• 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!

Quick measurements of Behringer Ultracurve DEQ2496

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
DEQ2496 is a very complex device with many features and functions. It can work in digital domain entirely, or convert between digital and analog for both, inputs and outputs.

First up is digital performance. This is measured using optical input and output, so the signal never leaves digital domain. Source signal was generated by REW and dithered to 24 bits there.

Let me know if you have any specific tests you want to run. Considering the number of functions and features and their combinations, I might be measuring this thing for the next year :)

This particular DEQ2496 was manufactured in 2008 and has firmware version 2.4.

1kHz with all features bypassed, no dither enabled on DEQ2496:
1khz-bypass.png


24-bit dither and noise shaper:
1khz-bypass+24dither-noiseshaper.png


1Khz, PEQ 100hz Low-Cut filter:
1khz-lc-at-100hz.png

1KHz, PEQ 100Hz, BW=1OCT, Gain=-5dB:

1khz-peq-100hz-oct1-5db.png


PEQ High-Cut @30Hz filter applied to white noise signal:
peq-hc-5db.png



PEQ @30Hz High-Cut (HC) 12dB/Oct filter:
peq-HC12-5dB.png


PEQ BW=1 Oct, Gain=-5dB:
peq-oct1-5dB.png


Frequency sweep with PEQ 30Hz HC12 -5dB filter:
sweep-hc12db-5dB-30Hz.png



Sweep of PEQ 100Hz, BW=1OCT, Gain=-5dB:
sweep-oct1-5db-100hz.png


Impulse response of PEQ 100Hz, BW=1OCT, Gain=-5dB filter:
IR-oct1-5db-100hz.png



SMTPE dual-tone signal with PEQ 100Hz, BW=1OCT, Gain=-5dB filter:

smtpe-100hz-oct1-5db.png


SMTPE signal with no filters:
smtpe-no-filter.png


15kHz High-Cut PEQ filter sweep:
15khz-hc-fq-phase.png


Impulse response 15kHz High-Cut PEQ filter:
15khz-hc-ir.png
 

Attachments

  • 1khz-bypass+24dither.png
    1khz-bypass+24dither.png
    106.5 KB · Views: 234
Last edited:

Wombat

Master Contributor
Joined
Nov 5, 2017
Messages
6,722
Likes
6,459
Location
Australia
Please measure the analog ADC/DAC performance separate and overall if you have the time.:)

Could you briefly comment on the performance charts, above? They appear to contradict much criticism, on the WWW, of this item.
 
Last edited:
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Please measure the analog ADC/DAC performance separate and overall if you have the time.:)

Could you briefly comment on the performance charts, above? They appear to contradict much criticism, on the WWW, of this item.

Considering these were all digital in/out captures, I don’t see anything terribly wrong in the DSP part of the PEQ implementation. Don’t really know what is being criticized, perhaps it’s the analog performance? I’ll do some quick measurements, but I never used DEQ for analog signal processing, its DSP in the digital domain is what first attracted me to it... and the low price, of course ;)
 

StevenEleven

Addicted to Fun and Learning
Forum Donor
Joined
Dec 1, 2018
Messages
581
Likes
1,188
This is awesome. I’ve had the DEQ2496 since, it seems like, the early to mid 2000s, maybe 2003 or 2004. I still remember the guy at Guitar Center telling me all of the things it could do and I was just floored, it pretty much sold itself. I never did the big firmware upgrade, I forget when that was. I will do that some time. I have used it mainly for headphone EQ and crossfeed, but I’d like to maybe get some benefit out of the RTA / mic interface some time and try out some of the other stuff more extensively.

Anyway, thanks for the measurements, and if anyone can translate them out of geek speak to laymen‘s terms I’d be very appreciative. This is fun. :cool: Thanks!!!
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
This is awesome. I’ve had the DEQ2496 since, it seems like, the early to mid 2000s, maybe 2003 or 2004. I still remember the guy at Guitar Center telling me all of the things it could do and I was just floored, it pretty much sold itself. I never did the big firmware upgrade, I forget when that was. I will do that some time. I have used it mainly for headphone EQ and crossfeed, but I’d like to maybe get some benefit out of the RTA / mic interface some time and try out some of the other stuff more extensively.

Anyway, thanks for the measurements, and if anyone can translate them out of geek speak to laymen‘s terms I’d be very appreciative. This is fun. :cool: Thanks!!!

Mine has the date code of 2008. The digital in and out measurements are just showing what the various filters do to the signal, and whether or not the digital signal is handled properly inside the unit. From a quick look over the charts I posted above, everything looks reasonable and should perform well when used for all digital signal. I'll post the analog results a bit later, these will tell us how well the DAC and ADC portions perform.
 

Ron Texas

Master Contributor
Joined
Jun 10, 2018
Messages
6,077
Likes
8,913
It's probably there, but the print is small. What's the digital in, analogue out SINAD?
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Adding analog tests for the DAC. ADC measurements - later.

Frequency response:
1581309606566.png


Distortion vs level (THD+N is the grey line overlapping the green=noise). THD+N is pretty constant at around -98dB and dominated by noise until about -20dBFS, at which point the harmonic distortions start to push it up, getting to a maximum of -78dB at 0dBFS.

1581310641809.png


Here's what it looks like at 1kHz and -20dBFS:

1581309954259.png


1kHz @ 0dBFS:
1581309890440.png


Distortion vs. Frequency shows some increasing distortion (up to THD+N of -83dB at about 14kHz):
1581311099933.png


CCIF signal shows a high level of IMD components:
1581311278199.png


Jitter J-Test result shows a good amount of spurious tones and jitter sidebands, but all below -105dBFS and most below -120dBFS:

1581392026746.png
 
Last edited:

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,579
Likes
38,280
Location
Gold Coast, Queensland, Australia
1kHz @ 0dBFS:
index.php


There's no doubt that is harsh clipping at 0dBFS. Back off the generator a fraction of a dB and retest it?

Can you post the 0dBFS 1KHz REW generated file or check it for consecutive clipped samples?
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
There's no doubt that is harsh clipping at 0dBFS. Back off the generator a fraction of a dB and retest it?

Can you post the 0dBFS 1KHz REW generated file or check it for consecutive clipped samples?

No, no clipped samples, but the harmonic distortion does increase rather consistently above -30dBFS level. I watch as it goes up when doing the level vs distortion sweep, and all those "extra" harmonics are there at lower levels, not just at 0dBFS. Here it is at -5dBFS:
1581310818072.png
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East

Wombat

Master Contributor
Joined
Nov 5, 2017
Messages
6,722
Likes
6,459
Location
Australia
Adding analog tests for the DAC. ADC measurements - later.

Frequency response:
View attachment 49514

Distortion vs level (THD+N is the grey line overlapping the green=noise). THD+N is pretty constant at around -98dB and dominated by noise until about -20dBFS, at which point the harmonic distortions start to push it up, getting to a maximum of -78dB at 0dBFS.

View attachment 49522

Here's what it looks like at 1kHz and -20dBFS:

View attachment 49517

1kHz @ 0dBFS:
View attachment 49516

Distortion vs. Frequency shows some increasing distortion (up to THD+N of -83dB at about 14kHz):
View attachment 49524

CCIF signal shows a high level of IMD components:
View attachment 49525

That should be OK for my old ears.
Don't%20tell%20anyone.gif
 

Hipper

Addicted to Fun and Learning
Joined
Jun 16, 2019
Messages
753
Likes
625
Location
Herts., England
This is awesome. I’ve had the DEQ2496 since, it seems like, the early to mid 2000s, maybe 2003 or 2004. I still remember the guy at Guitar Center telling me all of the things it could do and I was just floored, it pretty much sold itself. I never did the big firmware upgrade, I forget when that was. I will do that some time. I have used it mainly for headphone EQ and crossfeed, but I’d like to maybe get some benefit out of the RTA / mic interface some time and try out some of the other stuff more extensively.

Anyway, thanks for the measurements, and if anyone can translate them out of geek speak to laymen‘s terms I’d be very appreciative. This is fun. :cool: Thanks!!!

Bought mine in 2006 and used it to listen to Red Book CD since then, set between the Transport and DAC so operating only in digital.

Lots of comments on the internet say it's noisy with DEQ/PEQ or bypass, unreliable and generally not much good. That is not my experience. I don't notice any noise, it's proved reliable (fourteen years now) and it works brilliantly for what I want it to do.

I used to use the RTA but now use REW to suggest filters and apply them manually to the DEQ2496. I've found that the best sounding solution.

Thanks pkane for these measurements. Whilst I don't understand it all they generally seem to confirm what long term users know - it works well in digital but probaly OK using its ADC and DAC.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
1KHz, PEQ 100Hz, BW=1OCT, Gain=-5dB:

View attachment 49491

Many thanks for posting those measurements. :)

Clear rise in LF noise but doesn't look too bad.

Here's a filter identical in frequency response but with an "ill-behaved" implementation (an old VST plug-in):

(-3dBr 1kHz sine input.)

1581352839879.png



I left the process running for a few minutes. It settled into different states--so here's the same filter AGAIN:

1581353048904.png


And the same filter AGAIN:

1581352995794.png




BTW, the sine alone would appear on the above spectral plots as absolutely clean, as it would after passing through a 64-bit float implementation of the same filter.

Impulse response of PEQ 100Hz, BW=1OCT, Gain=-5dB filter:

View attachment 49490

Hmm. Symmetrical impulse response?

15kHz High-Cut PEQ filter sweep:
View attachment 49494

Not decrampled so the frequency response is warped toward Nyquist--i.e. the response at Nyquist goes to minus infinity.

Here is the response of two LP filters, one is an FIR filter vs. a (regular biquad) IIR filter:

1581353552065.png


The FIR is red, the IIR is green.

It can be seen that the response is grossly warped towards Nyquist--fortunately at 96kHz sample rate, this is not too catastrophic in terms of magnitude response within the audible range.

I suggest testing the dynamics processes of the DCX2496. Here is an example spectral plot of a (digital) compressor fed with a 998Hz sine, fast attack/release setting on the compressor:

1581353732809.png


And if you think that's bad, you should see it at 44.1 or 48kHz... more aliasing...
 
Last edited:

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
This is awesome. I’ve had the DEQ2496 since, it seems like, the early to mid 2000s, maybe 2003 or 2004. I still remember the guy at Guitar Center telling me all of the things it could do and I was just floored, it pretty much sold itself.

If the world made sense there would be no reason to be "floored"--and silly mains cables costing 10x the price wouldn't exist, either. ;-)
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
I suggest testing the dynamics processes of the DCX2496. Here is an example spectral plot of a (digital) compressor fed with a 998Hz sine, fast attack/release setting on the compressor:

View attachment 49563

And if you think that's bad, you should see it at 44.1 or 48kHz... more aliasing...

Here's the compressor set to 2.5 ratio, -10dB gain, -10dB threshold, 1ms attack, 20ms release (at 24/44.1k):
1581393613584.png


Nasty, but I guess I'm not sure why you'd expect a compressed waveform to look pretty :)
 
Top Bottom