• 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!

Review and Measurements of Benchmark DAC3

Vasr

Major Contributor
Joined
Jun 27, 2020
Messages
1,409
Likes
1,923
I use five subs because more subs substantially reduces measured THD at the main listening position (MLP).

Is there a technical reason to support this? You are suggesting more subs decrease non-linear distortions? I am not an expert on this so would really like to know if this is true.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Is there a technical reason to support this? You are suggesting more subs decrease non-linear distortions? I am not an expert on this so would really like to know if this is true.

Each bass driver can play at a lower amplitiude, which typically reduces a speaker's distortion faster than the reduction in volume of the fundamental.

I have cheap subs assisting/duplicating the woofer in the mains.

The combination of three drivers on each side allows a 9dB reduction in the output of any one of the drivers.
 

anmpr1

Major Contributor
Forum Donor
Joined
Oct 11, 2018
Messages
3,722
Likes
6,406
Is there a technical reason to support this? You are suggesting more subs decrease non-linear distortions? I am not an expert on this so would really like to know if this is true.
Generally, the louder a single speaker plays, that is, the more air it moves, the higher will be its distortion. By moving the same amount (or more) of air with multiple speakers than with a single speaker, I can understand how overall distortion would be lowered. Especially with bass since it's usually a non-directional thing. Multiple drivers mounted on a single baffle in a 'full range' loudspeaker can, however, result in different problems other than distortion at loud SPL. With subs, at their intended frequencies, you don't have the same problems that you could experience with a large-baffle full range loudspeaker comprised of a bunch of mids and/or tweeters lined up.

Certainly @etc6849 has gone to lengths to integrate multiple woofers into his listening room, evidently eliminating most typical subwoofer problems, and I'm pretty impressed on a 'theoretical and technical' level. I never thought about such a large-scale system (five subs), but would certainly like to hear it, or another like it. I am not convinced that on most acoustic music this sort of setup would offer a significant benefit over something simpler, but I am happy to be disabused about it.
 

etc6849

Active Member
Joined
Apr 28, 2018
Messages
132
Likes
241
Location
Columbia, SC
I don't want to go into too many details here as this has already derailed this thread for an outstanding product that I also own and love...

But, my post said multiple subs reduces THD and concurrently gets rid of unwanted room issues. Your post is only focusing on distortion from the speaker, and ignoring issues from small room acoustics (major contributor that should always be tackled first before buying new equipment).

I would encourage you to read the help files on this site then study the two charts below...
https://www.roomeqwizard.com/betahelp/help_en-GB/html/graph_waterfall.html

When I compare my setup to any I've heard, the ringing in the bass region versus what I'm used to is easy for me to hear. Note my setup has a decay time of ~120ms for a -45dB decay, and this is consistent for frequencies above 40 Hz. In case you are wondering, yes Dirac can improve this even more at the expense of digital headroom. I don't use an analog volume control, so I like the extra headroom though.

Here is the waterfall plot for my front right speaker crossed to the five subs with no EQ:
Front Right Waterfall No EQ 10.28.2018(3).png


A lot of systems have decays two, three or even more times what I'm showing above, and most plots will look very uneven in the decays too with ringing at several frequencies. Here is a random plot I pulled from the REW help file. If you look -45dB down from the localized peaks, you can indeed see ringing is 2-3 times my setup and very uneven.
1596498886824.png


The affect of multiple subs is well researched and there are multiple white papers if you look around on google. I also have a lot of room treatments too. Many of the best recordings I have already include echos from the space or venue they are recorded in, why add all the ringing in the bass region during playback that shouldn't be there as it's from your room? If you are ever near South Carolina, feel free to drop me a line. I promise you can hear the difference ;)

@anmpr1
 

misteracng

Member
Joined
Jul 21, 2020
Messages
39
Likes
25
Location
East Bay, CA
If using the dac3/ahb2 combo, I assume there would really be no advantage in adding an LA4 or HPA4, if no other inputs are needed? The dac3 being used as essentially the preamp.
 

Inner Space

Major Contributor
Forum Donor
Joined
May 18, 2020
Messages
1,285
Likes
2,938
If using the dac3/ahb2 combo, I assume there would really be no advantage in adding an LA4 or HPA4, if no other inputs are needed? The dac3 being used as essentially the preamp.

Theoretically (and perhaps counterintuitively) inserting the LA4/HPA4 can further optimize the gain structure, improving the already-stellar numbers a point or two; and the LA4/HPA4 gives you a better-resolving volume control. So an absolute perfectionist might do it ... but I didn't.
 

anmpr1

Major Contributor
Forum Donor
Joined
Oct 11, 2018
Messages
3,722
Likes
6,406
If using the dac3/ahb2 combo, I assume there would really be no advantage in adding an LA4 or HPA4, if no other inputs are needed? The dac3 being used as essentially the preamp.
In addition you will gain a line level balance control (it looks like), and more powerful headphone amplifier with the 'H' version. The DAC3 B would seem to be the logical choice if using a LA4/AHB.
 

misteracng

Member
Joined
Jul 21, 2020
Messages
39
Likes
25
Location
East Bay, CA
Theoretically (and perhaps counterintuitively) inserting the LA4/HPA4 can further optimize the gain structure, improving the already-stellar numbers a point or two; and the LA4/HPA4 gives you a better-resolving volume control. So an absolute perfectionist might do it ... but I didn't.
I was thinking about this as well. I like the idea of better volume control.
 

audiopile

Active Member
Joined
Jun 27, 2019
Messages
161
Likes
125
Anybody know if the UltraLock 3 jitter reduction feature is implemented if you use the second SPDIF as a digital passthru ? IOW -when jumpered to make this input a output -does it just take whatever is comin in and pass it back out or is anything done to reduce jitter from playback device or cabling ?
 

¥€$

Member
Joined
Sep 2, 2020
Messages
44
Likes
12
There's something perturbing about how others have reported loss of audio quality when using the attenuator pads in -10/20 dB setting. So I decided to test them both and indeed, at least the -20 dB setting seems to wipe out detail, even with the volume pot @ 12 o'clock. It makes sense that signal attenuation that changes impedance will affect the sound, but how bad exactly would the measurements be in comparison with 0 dB setting? If only @amirm would revisit this device again and tested its balanced outputs with the other gain settings.

https://forums.stevehoffman.tv/threads/benchmark-dac2-hgc-interesting-settings-discovery.687368/

Just as the original poster there mentions, there are a couple of older posts elsewhere with similar observations. This is disturbing because Benchmark have always stressed the importance of correct output level when these are used as preamps.
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Anybody know if the UltraLock 3 jitter reduction feature is implemented if you use the second SPDIF as a digital passthru ?

My guess would be "no".

"Digital Pass-Through
The second coaxial input (D4) can be reconfigured as a digital output. When operating as an output, any selected digital input is passed through to D4 without any processing."

My understanding:
Ultralock 3 takes the input clock (or creates one, in the case of USB), performs ASRC with non-integer rate difference, and creates a low-jitter 211kHz clock for the converter.
 
Last edited:

Inner Space

Major Contributor
Forum Donor
Joined
May 18, 2020
Messages
1,285
Likes
2,938
There's something perturbing about how others have reported loss of audio quality when using the attenuator pads in -10/20 dB setting. So I decided to test them both and indeed, at least the -20 dB setting seems to wipe out detail, even with the volume pot @ 12 o'clock. It makes sense that signal attenuation that changes impedance will affect the sound, but how bad exactly would the measurements be in comparison with 0 dB setting? If only @amirm would revisit this device again and tested its balanced outputs with the other gain settings.

https://forums.stevehoffman.tv/threads/benchmark-dac2-hgc-interesting-settings-discovery.687368/

Just as the original poster there mentions, there are a couple of older posts elsewhere with similar observations. This is disturbing because Benchmark have always stressed the importance of correct output level when these are used as preamps.

1) A guy has his pads at -20, and "everything sounded great".
2) He reads online that 0dB is better.
3) He resets to 0dB ...
4) ... OMG it sounds better.

You did the same in reverse. Stop me if I'm going too far, but could expectation bias be playing a tiny role here?
 

¥€$

Member
Joined
Sep 2, 2020
Messages
44
Likes
12
1) A guy has his pads at -20, and "everything sounded great".
2) He reads online that 0dB is better.
3) He resets to 0dB ...
4) ... OMG it sounds better.

You did the same in reverse. Stop me if I'm going too far, but could expectation bias be playing a tiny role here?

Of course it could be, but then it would've also played a part in my expecting that raising the volume from 9.30 (original position with 0 dB attenuation) to 12 o'clock and above after enabling attenuation would improve the sound - it didn't. Were there two competing expectation biases? :)

This is why you want hard data, and so would I, but Amirm is done with this one, I'm sure. I'll just keep on listening to it for a while, then maybe I'll switch back to 0 dB for the sake of experiment.
 
Last edited:

Inner Space

Major Contributor
Forum Donor
Joined
May 18, 2020
Messages
1,285
Likes
2,938
Of course it could be, but then it would've also played a part in my expecting that raising the volume from 9.30 (original position with 0 dB attenuation) to 12 o'clock and above after enabling attenuation would improve the sound - it didn't. Were there two competing expectation biases? :)

Did I misread? The guy on the Hoffman forum went from -20 to 0 and concluded 0 was better. You were already on 0 and went to -20 and confirmed his finding. You were agreeing, surely? Expectation bias ran in the same direction.
 

¥€$

Member
Joined
Sep 2, 2020
Messages
44
Likes
12
Did I misread? The guy on the Hoffman forum went from -20 to 0 and concluded 0 was better. You were already on 0 and went to -20 and confirmed his finding. You were agreeing, surely? Expectation bias ran in the same direction.

I respect the engineers of Benchmark so one could also imply that I'm biased to believe that their recommendations are more valid than some anonymous internet poster's. I had long planned to test the other settings to see if there were sonic benefits to be had, it was that one post that made me take the plunge. Benchmark have made it especially difficult to A/B test this feature by hearing, for obvious reasons. In the end what counts is which setting sounds best at that moment, bias or no. I don't need this piece for business, only pleasure, so in the absence of measurements that will have to be the deciding factor.
 
Last edited:

¥€$

Member
Joined
Sep 2, 2020
Messages
44
Likes
12
I've settled for the - 10 dB setting for now since I can't hear a difference between that and 0 dB, but now I'm using a slightly lower volume than what it was with 0 dB, even though the volume pot's setting is higher. This makes the sound softer, lower volume always seems to curb treble and I want to see if I could get used to it. I never touch the pot again after finding a sweet spot, I only control volume with software afterwards, and the programs with a fixed volume setting now sound quieter. I've been having more listening fatigue lately anyway so I doubt my earlier findings were particularly sound, but I'm not going to test the - 20 dB anymore now that I've found a satisfactory setting (OCD), too bad I don't have the skill or equipment to properly test this theory.

The Benchmark manual states that the -10 dB setting increases the balanced output impedance more than the - 20 dB one. There is at least one typo in there but I doubt this is a mistake. I'm no electrician so why is this so?
 
Last edited:

Kaiede

Member
Joined
Aug 3, 2020
Messages
39
Likes
29
The Benchmark manual states that the -10 dB setting increases the balanced output impedance more than the - 20 dB one. There is at least one typo in there but I doubt this is a mistake. I'm no electrician so why is this so?

As a nitpick, I don’t think an electrician would have much insight either. A good EE might though. ;)

I noticed that in the manual as well. From the shot of the PCB in the manual, the attenuators appear to be purely passive, so there’s nothing terribly obvious there for me to latch onto. My EE is too rusty to be much good here, and my signal processing classes are a distant memory by now.

Earlier in the thread, amirm did mention there were differences he measured in the output when using the different attenuator settings. He didn’t go into what those differences were, or how large, sadly. All the measurements were done at the -10dB setting.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
Last edited:

¥€$

Member
Joined
Sep 2, 2020
Messages
44
Likes
12
As a nitpick, I don’t think an electrician would have much insight either. A good EE might though. ;)

I noticed that in the manual as well. From the shot of the PCB in the manual, the attenuators appear to be purely passive, so there’s nothing terribly obvious there for me to latch onto. My EE is too rusty to be much good here, and my signal processing classes are a distant memory by now.

Earlier in the thread, amirm did mention there were differences he measured in the output when using the different attenuator settings. He didn’t go into what those differences were, or how large, sadly. All the measurements were done at the -10dB setting.


Here it is (the mysterious conclusion in bold):

Measurements

Please note that all measurements are made with USB input of the DACs. Two posts down there is a comparison of Toslink to USB.

I always like to match levels of DACs where possible to make a fair comparison. Alas, even though Benchmark documents that the unit by default outputs 24 dBu (12 volts RMS), it puts out copious more output than that. I had to use a "pad jumper" inside the unit to dial in -10 dB of attenuation to match that of RME ADI-2 which indeed maxes out at 24 dBu. In this dashboard view of DAC3, I have one channel with default output and the other with -10 dB:

index.php


As you see, the non-attenuated channel outputs nearly 20 volts RMS, corresponding to whopping +-28 volts or peak to peek! :eek: With that kind of output swing if you are bored, you can hook up the DAC3 to drive your Christmas tree lights! :D

Dallasjustice uses the much higher output to good results in room equalization. The most effective way to deal with troughs in room response is to simply bring down the levels of all the other frequencies. This gives you a flat response but now you have much less gain to drive your speakers. Having so much headroom and output capacity in the DAC3 helps mitigate this.

For the rest of these tests, I ran with the attenuation at -10 dB to match the RME as mentioned. Note that performance without this attenuation is different (sometimes better, sometimes worse). Such is life when we try to make apples vs apples comparison.
 

¥€$

Member
Joined
Sep 2, 2020
Messages
44
Likes
12
After using the -10 dB setting for a while I began to notice that when the DAC turns on, a few seconds after the lamp test a click comes from inside it. It sounds like a relay. Has anyone else noticed this? I never heard it when the pads were set at 0 dB, and my trial run of -20 dB was so brief that I may have missed it. Perhaps it comes from the volume pot's motor now that its position is higher?
 
Top Bottom