that it is indeed that the DAC itself is really not up to the job.
Interesting! May I just ask if you boost the bass with the unit on - is it the Dirac house curve? Have you tried having it on, but keeping the frequency response more or less exactly as it was without DSP? My experience is that boosting the bass may indeed affect the perceived clarity of the mids and highs - more bass can imply less "clarity". So it may be as simple as that.

Or have you tried turning off dirac for all frequencies above 150 or 200 hz? I'm ambivalent about room correction in higher frequencies. I couldn't live without in the bass, given my room, but I'm not sure what I think about it higher up. The change you perceive could be due to the Dirac algorithm, not the hardware itself.

But it could also be the chip itself, of course.
 
Joined
Aug 25, 2017
Messages
13
Likes
4
Location
Long Island
#39
Interesting! May I just ask if you boost the bass with the unit on - is it the Dirac house curve? Have you tried having it on, but keeping the frequency response more or less exactly as it was without DSP? My experience is that boosting the bass may indeed affect the perceived clarity of the mids and highs - more bass can imply less "clarity". So it may be as simple as that.

Or have you tried turning off dirac for all frequencies above 150 or 200 hz? I'm ambivalent about room correction in higher frequencies. I couldn't live without in the bass, given my room, but I'm not sure what I think about it higher up. The change you perceive could be due to the Dirac algorithm, not the hardware itself.

But it could also be the chip itself, of course.
I tried a lot of different curves, everything from their house curves to a few recommended on other forums to my own, including no curve just having the signal go through the miniDSP.I had the curve stop at as low as 200hz. My DAC has two analog outputs active at the same time, so after a while working to get levels the same I was able to switch, on my preamp, between the DAC going into the miniDSP to the preamp and just my DAC going to the preamp. The difference in the two was very apparent, the miniDSP absolutely degraded the sound of the mids in my system. My DAC is the Wyred4sound DAC-2v2 SE FWIW.
 

oivavoi

Major Contributor
Joined
Jan 12, 2017
Messages
1,156
Likes
558
#40
I tried a lot of different curves, everything from their house curves to a few recommended on other forums to my own, including no curve just having the signal go through the miniDSP.I had the curve stop at as low as 200hz. My DAC has two analog outputs active at the same time, so after a while working to get levels the same I was able to switch, on my preamp, between the DAC going into the miniDSP to the preamp and just my DAC going to the preamp. The difference in the two was very apparent, the miniDSP absolutely degraded the sound of the mids in my system. My DAC is the Wyred4sound DAC-2v2 SE FWIW.
Ok, thanks! It may seem as if the minidsp chip indeed does something then. Perhaps it's the resampling? Or the jitter etc. Hard to know.
 
Top Bottom