The No.1 Website for Pro Audio
Sonarworks Reference 4

Sonarworks Reference 4 Studio

4.4 4.4 out of 5, based on 3 Reviews

Sonarworks Reference 4 Studio Edition with Microphone is a complete and easy to use room and monitor correction system that includes a headphone correction mode and profiles for over 200 headphone models.

9th November 2018

Sonarworks Reference 4 Studio by Sound-Guy

Sonarworks Reference 4

Sonarworks Reference 4 Monitor/Room and Headphone Calibration System

The Sonarworks Reference 4 system is one of several available sound calibration systems that aims to correct speaker/room variations in frequency response, and in addition has a headphone mode with a library of generic corrections for over 200 headphones, as well as a unique service to calibrate your headphones to provide extremely accurate and repeatable calibration. There are several versions of Reference 4 available, from software-only to software with a calibrated measurement mic (that I review here) to the full blown Premium Bundle that also includes an individually calibrated Sennheiser HD650 headphone. I set out to see (hear?) if it does what it claims.

The Claim
Sonarworks has researched sound reproduction and the physiology of sound in humans to determine how perception and sonic spectral distribution relates. As they state in their Technical Overview of June 2018, “Human beings do not perceive sound the same way a measurement microphone does. Whereas for a measurement mic the room’s frequency response changes significantly as it moves around, it normally sounds much more even for a human listener due to brain interpretations of the sounds that we hear.” And as every mix engineer knows, “The acoustic properties of a music listener’s playback system more often than not have little similarity with the setup the artist used when creating the song. Some of the music always gets lost in translation.” Anyone running a small project studio is no doubt familiar with this and has likely dealt with the mix translation issue by listening on multiple speakers and/or headphones. Reference 4 aims to remove unwanted coloration from studio speakers and headphones to minimize tonal distortion in your studio to improve translation to other listening environments.

What Do You Get?
The Studio Edition arrived in a well-packed shipping box . . . from Latvia! Since I am in the middle of the USA, I was surprised how quickly it arrived. The product packaging looks very professional and has the expected “information-advertising” displayed on the product box itself (which has a magnetic latch). The product box contained only two things: an XREF 20 microphone and a card with the license ID on it for the software. I actually approve of not sending DVDs or USB drives unless the software is huge since you inevitably need to update it anyway! The software download was easy, and the instructions for activation after installation were clear for both on-line and offline activation. Since my studio is offline unless I run a ten metre cable from the office, I always try to use offline activation. It worked well (unlike some other programs I’ve installed) and I was ready to measure my room and check out the correction plug-in. I did not try the online activation, but in my experience that always works better than offline procedures.

The XREF 20 mic is a slim measurement mic with a small diaphragm condenser element that requires the usual 48 Volt phantom power. There is no mic clip included, which wasn’t an issue for me since I have at least three small condenser mic clips that fit fine if I need a clip, and the procedure, as you will see, can be run holding the mic by hand. Also, there is no storage case for the mic other than the nice box it comes in, which is larger than necessary to store one small mic. However, since I have three other measurement mics, and only one of them has a dedicated storage case, I now have a box holding three measurement mics in my mic closet.

While the frequency response of the mic itself is not perfectly flat (which would make it cost much more) it has been tested at Sonarworks and an exclusive calibration curve is available. I expected this to be included as a chart, but it is actually obtained by a download which includes a txt file and a proprietary swmic file. The text file shows the on-axis frequency response of the mic, and mine had a notable bump centered at about 9 kHz, but using the calibration data (which has a resolution of 0.01 dB and a claimed absolute accuracy said to be better than 1 dB) should result in a corrected frequency response about as good as the best costly measurement mics out there. I compared the corrected response to a more accurate measurement mic I own and it appears to be “spot-on”. The swmic file is used in Reference 4 (R4 from now on) and is actually an off-axis calibration (30 degree) that R4 is calibrated to use since the mic is nominally pointed between the speakers and picks up off-axis sounds in use.

Measurement Procedure
Sonarworks is designed for use by anyone, novice through professional, in terms of the set-up and measurement processes - it leads you through each step with audible explanations (spoken in English in my version) and plays beeping/clicking signals before each measurement is made, although this guide can be turned off. First it has you specify the mic being used and the correction file if you have one. If you already have a flat measurement mic, you can specify the mic to be used without a correction file, but as mentioned, the Studio Edition comes with the XREF 20 mic and instructions for downloading its calibration file based on its I.D. number.

This download worked quickly and when I entered the provided swmic file the program showed the mic data

Display of swmic curve provided for my XREF 20 mic

After the mic is specified you are ready to make the actual measurements. Note that if you use Windows the measurement program normally uses an ASIO driver, but can be switched to a WASAPI Mode if you don’t use ASIO.

The Settings button in the lower left opens a window that enables viewing license info and setting a few parameters. You shouldn’t need the Debug Mode, but it’s available if you have problems. You can turn the Tutorial explanations on or off here (good to run first with them on) or later in the program. And you can select from three different Microphone localization signal types, although this adjustment does not affect the measured frequency response of your system. The signal type only affects how the measurement software locates the microphone, and if the software has trouble detecting the location of the microphone, you can try changing the signal type to one of the other modes. I had no problem with the default setting.

Settings window

Measuring Up
There are 37 measurements required for the room/speaker analysis, which Sonarworks has determined is a minimum number needed to “map” the speakers in a room to better than 1 dB of uncertainty. Before running the 37 measurements there are brief set-up measurements which determine, strictly from sound, how far apart your monitors are. This is all explained as you proceed. The measurements came within a centimeter of the actual center to center measurement of my speakers! Next it determines how far your “sweet spot” is from each speaker (you place the mic where the middle of your head would be). That was also spot-on. If these measurements come out a bit off, you can easily adjust them on the screen. I use a subwoofer for final tuning of most mixes, and R4 works fine with one as long as it has been properly balanced against the main speakers. There is no need to make any special set-up measurements if a subwoofer is employed. After this set up process the fun begins.

The 37 measurements start with one at the sweet spot and have you then move about the room, positioning the mic left and right, forward and back from the sweet spot. They suggest hand-holding the mic, which leads to a few concerns since your body will change the sound field in your room (unless you have a very large studio!) and it’s a bit of a challenge to keep the mic at the same “ear level” height, although variation of 10 - 20 cm in height should not cause a problem. You certainly should not stand between a speaker and the mic - I expect it’s best to stay behind the mic, that is between the mic and back wall, so that you only affect reflections off that wall. Since low frequencies (below a few hundred Hertz) are effectively not blocked by your body and are usually the biggest problem in small rooms, this should minimize any effect of your body during measurements. And although the software and manual indicate “It is not necessary to use a microphone stand to make measurements” a Sonarworks Zendesk FAQ says, “For the most precise results we suggest using a microphone stand during the measurement process.” I tried it both ways, and results I obtained were nearly identical between the two methods.

While this measurement process is easy in concept and guides someone without experience through the measurements, I’d like to be able to position the mic first and then request a measurement, at my own pace. As it is, it’s difficult to pause the measurements when you are standing across the room.

When the measurements are completed, you are congratulated with a “Nice job” message and the measuring window shows all the actual measurement positions, which will rarely be right in the center of each test zone, but the software uses timing signals to figure exactly where the mic was and completes its calculations accordingly. Unlike some room correction systems, Sonarworks leaves no doubt where each measurement is to be made, which may be considered a positive or negative feature depending on your needs.

Completion of measurements window

And the Results Are
Clicking “Show Results” brings up the Results window:

Results window

This Results window shows lots of details and moving your cursor in the plot shows details at any frequency. My small treated room is fairly even above 200 Hz with small +/-2 dB variations, but due to standing waves (and in spite of several bass traps) has significant build-up below that frequency. At the bottom of the Results window the average difference of the left and right speakers is shown. In my case the right speaker level in down 0.18 dB from the left - nothing to worry about! At this point you can save the measurement data with a profile name for use in the R4 plug-in. If you use a subwoofer for some work and not for other projects, you might want to include that in the profile name, and make a separate set of measurements without the sub. It will be drastically different than the profile with your sub. When the profile is saved you close this window, start your DAW, and insert the plug-in as the last device in your master bus. Or you can launch the Systemwide application, if it works in your system (see comments later).

Reference 4 plug-in window

Comprehensive Analysis and Correction Control
As you can see in the figure, the R4 plug-in window shows a lot of information. The relative sound level from each speaker is shown as before (small orange window, this time to the nearest tenth of a dB), and the measured response of each speaker in the room (“Before”) and a “Simulated After” correction is shown (purple line). Note that the “Simulated After”plot for my room shows some small bumps below 100 Hz due, I expect, to standing waves in the small room. R4 attempts to correct speaker variation and room contributions, but strong standing waves may result in some residual peaks and dips, though greatly reduced from the uncorrected response. Without correction these bumps measured 8-12 dB more than with correction (confirmed using a measurement mic measuring pink noise). That’s about ten times the energy of the corrected level. R4 reduced the bumps from a major anomaly that really “clouded” the sound to a minor level variation that isn’t really noticeable in a mix.

You can view several other measures including the applied correction curve, the target curve, and filter phase (if you use the zero latency filter mode - there is also a linear phase mode that introduces latency and eliminates phase shift). There is a Safe Headroom control that attempts to prevent over driving your speakers or headphones when significant gain has been required at some frequencies - you can determine if it is needed and turn it off if you wish since it reduces the outgoing signal level. A Limit control adjusts the amount by which the Reference software is allowed to boost when correcting your speakers, a frequency related limit different than the Safe Headroom. There is a Mono button which does just that, and a Bass Boost and Tilt control as well as a Predefined Target Curves. The Bass Boost and Tilt do as expected, and gently boost or cut frequencies up to 6 dB (while the plug-in still corrects for wild swings in frequency response) but the Predefined Target Curves needs a comment. Actually, there are only two predefined curves, B&K 1974 speaker target and X-Curve. These curves are applied “on top” of the Reference 4 correction like the Bass Boost and Tilt, and the B&K curve supplies a specific bass boosted mode which has been used over many years in audio installations for more natural sounding audio. The X-Curve is an old standard (ISO 2969 or ANSI/SMPTE 202M) used in movie theater sound management and by mixers in the film & TV industries.

There is currently, as I write this, a concern over missing simulations that were formerly a major feature of Reference 4 (4.0.x and earlier). These simulations changed the frequency response of the speaker correction to ‘match’ a number of well-known speakers such as the (in)famous Yamaha NS10. The feature was removed due to worries over legal issues - yet you can still download a previous version (4.0.115) that includes this feature. It’s not something I really miss since I have real NS10 M Studio monitors. However, other speaker types might be useful to have and Sonarworks are hoping to bring this feature back.

Does It Work?
In short, Reference 4 works as expected as long as you don’t expect full correction for strong room modes, ringing and long signal decay caused by a poorly treated (or untreated) room. As Sonarworks themselves say, “In an ideal world the After curve would coincide with the Target curve. However, depending on the capability of your speakers, the correction limit settings, and filter phase type setting, the After curve could differ from the intended Target curve”.

As Sonarworks stated, “human beings do not perceive sound the same way a measurement microphone does” and the proof of the pudding is in the eating as they say. Listening tests with the Reference 4 plug-in switched in reveal much improved clarity and a ‘flat’ response over the full audio spectrum. However, while flat is a good target for consistency (drum roll please . . .) flat does not sound all that “hi-fi”! Especially the lowest octaves. This is of course due to the Fletcher Munson effect, that humans do not perceive the lowest octaves (or the highest) as being as loud as the midrange when played at the same Sound Pressure Level (SPL), and thus almost everyone ends up listening to audio using a “smiley” curve on their EQ! Unfortunately, everyone has a different idea of how much to boost the lows and highs, so the final consumer of music, the listener, may change the frequency spectrum from what the artist/mix and mastering engineer intended.

One solution for the Fletcher Munson effect is to listen at a very high level since the higher the SPL, the flatter the perception across the audio spectrum. BUT DO NOT ATTEMPT THIS if you value your hearing! I use 74 dBC SPL per speaker with band-limited (500 Hz - 2 kHz) pink noise as my reference level. Guidelines such as proposed by Bob Katz, indicate a range from the low 70’s to mid 80’s depending on room size (louder for larger rooms because a larger room needs more energy to sound as loud as less energy sounds in a small room).

R4 includes the Boost, Tilt and B&K curves to boost the bass/treble if it feels weak to you. If your room has significant standing waves in the 40-200 Hz range (which is likely), turning on the R4 correction will at first appear to “cut off the bass”. But it is cutting false bass signal peaks that will not translate well to other rooms, and possibly cause you to reduce level for a frequency that in another room may already be a dip, creating a horrible frequency imbalance. And other frequencies may dip in your room and peak in another, again leading to poor mix adjustments. With R4 engaged, these false peaks and dips are corrected, and mix decisions will translate much better to other rooms since you will not have exaggerated any false peaks or dips.

I sometimes use a separate EQ after R4 to create a “smiley” curve, and this does not negate the correction of R4 for dips and peaks in short spans of the spectrum, but will shape the overall audio spectrum like Boost, Tilt and the B&K curve do. I do this to get the bass “shaking” for clients who expect to feel a good bass rumble, but during mixing itself I like to keep things “flat”. And as always, it’s good to listen to reference tracks when mixing. Using R4 will assure the final balance is good from the lowest to highest frequencies.

Does It Work Part 2
In addition to the Reference 4 DAW plug-in, there is a module called Reference Systemwide intended to be used for all audio use in your computer. It installs a virtual audio driver on your device capable of calibrating all outgoing audio from your system. Unfortunately, Systemwide does not work with an ASIO driver in spite of the R4 measurement program using ASIO as the default! This is an unfortunate situation, and it would be useful if it could be fixed, but if your audio interface uses ASIO like mine, the current version does not support it. When I tried running Systemwide, some of my audio programs had their driver assignments changed, and I had to change them back to ASIO before I could use them again. So be warned if you use an ASIO driver that Systemwide is not ASIO compatible. In most project studios this is likely not to be a concern since your DAWs and audio editors can use the R4 plug-in.

And There is More
One great feature of Reference 4 is that it includes the headphone correction mode and “average” correction curves for over 200 headphone models. These curves are estimated to be accurate within 3 dB, but if you want to really “get flat” you can send your headphones to Latvia for an individual calibration. If you live close enough, you might just want to take them to Latvia since it is a rather beautiful country from what I’ve seen! I ran tests using my AKG K240 Studio cans and was surprised at how relatively muffled they sound without R4 correction! I was also fascinated trying other headphone curves while listening to the K240’s - you can switch calibration profiles while listening and there are no pops or clicks - it just switches quietly to the new profile. This is quite educational since the variation in tonal quality is very pronounced, though it doesn’t actually replicate how another headphone would sound listening through yours - that might be a nice feature for Sonarworks to add in the headphone mode!

At any rate, the headphone mode is very useful and will provide even more consistent correction for different headphones than the speaker correction can provide in different rooms since with headphones there are no room reflections and standing waves to interfere with the sound. Note that R4 remembers both your last speaker profile and your last headphone profile, and you can switch instantly between them.

In Summary
Sonarworks Reference 4 provides an easy to follow method to measure the response of your speakers in your room, and a flexible plug-in for use in your DAW or audio editor. It also provides a large selection of “averaged” calibration curves for over 200 headphone models that provide correction to within ~ 3 dB across the audio spectrum, and the option to send your own headphones to Latvia for a precise calibration. The Studio Edition with Mic includes a measurement mic that has been individually calibrated, so you get a mic that will be useful for various measurement tasks, and can be used for audio recording if you don’t need a very quiet mic (small diaphragm mics tend to have relatively high self noise). The XREF 20 is rated at a sensitivity of -37 dB/Pa (14 mV), 24 dB self noise level with a snr of 70 dB, maximum SPL of 132 dB, and a dynamic range of 106 dB.

On the not-so-good side, the Systemwide module does not work with ASIO drivers, which is unfortunate. However, not a deal-breaker for me. Another minor ‘minus’ is the XREF 20 mic comes without a mic clip, but since the measurement procedure can be accomplished holding the mic by hand, this is not a real concern for R4 use.

No software solution can solve significant acoustic problems in a room - strong standing waves, long ringing decays, and faulty speakers cannot be fixed using Sonarworks or any other software solution. However, if your room is reasonably treated with sound absorption and diffusion, and sounds “pretty good” already, Sonarworks will take you to a new level of audio clarity. The headphone correction mode is excellent, and the XREF 20 mic with it’s factory calibration curve is an excellent measurement mic if you have need of one for other activities (don’t we all!).


  • Easy to use
  • Includes 32 bit and 64 bit versions
  • Results are impressive, especially in a room with moderate acoustical treatment
  • Studio Edition with mic includes an individually calibrated microphone for precise measurements
  • All editions include headphone calibration mode with general calibration data for over 200 models
  • Individual headphone calibration available (but must send to Latvia at this time).
  • Boost and Tilt or Predefined Target Curves can be used to adjust tonal balance to taste
  • Systemwide standalone module works outside of a DAW host if you don’t use ASIO drivers

  • Systemwide mode will not work if your audio interface requires an ASIO driver
  • Fixed number of calibration measurements in a predetermined area may not fit everyone’s needs
  • At install, it loads plug-ins in default folders that are likely not where you want them unless you click on 'Customize' under the Install button before you select the product to install, which is not very intuitive
  • Current version is missing speaker simulations that were previously a touted feature

I started testing ver and updated to when it was released during the testing period. I had no issues installing the new version, it just ran and recognized my calibration curves after the update.

Mac (OS X 10.9 or later), or Windows (WIN 7 SP 1 or later)

Plug-in formats: AU, AAX Native, RTAS, VST

Sound Calibration Technology | Sonarworks

Software is always a download, even with the boxed edition with microphone.
Reference 4 Headphone Edition - Download $99 list price
Reference 4 Studio Edition - Download $249 list price
Reference 4 Studio Edition with Microphone - Boxed $299 list price
Reference 4 Premium Bundle with Microphone and Pre-Calibrated Sennheiser HD 650 Headphones - Boxed $699 list price

Attached Thumbnails
Sonarworks Reference 4-mic-cal.jpg   Sonarworks Reference 4-ref-4-measure-settings.jpg   Sonarworks Reference 4-ref-4-mic-final.jpg   Sonarworks Reference 4-ref-4-results.jpg   Sonarworks Reference 4-ref-4-plugin-1.jpg  

Last edited by Grahamdwc; 9th July 2019 at 08:01 PM..

  • 5
9th April 2019

Sonarworks Reference 4 Studio by OK1

  • Sound Quality 3 out of 5
  • Ease of use 4 out of 5
  • Features 3 out of 5
  • Bang for buck 4 out of 5
  • Overall: 3.5
Sonarworks Reference 4

I tried Sonarworks Reference 4 Studio (SR4S)on

1. Alesis M1 Mk 2 - a soft(silk) dome tweeter, with 6.5 inch woofers. Active speakers

2. Alto TS310 PA speakers - a compression driver tweeter and 10 inch woofer. Active speakers

3. AKG 702 headphones using the average profile provided.

I used a Dayton Audio EMM-6 measurement microphone for 1 and 2, which comes with an individual calibration file. (minor modification of this file is needed to provide a suitable format compatible with Sonarworks), and Reaper 5 is my DAW.

In general the process of conducting the measurements is easy or difficult, depending on your experience with audio gear and room correction software.

I'll focus here on the results. As there is another review on gearslutz from which you can glean the gory details, no point repeating this here.

With the headphones especially when comparing them to the result of the speakers, it highlighted the fact that headphones are IMHO seriously compromised in the region above 8Khz, and you cannot correct what it not there..No amount of correction can fix the compromised high end..

It also calls to question the physics - how credible is a single dynamic driver on a headphone capable of full frequency presentation - it simply is not.

So SR4S simply highlighted the fallacy - and so many of is invest in these things, expensive headphones - studio grade, bought under advice from highly respected magazine and forum reviewers - and I wonder - what the heck are these esteemed people hearing...

The sound quality on the headphones was better than default - no doubt, but was it good enough for more than listening e.g actually mixing a track with it - not so. I will never ever bother attempting any mixing on headphones - the resolution and flat frequency response is not there - SR4S could not perfect this turd.

With the other speakers it became clear to me that my beloved Alesis also bought on good advice of an esteemed magazine reviewer, does not cut the mustard anymore. The drivers on a PA speaker designed more recently were better, had less distortion better frequency response by default, and with SR4S. the ALto's sounded better.

So room/speaker correction depends on what you give it to work with, its the final icing on the cake, not the cake itself.

In comparing SR4S to another room correction method it became apparent that SR4S gives focus to correcting certain areas of bass - what I would call the upper bass, between 80 and 250 hz, while lower bass below 100hz is recessed.

This gives an impression of a heightened bass

And with the higher frequencies anything above 12K is recessed and focus is given to frequencies between 4 and 10K.

So you end up with a smilyish EQ, with tapering off a downturn at the extreme ends.

It's a nice educational tool, which I would not use for critical mixing, on any speaker cos what you hear is almost like listening to a budget hi-fi which was trying too hard to please..

Thankfully you can demo this for 21 days (20 really) to form your own impressions. A measurement microphone is needed for use with speakers.

I found the tone controls useful, but compared with using other speaker equalisation methods, Felt that the final result was made to sound better than it should, and will lead to lazy mixing that does not translate well into other speakers.

Its obvious to me that Sonarworks attempts to "interpret" what a flat frequency should sound like from a phychoacoustic perspective and IMHO opinion, this is where they fail, I find correction on speakers with SR4S causes me to want to keep the volume lower than a certain threshold. Any higher and it starts to sound shrill and over processed.. It attempts to overcome the Fletcher Munsen effect at low volumes by boosting aspects of the mid/upper bass and the mid highs.... and effect heightened by also lowering the volume of the extreme ends of the frequency that humans can hear. Voices sound even more processed than they should - nice but not accurate. I might use SR4S for playback as a DJ, as an occasional effect, but not for accurate listening or playback...

In contrast using other room correction tools, do not alter this relationship between frequencies, in any psycho-acoustic way, leaving a presentation that is much more natural and believable....

What gives the SR4S away is - rather than providing separation between voices, and instruments, so it is easier to hear each one distinctly, it blends them together making them difficult to hear on their own, almost like you have already mastered it, like you have added mastering eq and limiting, pretty unnatural....which explains the tendency to want to turn the volume down... And picking out voices and instruments in the stereo field becomes more difficult. Compared to other room correction tools, SR4S compromised the stereo image, and the center image of lead vocals was also compromised, as all non centered instruments and voices appeared more centered.. In general the soundstage with SR4S was smaller, even smaller than the uncorrected speaker.

I do wonder if as Sonarworks claims, this software has been bought in the 10's of thousands, I do hope Sonarworks is not responsible for the terrible sounding music I hear increasingly, by altering the truth and presenting a version of truth that makes mixing engineers lazy - thinking the mix is finished....!!

I think in general certain activities which larger studios take for granted, including adjusting monitor eq, sound proofing, acoustic treatment, have become oversold as being easy to achieve in the small studio, leading to a whole new profitable industry for some, selling dreams that are rarely realised in full. Especially where software is involved which cannot be refunded.

I had high hopes based on all the advertising and positive reviews.

Unfortunately this has made me question everything I have read not just about Sonarworks, but other products reviewed in esteemed magazines. And the need for positive, unbiased reviews abounds, in a purist format, that has absolutely to dependence on advertising revenue from music technology manufacturers...

Here's one of the most telling aspects of the Sonarworks psychoacoustic alterations - listen to any section of audience handclaps on a live event. It just sounds many respects audience handclaps are a bit like noise, so you know what it should sound like, but via Sonarworks correction, you know that something is not quite right with those handclaps......

Or listen to an acoustic piano recording, jazz or classical, typically any instrument whose tones are not too complex like a vibraphone, or xylophone and you know straightaway, that Sonarworks is definitely bending those overtones......beyond the truth - creative license.

  • 2
Review Tools
Search this Review
Search this Review:

Advanced Search