People often ask me to opine on the most critical studio components in which to invest. Many are surprised by my answer: microphones and studio monitors. Great mics never have to be replaced. Therefore they never lose their value. Studio monitors are similar, but you'll never know how something sounds in the consumer world unless you trust your speakers to be faithfully revealing. But even the most expensive monitors aren't necessarily neutral-sounding, nor are most project studios sonically treated to produce a flat environment. That's where Reference 4 comes in, and it can give any speaker or headphone an uncolored response. (More on the headphone aspect later.)
There are four varieties of Reference 4. Premium Bundle comes with a pre-calibrated pair of Sennheiser HD650 headphones. Studio Edition with Mic bundle comes without headphones but includes the Sonarworks XREF20 measurement mic. Studio Edition is the software for headphones and monitors only (no mic.) And Headphone Edition is the software that works with 112 pre-profiled headphones, but not studio monitors. If you buy any version and want to upgrade the software or hardware, you can do so at a reduced cost, and you can buy the XREF20 mic or any pair of pre-calibrated headphones later. All versions include the Systemwide app as well as the plug-ins in AAX, Audio Unit, RTAS, and VST formats, and each license allows installation on up to three computers. (I installed it on both edit bay machines, as well as the laptop with which I travel.)
Audiophiles like me like to listen to music with a high-performance DAC (digital-to-analog converter) and a pair of open-back headphones. When I take time to listen critically to my music library (something I try to do 3 to 4 hours a week), I'll plug my Oppo HA-2SE DAC into my computer, then choose a pair of headphones. I like the Beyerdynamic DT990 Pro 600 ohm headphones. They're quite bright above 12kHz but have a richer low end than most open-back models. I also have a pair of Philips SHP9500S cans. They're not quite as bright, but they sound thinner in the low end and a somewhat more boxy sound, overall.
Correcting headphones is more standardized than speakers because the headphones always sit on the head of a human, so there are no other environmental variables to consider. To that end, Reference 4 currently comes with 112 popular headphone calibration profiles. Referring to the top of Figure 1., you'll see the measured response of the DT990 Pro headphones in blue, and the resulting frequency correction in red. The result is a neutral response that not only sounds wholly balanced but also lets me listen longer without ear fatigue due to bumps and dips in the uncorrected frequency curve.
Putting a speaker in any room will naturally alter its frequency response. That's why you have to profile the speakers in your environment rather than use a pre-defined curve. Profiling requires your speakers, an audio interface, a measurement microphone, and a Mac or PC running Reference 4. The software connects to the audio interface and the speakers attached to it to play a series of pulses that measures distance via Doppler shift, as well as full-range frequency sweeps to analyze the frequency curve of the monitors. The sound is picked up by the mic and fed into the software that processes the results. It then applies the profile to the audio playback of either the entire computer audio system or in any DAW via the Reference 4 plug-ins.
After you've installed Reference 4, you'll find an application called Reference 4 Measure. When you launch the app, it will take you through the profiling process. While you can use microphones recommended by Sonarworks (like the Behringer ECM8000), it's critical that each microphone is individually profiled. After all, if the microphone has bumps and dips in its frequency response (and all mics do), it can't accurately measure for a flat curve. To that end, I strongly recommend getting Studio Edition with Mic because Sonarworks not only makes their own measurement microphone but they also individually profile each one so that you can download the curve of your mic from their website before you profile your speakers. The serial number of the mic is required to download the correct calibration profile, but be aware that the numbers include letters. (I thought the 'I' was a '1', and that profile obviously couldn't be downloaded because it didn't exist.)
After you plug the mic into your audio interface, the software will prompt you through the provisions you'll need make for an accurate profile, including turning on phantom power, turning off direct monitoring, and setting proper input/output levels. (There's a nice lady with a hint of Latvian accent that helps you set levels.)
The measurement process will require that you move the mic around the 'sweet spot' (usually the engineer's 'chair' position) of the room. I was amazed at the accuracy of the software as it accurately measured my monitors as 4' 1" apart from each other. The original measurement of the monitors to my ears was 2' 8", but I manually adjusted it to the actual distance of 2'10". (It's a good idea to have a tape measure to verify the distances.
Since the mic will need to remain at the same height as your ears, I measured 47" from my ears to the floor, then mounted the mic in a stand at that height. (Note: The mic doesn't come with a clip, but any substantial clip will work.) When the measurement sounds are played through the speakers, it's important that your physical body (and the sound-absorbing materials you'll probably be wearing) not be near the microphone. You can program the software for 3, 5, or 10-second delays, which gives you time to get out of the room or at least a good distance from the mic and speakers.
When it comes time to measure around the sweet spot, you may need to take the mic out of the stand. I used the stand as the boom arm kept my body further from the mic. (You don't need to use a stand at all, but I feel like doing so provides a more accurate measurement.) However, when I first profiled my big edit suite, I did so in such haste that I didn't notice the instruction to always keep the mic pointed at the centerline between each speaker. Therefore I had to make a whole new profile. (The results were similar but not identical to the first.)
The sweet-spot analysis is quite dazzling to watch because the software repeatedly uses Doppler shift to measure the left-right / front-back position of the mic. It's a little eerie to watch the on-screen mic move in concert with you. (Note: It does not measure mic angle or height, which is another reason I recommend a stand.)
A total of 24 measurements are taken, and the whole process took about 18 minutes to complete. (16 minutes my second time to correct the mistakes I'd made the first time.) When you're done, save the profile to your computer.
Systemwide is an app that routes your computer audio system through the assigned profile. That means that any general computer application (like iTunes, Spotify, YouTube, etc) can be played through your speaker correction curve. You can also choose between a flat response, or customize the bass and tilt settings, or even simulate different headphones, studio monitors, and consumer speakers included in the app. (Old timers like me will appreciate the "Japanese white cone Studio (sic) monitors" simulation.) You can also adjust the output level, calibration on/off, enable the Avoid Clipping feature, and mix the dry/wet balance.
While the Reference 4 plug-ins can be used in any pro audio or video application, its placement in the signal flow is imperative for proper operation. Essentially, Reference 4 must be inserted at the very last plug-in through which your final output signal flows. You must also remember to disable or remove the plug-in before you render your mix or master because you don't want your speaker or headphone profile to be encoded on the final product. Some programs like WaveLab have specialized insert slots that only process the sound during playback and not during rendering. However, Reference 4 may not automatically appear in those slots, so check your application manual to see if you need to activate the plug-in for those slots.
With my speakers corrected, I could hear a lot more detail than ever before. I recognized the midrange to be smoother and much more revealing. My small system has mismatched components (Yamaha MSP5s with an HS8 subwoofer), and the low-end has always been a little quirky but sound accurately integrated when corrected.
You might already have wondered at which point during your creative process would you insert Reference 4. The answer is all the time. That's because there are three different settings for balancing latency with phase linearity.
If you've just started your audio and MIDI recordings, set the Filter Type to Zero Latency, but the correction won't be phase linear. The Optimum setting has more phase linearity but introduces a little latency, so that setting is best for mixing and on slower computers. The Linear Phase setting adds significant latency (109 ms in my case), so it's best used during mixing and mastering when phase linearity becomes more critical, but tracking has wrapped. You'll find some other nice settings in the Advanced section where you can customize the L/R balance correction (great for monitors that don't have detented level controls), and you can further limit or expand the correction level and low/ high-frequency limits.
I'm a huge fan of speaker correction products like Reference 4. But there are a few things I wish it could do. First, it only analyzes the sweet spot, so it doesn't account for other listening positions (like a client couch or producer's desk) in the control room. Second, while it does have four custom profiles for consumer speakers, you cannot further modify the frequency correction curve other than the bass and tilt. Since I get regular hearing tests and I know what my audiogram looks like, I would love to be able to 'flatten my ears' by entering my own curve. While some programs like IK Multimedia ARC 2.5 do allow for broader area analyses and customizable frequency curves, it doesn't have an app like Systemwide to apply the correction to the entire computer audio output, nor does it work with headphones. Maybe the good folks at Sonarworks could add those features in the future.
If you have relatively cheap monitors and are thinking of upgrading them for better mixing and mastering, you'll have to spend quite a bit of money to make a vast improvement. Even then, your new expensive monitors won't be flat. To that end, I recommend that you download the trial version of Reference 4, which will run for 21 days. Granted, you won't really hear what it's capable of providing to your studio monitors until you've analyzed your control room and saved your custom profile. But there's a high likelihood that your headphones are on the pre-profiled list*, and that will give you a good idea of the advantages to Reference 4. Once you've tried it, you won't be able to imagine living without it.
Price: Software 249 Eur
Software with mic 299 Eur
Premium Bundle 699 Eur
Pros: Excellent correction for both speakers and headphones. Easy analysis process without the need for a manual. Systemwide applies the correction to the entire computer audio output.
Cons: Only bass and tilt user curve customizations. Analysis process for engineer position only.
*If your favorite headphones aren't on the list, you can send them to Sonarworks for individual profiling. The cost is $108.00 including shipping to Riga, Latvia, and usually takes about 10-days. More information at https://storeus.sonarworks.com/collections/reference-4/products/individual-headphone-calibration