• 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). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Behringer X32 Measurements (Digital Mixing Console)

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397
This is my 'review' of the Behringer X32 (Compact), one of the most popular Pro Mixing Consoles in the live sound business.
It's a 32 channel desk (hence the name) with 23 local (analog) inputs and 16 outputs. It comes with a USB expansion card, which made it possible to test its ADC and DAC sections separately.
The console was released in 2012 (almost 10 years ago!) and I bought this unit in 2018. It can be had for ~1500€/1900USD now.
My tests were performed with the desk running at 48khz.

JnjAhIb.jpg


Let's start with the DAC section. This means I'm sending a signal from the PC to the X-USB interface (the digital expansion card inside the console), routing that straight out of the XLR port in the back.

DAC, 4V, 98,4dB THD+N:

Hss18kc.jpg


Just shy of 100dB THD+N. That's a long shot compared to modern state of the art DACs, but it isn't bad and close to the performance of the Focusrite Scarlett 2i2.
I don't know if the desk has seen hardware improvements over the years, but if a new console measured like this in 2012, it would've been a pretty nice result.

Frequency response:

x1SYfyV.jpg


Dead flat in the audible band, that's how it's supposed to be.

THD vs frequency @ 4V:

mg5SH7r.jpg


THD rises with frequency, but I've seen worse. It's worst at 10k (and then drops to nothing, since HD above 10k is inaudible and filtered out), but even there it's ~95dB quieter than our signal.

THD vs output level:

x7uzJyC.jpg


Nothing unexpected here. THD+N is noise dominated until ~5V output, when the 2nd harmonic starts coming out.


That's it for the DAC, let's have a look at the ADC section now:

ADC, 4V, 89,6 THD+N:

fmJjWi5.jpg


Noise is a bit lower compared to the DAC side, but we're fighting against odd order harmonics here. THD+N is dominated by distortion, which we will later see in the distortion vs output graph again. 90dB is nothing to write home about, but not a deal breaker. Hey, maybe the preamps were optimized for mic input levels, which are much lower? Let's see if that is the case, but first we'll have a look at the frequency response.

Frequency Response:

d6MWfuW.jpg


Not as flat as the DAC side, but -1db at 20khz isn't an issue.

THD vs frequency, 4V:

Jg5qNCA.jpg


Well this is a puzzle... The cursor is at 170Hz, where THD is at -109dB. SINAD would be noise dominated in this case and around 100dB. This is 10dB better than at 1k and ~25db better than at 7k. What's going on here? The odd order harmonics are haunting us again.

Let's see if distortion is optimized for lower input levels now.

THD vs input level:

s7qmNJp.jpg


Sadly, this is not the case. Best case SINAD seems to be between 1 and 2V. Distortion sets in very early, but I have observed some odd behaviour on this during my testing, which I will report on later. I have a few more measurements I want to show first.
I also want to note that I could only measure to 4V input, since my DAC can't output more. Upgrades to 10V coming in the future :)

Noise floor XLR output:

JvqXcmB.jpg



Noise floor with 16 input channels unmuted and at unity gain:

R88CqaI.jpg



Same thing with all channels with an active gate:

v5gRfvP.jpg


Well the gates work, that's good. They clean up our output spectrum while there is no signal present. Good to use em, even when nothing is plugged in, so you get rid of the preamp noise.

Round trip through analog in → analog out:

2lDAoiZ.jpg



Same thing on different channels:

Fc88g0b.jpg



Crosstalk:

MrZFmdO.jpg


I couldn't get good results on this doing a sweep, so I pulled up the RTA and enabled infinite peak hold (hence the red colour). Test frequencies are as shown (100, 1000, 10000Hz).
I sent a 4V input signal into Ch1, and out Ch7. The RTA shows the output of Ch8 with the input Ch2 open to it at unity gain (to basically get double crosstalk on input and output combined). I'm typically not interested in crosstalk measurements, as I have never seen a device where that would've been an actual issue, but since we could have 40 signals hooked up to this console, crosstalk might add up and become an issue on a quiet channel with lots of gain. This is very unlikely.



During my tests of the ADC I noticed a wierd phenomenon: When I gave it an input signal SINAD would be 80db, but improve to 90db withing half a minute or so.
Leave it alone for a while and you'll measure 80db again, which then improves to 90db just like before. Leave it on for longer or start driving it harder and it'll get worse again. Maybe this is temperature related?

Conclusion: The X32 revolutionized the live scene, as it more or less kicked off the era of affordable digital mixing consoles, which were easy to use at the same time.
I've started working with this console 7 years ago and immediately liked the work flow and layout. My current desk has possibly 'blessed' nearly 100 thousand ears over the last few years and never has anyone complained to me, that the desk (specifically) sounded bad (after the show was over). I just heard that comment once before a show, by a sound engineer who quickly disappeared after his show on his Yamaha QL5 sounded pretty bad compared to mine on the X32. Classic.
Measurements look pretty good if you disregard the distortion of the input section. Is the distortion bad, is it an issue? No. I'm much more bothered by the compressor in the channel strip, which often doesn't react as fast as I'd like (even on 0ms attack time it still sounds like 5ms). This console packs one hell of a feature set into one device. I use it with a X32 rack as stage box. The desk stays at home for smaller shows and I just take my tablet with me. Convenience and ease to use it very important imo and the X32 got that down.
Until I get booked for jobs that pay well enough to justify a DiGiCo S21 or something Soundcraft, I'm gonna keep the X32. It's a good ecosystem and performs well, if you know how to work with its limitations. Measurements revealed that you won't have to worry much about the technical performance of the ADC or DAC. If it sounds bad the fault lies somewhere else.

Merry almost Christmas y'all, have a nice holiday season :)
 
OP
peniku8

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397
If anyone wonders: afaik the entirety of the X32 line share the same internal components (including the stage boxes), while the XR line seems to use different stuff. At least I got the impression that the XR18 I used had a higher noise floor (because it was the only device I ran into noise-related issues with). Apart from that, the XR line also has less processing power and features. When I worked with it, I missed a few things I had in the X32, but can't quite remember what it was, since that was several years ago.
Happy new year, all! :)
 
OP
peniku8

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397

weesch

Member
Joined
Apr 28, 2021
Messages
79
Likes
10
The X32 Core should accept the X-ADAT card
Hi
You are really sure which accepts the card adat ?
Tu es sur qu'elle accepte la carte adat ?
I think i will ask the question to behringer ....
Je pense que je vais poser la question a behringer....
parce que si c'est le cas ça serai une bonne option pour remplacer ma 01v96 yamaha ....
because if it is the case it could replace ma 01v96 yamaha.
thnaks for your reply
best regards
weesch
 
OP
peniku8

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397
Hi
You are really sure which accepts the card adat ?
Tu es sur qu'elle accepte la carte adat ?
I think i will ask the question to behringer ....
Je pense que je vais poser la question a behringer....
parce que si c'est le cas ça serai une bonne option pour remplacer ma 01v96 yamaha ....
because if it is the case it could replace ma 01v96 yamaha.
thnaks for your reply
best regards
weesch
I have not tried it and neither am I a Behringer employee, so I can't really say for sure. I don't have an X32 core, since I have an X32 rack.
 

weesch

Member
Joined
Apr 28, 2021
Messages
79
Likes
10
Hi
Oky
but if it work on x32 rack it should work on x32 core...
the better i think
is to ask behringer if it work...
Thanks for reply
best regards
weesch
 

Eldus

Member
Joined
Jun 8, 2021
Messages
58
Likes
52
Thank you so much for your measurements! I would wager that most DI boxes have higher distortion than the ADC. I agree that not being able to use the channel compressor as a true limiter at 0ms attack is annoying.

Would you be interested in testing the effect some of the FX like distortion of the classic compressor emulations without GR? I would love to try paralell bus compression with the faircomp but the added latency means taking up another half of an FX slot to avoid comb filtering.

Or the Pultec clone with all knobs at default? I wonder if the GEQs add color.

I am also passivley curious about the headphone monitor out on the console. I suspect it is not stellar based on my subjective listening but I could be wrong. Might be differences between models as well.

Sorry about all of the questions.
What FX do you like using?
 
OP
peniku8

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397
I might test these things when I get the time to. The headphone out feeds my Moondrop Aria pretty well, I've never had any issues with it.

As for FX that depends on what I'm doing. If I have stage monitors and no separate monitor desk I have two slots occupied with GEQs for 4 wedges to fight feedback.
The usual: hall verb, plate/room verb, delay, GEQ (PA)
The FX that deviate from the regular set of FX are probably the multiband compressor (insert on main vocals), the de-esser (usually insert on the vocal bus), suboctaver (send effect, great way to enhance any drum that needs more oompf or to add some extra sauce to a bass guitar), a 2nd compressor for the bass guitar (insert), if I want to side-chain the channel dynamics with the kick drum, or the guitar amp (send effect, also good for extra sauce on bass guitar, sometimes vocals).
For my own band's IEM rack I use the guitar amp sims as guitar (and bass) amps and use the TruEQ to approximate an impulse response loader. Without the EQ, the X32's amp sim (high gain) sounds like trash, but with it it sounds surprisingly decent.
 
OP
peniku8

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397
@Eldus I had the desk readied for a gig which was cancelled last-minute, so I went to measure some other aspects of the console, including some of the effects.
I tested all of the effects in slot 8, which means I couldn't test the CPU intensive effects like the multiband compressor, but my testing showed that unless an effect is based on some hardware or specifically distorts the signal with an adjustable parameter, it just passes the signal through without degredation (almost).


The first thing I tested was the desk's internal generator, because I was too lazy to connect it to the PC and integrate it into flexASIO again.
This failed miserably, as the internal generator performs way worse than REW through X-USB:

AuT5vnr.jpg



So I grabbed a usb cable and set it up with FlexASIO and sure enough, we were back to almost 100dB THD+N at 4V output.
Edit: the headphone output measurements were flawed because I used a 3pin "balanced" cable into the ADC, which nullified most of the 'badness' in the signal. I noticed this when I went to take the power measurements, so here are the all-new ones:
Headphone output at 4V, with no load:

8dANmGy.jpg



360Ohm power is about 120mW (that's all I could find in resistors in that range and it includes the ADC's impedance in parallel) (one channel driven):

I7DwLBg.jpg



33Ohm power is just shy of 200mW (one channel driven):

qF22dXU.jpg



Onto the effects section. First I tested the channel EQ. Wierdly enough, it didn't affect the noise spectrum nor distortion, until I started putting filter below 100Hz. Worst I could get it was with all (6) EQ points maxed (all maximum negative gain, no boosts!) between 20 and 100Hz:

tuwancd.jpg



I have the Stereo TrueEQ on my PA bus, which performed the same as the regular GEQ (settings didn't change the noise much at all):

FqNYOoH.jpg



The Xtec EQ didn't add any colour, unless you switched the "Transformer" switch on:

eXvExI3.jpg


GKniA80.jpg



Now follow a bunch of other effects, which I measured to show what kind of 'colour' they add:

jONwVvZ.jpg



q7IGSYd.jpg



RdSWS9H.jpg



ZPYCBXg.jpg



KGvyns0.jpg



Eo9tfmF.jpg



VCqyy4y.jpg



FM0bKw5.jpg


I hate the fact that the SubOctaver adds such an extreme amount of harmonic distortion. The fact that we have the 2nd harmonic in the -1Oct signal, means that's just the dry signal again (even when turned to 100% wet!). This annoys me literally every time I use that effect, since I have to cut those frequencies in the FXreturn channel per EQ. That's a terrible oversight imo.
 
Last edited:
OP
peniku8

peniku8

Active Member
Joined
Dec 9, 2020
Messages
232
Likes
397
I went to measure power for the headphone output and added them to the post above. I also noticed a mistake in my connection when taking these measurements and fixed that, which I also described in the post above.
 

Eldus

Member
Joined
Jun 8, 2021
Messages
58
Likes
52
I might test these things when I get the time to. The headphone out feeds my Moondrop Aria pretty well, I've never had any issues with it.

As for FX that depends on what I'm doing. If I have stage monitors and no separate monitor desk I have two slots occupied with GEQs for 4 wedges to fight feedback.
The usual: hall verb, plate/room verb, delay, GEQ (PA)
The FX that deviate from the regular set of FX are probably the multiband compressor (insert on main vocals), the de-esser (usually insert on the vocal bus), suboctaver (send effect, great way to enhance any drum that needs more oompf or to add some extra sauce to a bass guitar), a 2nd compressor for the bass guitar (insert), if I want to side-chain the channel dynamics with the kick drum, or the guitar amp (send effect, also good for extra sauce on bass guitar, sometimes vocals).
For my own band's IEM rack I use the guitar amp sims as guitar (and bass) amps and use the TruEQ to approximate an impulse response loader. Without the EQ, the X32's amp sim (high gain) sounds like trash, but with it it sounds surprisingly decent.
Cool. I am mixing for streaming. Currently using Dessers inserted to each vocal and using the Faircomp on the bus. I do use the channel comp on each bus as a light limiter that I make sure to not engage often. I love to use the Mid/Side faircomp variant on my stereo drum bus which allows for more control with a single slot used. I use the Rich Plate for vocals and the stereo delay set shorter with vocals and some instruments lightly into it. I have found for more natural compression for a live stream using paralell and serial compression in combination less aggressively than you would with just one gets good results with our band. The in channel comp Mix knob is a nice feature I use with vocals esp. I should look into using the multiband comp. I also use the Extec EQ on Bass guitar. Sometimes the LA2A comp. I am still learning.
 

Eldus

Member
Joined
Jun 8, 2021
Messages
58
Likes
52
Thank you so much! I never thought I would get to see measurements like this of an x32. I bet the guys over at Gearslutz. com would find this interesting.

Thank you!
 
Top Bottom