This page aims at providing the procedure, raw measurement datasets and results gathered to prepare the following comparison video:

https://www.youtube.com/watch?v=NYbtzWdxxuo

This document proposes for this purpose

• a test procedure
• RAW measurements
• Test results comparing Behringer XR18, Midas MR18 and Behringer WING performance
• An XR18 filter configuration to compensate its frequency response

Starting with a general statement: Measurements are done as close as possible to what would be a scientific measurement process (Configuration and tests are documented, are reproducible by a third party, tests results are published, known assumptions and limitations are mentioned as far as possible).

Now the claims:

• Statement #1 : Measurements are done in recording studio environment. Namely, there is no control on CEM/EMI interfering with measurements apart from standard studio installation best practices in terms of proper ED grounding and shielding.
  • As a consequence, comparing the results is more important that dealing with the RAW value.
• Statement #2 : One input and one output are used on each unit, there is only one pair of interface sample tested. Having a pair allows to check for a similar performance on both interface, to rule out obvious erroneous failures on one interface. It is understood that statistically processing those measurement among several interfaces and unit S/Ns would be much better, but is out of scope of this measurement procedure.
  • As a consequence, these results are (1) immune to critical failures on one channel but are (2) prone to specimen bias.
• Statement #3 : XR18 has two high impedance instrument inputs which MR18 does not have. For the sake of comparing apples with apples, Input 9/10 and Output 1/2 are used as references on both units.
• Statement #4 : There is a 10 year gap between XR/MR design and WING design. Consequently, the analog part is significantly less noise-sensitive on WING which directly derives from the technology gap.
  • As a consequence, comparison will focus on analog comparison ONLY, and will not take into account the difference in digital signal processing that units are capable of. The technological gap in this field is even greater (processing bit depth, algorithm quality, etc…).

• UUTs
  • MR18 S211000533C8H Firwmare 1.22
  • XR18 S201004255BI8 Firwmare 1.22
  • WING 010063R0604AAE Firwmare 3.0.5
• Tools
  • Gen2 Apollo Twin USB
  • XLR Wires / Jack Adapters
  • Room EQ Wizard Rev 5.3 for measurement

Measure loopback RTA THD (+N) @1khz + total dynamic range.

• REW in loopback mode,
• Double shielded XLR as connection
• on both units, use In-9 and Out-1

Loopback diragram:

• Set Inputs Gain to 0dB
• Set Output level to match a -20dBfs input level
• Measure RTA and store

• Set Inputs Gain to 0dB
• Set Output level to match a -20dBfs input level
• Measure RTA and store

• Set Inputs Gain to 0dB
• Set Output level to match a -20dBfs input level
• Measure Transfer function and store

• Set Inputs Gain to 36dB
• Set Output level to match a 20dBfs input level
• Measure Transfer function and store

• Set Inputs Gain to 0dB
• Set Output level to match a -20dBfs input level
• Measure Transfer function and store

• Set Inputs Gain to 36dB
• Set Output level to match a 20dBfs input level
• Measure Transfer function and store
• (Note : -35dB is constrained by test setup, because of my attenuator being on/off, I could not reach exactly 36dB gain and -20dBfs signal)

You'll find everything I used here.

Open with Libreoffice or Powerpoint xmw_presentation.zip

Used for normalised dBfs analysis (open with Libreoffice or Excel) xmw_normalised_spreadsheet.zip

Opens with Room Eq Wizard tool: xmw_rew_dataset.zip

Here is an XR18 EQ preset you can load which compensates the poor XR18 frequency response at 36dB Gain.

Of course, you might need those EQ bands afterwards, it's just a modified starting point for those who want to have the sharpest start when configuring a setup.

xr18_compensation_eq.zip