True-Fi headphone software correction

A hard sell?

Software retailing for $79 USD that claims a ‘Studio Reference’ standard experience that can be enabled on more than 150 headphone models.

So what does this mean exactly?

As far as I can tell Sonarworks True-Fi is offering curated frequency-response software based on extensive testing of a large cross-section of cans.

It seems the goal is getting rid of the most concerning sonic issues which may plague a particular model; bass bump, attenuated treble, flabby or lacklustre midrange, etc.

Apparently they are EQing the response of the specifically selected pair of headphones (the idea being your headphones) based on a response curve determined by their in-house measurements and by the user inputting some simple parameters such as gender and age.

This seems to flatten out the worst dips or spikes in the targeted frequency response, the software also allows the user to fine tune the EQ to taste.

Personally I’m not a fan of EQing, preferring a straight-up presentation, but I’m not such a brick that I’m not open to having my opinion changed.

Sonarworks is offering a 10-day free trial of True-Fi on their website so you can try it for yourself. Be interesting to hear back from folks on what they think of the software in actual use.

Press release below:

Critical Mass, Studio Sound: Sonarworks Groundbreaking True-Fi Software Now Supports More than 150 Popular Headphone Models

With 27 new headphones supported and a mobile app on the way, Sonarworks is poised to bring ‘Studio Reference’ sound to the masses.

Wednesday, June 6, 2018 — Riga, Latvia — European audio software developer Sonarworks announced that its groundbreaking software platform True-Fi now supports more than 150 popular headphones, including popular models from AKG, Beyerdynamic, Focal, Sennheiser and others.

Additionally, the company has just launched a worldwide ‘Studio Reference’ standard, enabling music producers and consumers to experience a single, common audio reference — whether they intend to create or listen to music.

Further, Sonarworks is planning an imminent launch of its groundbreaking mobile application of its True-Fi software that will support both iOS and Android devices.  

The 27 new headphone profiles brings the number of True-Fi’s supported headphone models to 157, including two new noise cancelling additions: the Bose Quiet Comfort 35 II and Sony WH-1000XM2.

Each headphone has been meticulously measured and incorporated as a complimentary software upgrade for existing True-Fi users, which is available now at  Sonarworks The new supported models include:

  • AKG K553 Pro
  • AKG K72
  • Audio Technica ATH-MSR7
  • B&O Play Beoplay H4
  • Beats Powerbeats3
  • Beyerdynamic Custom One Pro Plus
  • Beyerdynamic DT 770 32 Ohm 
  • Beyerdynamic DT 770 M 80 Ohm
  • Beyerdynamic DT 880 Edition 250 Ohm
  • Bose QuietComfort 35 II
  • Focal Clear Professional
  • Focal Listen Professional
  • HyperX Cloud II
  • JBL E25BT
  • Master & Dynamic MH40
  • Philps Fidelio X2HR
  • Plantronics BackBeat Pro 2
  • Sennheiser HD 206
  • Sennheiser HD 280 Pro
  • Sennheiser HD 4.50 BTNC
  • Sennheiser HD 518
  • Sennheiser HDR 120 II 
  • Shure SE535
  • Sony MDR-XB50AP
  • Sony WH-1000XM2
  • Ultrasone Pro 780i
  • Ultrasone Signature studio

“We are proud to support an increasing number of headphones, and at no incremental cost to the consumer," commented Helmuts Bēms, CEO and co-founder of Sonarworks.

”Frequent travellers will appreciate that True-Fi now supports two of the best noise cancelling headphones we have ever measured: the Bose QuietComfort 35 II and Sony WH-1000XM2. We are also pleased that we now cover the entire Focal professional range, with the addition of the much requested Clear Professional and Listen Professional models.”

Sonarworks' True-Fi technology, which made its debut in the consumer market in November of 2017, removes unwanted sound coloration from headphones so consumers can experience the artists' true intentions — on a sonically balanced listening canvas.

The consumer-based software — which is available on both Mac and PC platforms and soon on mobile devices — is priced at just $79 USD and was awarded 'Best in Show' in the music listening category at CanJam Europe in Berlin last November.

Sonarworks SIA Smerla iela 3 Riga, LV-1006 Latvia

donunus's picture

Is this just an ad? Or is there a review after this? I personally didn't find the effect pleasing. I think the target curve has too much energy somewhere in the upper mids. I've tried it with two headphones and the results were similar.

metal571's picture

Glad to hear I am not the only one. Heard their calibration at CanJam NYC and it was very difficult to listen to any music on it that contained electric guitar, for example. Sounded like a 7506 in the mids.

donunus's picture

I did notice we have similar impressions of things based on your reviews on youtube.

donunus's picture

I think I've figured it out. When the clipping protection button is turned off and you rely on the jriver clip protection instead and turn the gain up full on the plugin, it sounds better. Somehow the lower db of the compensation makes the sound weak and shouty similar to a speaker setup using a passive preamp in sound hmm weird.

Lysdexic's picture

Their VLC plug-in was much better and more customizable.

JohnnyCanuck's picture

The link under Company Info is broken.

buckchester's picture

I think Tru-Fi is a great product. I think it improved the sound of my Hifiman HE400i headphones. It dramatically increased the sub bass, which I thought was over-powering at first, but once I got used to it I would never go back. The sound now sounds full range to my ears.

But I do have a complaint, which is similar to another poster above. I feel like there is a little too much energy somewhere in between 1khz and 3khz. I find this produces a slight "shouty" sounding quality. I would like the ability to manually EQ that area to my taste. They have said they are working on am EQ. I look forward to trying it.

Raffe - are you going to review this product?

MRC01's picture

Parametric EQ using gentle slopes (numerically small Q) can correct gross frequency response deviations without audible side effects.

But, EQ can't correct HD, IM or time domain distortions. And in some cases (EG with big Q, steep slopes) can worsen those.

So EQ is a powerful tool that can be used well, or poorly.

I tried the SonarWorks sample with my LCD-2 and was not impressed. I much prefer the EQ that I apply myself, which is +3 dB @ 4500 Hz Q=0.67. It is just enough to restore some of the LCD-2 upper mid / lower treble dip, making it more neutral overall, but gentle enough to be free of side effects and not change the character of the sound.

KaiS's picture

Time- and frequency response domains are strictly connected, so if you change one the other changes too.
So, correcting frequency response errors using EQ corrects for time domain (often referred to as phase) errors automatically, unless you are using special linear phase equalizers which decouple both.

Unfortunately lots of software equalizers today are built around FFT, where this connection gets lost and where the FFT introduces all kinds of strange artifacts due to it's windowing process. With these FFT based EQs you do not get the expected results, whatever you do, the sound gets worse.

With typical FIR or IIR filter designs this is not the case, although they have their own limits due to the bandwidth restrictions implied by the sampling rate.
Oversampling works around these problems, but takes a further step in the signal processing which needs to be done right.

Not to forget, if you equalize in the digital domain you change the signal's dynamic range. This can lead to annoying nonlinear distortions or not enough output level if you use systems without the option to change internal gain structures, like the usual smartphones.

Looking at all these problems I can understands people saying "I prefer using systems where I don't need to have EQ to make them sound good".

MRC01's picture

That's true, though time domain side effects are related to the slope of the EQ. Parametric EQ with gentle slopes (numerically small Q) minimizes these effects. Of course, gentle slopes apply to broad frequency ranges which limits flexibility. But limiting flexibility to avoid the worst side effects is a tradeoff I prefer. Some headphones have sharp Q problems that can't be corrected with gentle slopes. But in these cases the cure is probably worse than the disease.

KaiS's picture

QUOTE: "Some headphones have sharp Q problems that can't be corrected with gentle slopes. But in these cases the cure is probably worse than the disease."
Not with gentle slopes, but if you hit the frequency and Q of the headphone's peak exactly with your EQ curve you correct both frequency and time-response (if you do NOT use a Linear Phase EQ of course). Acoustic frequency peaks have the same properties as electrical or digital ones and they can compensate each other.
What you need is a tool for measuring, else you will not be able to do this exactly enough.
This can be done with cheap stuff you find almost everywhere:
You can use, e.g., the microphone of an Apple EarPod or EarBud, taped to a table, and a software like AudioTools for iPhone's FFT Analyser.

BTW: a lot of people have a wrong idea what are Linear Phase EQ does: it does not avoid time domain smearing, but distributes the smearing symmetrically to before and after the original signal, while leaving the phase of the signal intact.
Specially in the low frequency range this can sound quite annoying, because there is some audible sound before a drumbeat hits, which is much more unnatural than a longer decay.

DonGateley's picture

Time (phase or group delay) and frequency (magnitude) correction are in fact completely independent. The trick is to use fast (fft based) convolution for correction. A convolution kernel can be easily created from an arbitrary magnitude and an arbitrary time specification. Whether that is what this company is using is unknown by me at this time. The downside to using fast convolution is a delay in the sound equal to the length of the kernel but there exist methods to reduce the delay to zero while still doing fast convolution. However that has been under patent that is now pretty old.

steaxauce's picture

I've noticed I tend to adjust over time to the frequency response of whatever headphones I've been using. A bassy headphone may sound too dark initially, but eventually I'll get used to it, and when I go back to a headphone with a more neutral balance, the new one will sound wrong--too bright--even if the old bassy headphone I'd gotten used to was actually the one that deviated from neutral.

The same is true for other FR deviations. If the headphones you're used to has a dip in the mids, and you switch to something without that dip, you're likely to think the new sound is over-emphasized in that region.

This is especially true when listening to familiar music. We get used to hearing our songs in a particular way, with particular instruments and voices emphasized. If the FR of your usual rig emphasizes the frequency range of one instrument in the mix, and you switch to a more neutral setup, you might find yourself feeling that the new setup doesn't bring out that instrument in the way you like. (Or, you might think, in the way it ought to.)

If you don't like Sonarworks initially, I'd recommend using it exclusively for at least a few days, if not a few weeks, and then going back to your old non-SW configuration. You might find your old configuration is the one that sounds off.

I'm not saying Sonarworks is perfect, but based on my experience with it I'd say there's a 99% chance that what you hear when using Sonarworks is closer to what the artist intended for you to hear than what you get without it.

One thing that SW definitely has going for it is that the results you get are very consistent across headphones. Some headphones will still sound better than others, but this is due to factors other than frequency response, like distortion; the resultant frequency response is always pretty much the same. The big benefit of this is that, if you prefer a sound other than their conception of neutral, you can create an EQ profile to apply on top of SW, and that same EQ profile will work with every headphone in your collection. That's a huge deal!

gzost's picture

I think one of the reasons why some headphone enthusiasts may be reticent to try/accept digital correction of headphone sound is precisely that headphones sound much more similar using it.

To me it seems that part of the fun for these people is to pick headphones to listen to specific genres or even tracks where the headphone's "sound signature" to them fits best. This may then be used as a justification for having a large collection, since the matching of recording and headphone enhances the listening experience (and I'm not doubting it does bring these listeners pleasure).

Of course "sound signatures" are largely just the result of stressing or attenuating particular frequencies, whether intentional or as the best result that can be achieved with a particular hardware design.

To me it would be more rational to have a single headphone which performs at the optimum regarding the technical aspects (e.g. lowest possible distortion), digitally correct the frequency response (in the future also including a user-specific correction curve based on the ear shape etc.) and then apply an EQ to the user's liking on top of that. If somebody thinks that their favourite rock tune is enhanced by bringing the frequency range of the guitars to the forefront then that's perfectly valid. It's just that this is neither hi-fi in its core sense (uncolored reproduction of recorded sound), nor is having multiple (expensive) headphones an efficient way of getting there.

GNagus's picture

So if I use this on Bose Quietcomfort headphones, I am EQ-ing headphones with its own built in EQ

gzost's picture

With any active headphones, all corrections should of course be done at once instead of in multiple stages. Presently, however, that's not the case, so we make do with sub-optimal solutions.

I only had a very brief time with the QC35, so take this with a grain of salt, but my impression was that there is certainly room for improvement regarding their frequency response, so the double equalization here probably makes sense at the practical level.

Mrip541's picture

Sonarworks link in article is broken. It's 2 links in a row.

Rafe Arnott's picture
Thanks for the heads-up. It's fixed now.

It's weird it will only work in that field without https://

Good to know.

sszorin's picture

Did you mean to say that it is Parametric EQ versus EQ ?

Johan B's picture

How does it apply EQ?

gzost's picture

"Personally I’m not a fan of EQing, preferring a straight-up presentation"

I find that what True-Fi does is give me a straight-up presentation.

When listening to classical music, the result across several supported headphones is that instrument timbres are more realistic, instrument separation is improved and that I can generally concentrate more on the music rather than spending energy on mentally correcting the frequency response flaws of the headphone. The difference to me is that large that with anything acoustic I now only listen using True-Fi.

The target curve they are using may not be perfect, but to me it is markedly better than the stock frequency response of the Fostex T50rp mk III, Fostex TH-x00, Shure SE535 and Sennheise HD598 that I use with it.

Since they have a free trial, I suggest that anybody with a supported headphone and computer system give it a try. If you aren't convinced then deinstall and all you've lost is a bit of time. For me it took about ten minutes until I paid for the full version, and I consider this the best value for money purchase I've made in the headphone hobby.

Max_Minimum's picture

I've found that the HD598 compensation works very well for my old HD595 as well. They are very similar after all. I select the 598 compensation and add about 3db of bass. It sounds to me like it's arriving at something like the Harman Target Response curve, possibly with slightly weaker bass.

pwjazz's picture

It has a mobile version called "Morphit Mobile" that integrates with USB Audio Player Pro. UAPP is $7.99 and the Morphit plugin is another $3.99, so for anyone who has Android and is interested in trying out this sort of thing, this is a much cheaper option. Morphit is also interesting because in addition to providing a corrected "reference" curve, it also allows you to simulate other headphones. So you could simulate the frequency response of an HD800 while using your HD600, or perhaps more usefully do the reverse :)

南开米饭's picture

Tyll is a genius

carl99's picture

I have it and I liked it.
That's the past tense.
It quit working when TIDAL did its most recent update.
I contacted Robert at Sonarworks to ask for a cure...that was several days ago, and I haven't gotten a response as yet.
so beware of their customer service...assuming they have any.

buckchester's picture

I have had excellent customer service from Sonarworks. I have sent many emails about various things. Most were replied to promptly.

Rafe, i don't think you need to engage the detractors as much as you're doing. Obviously, if you do a good job, your work will speak for yourself.

I did ask a question in an earlier post that you have not replied to, so I will ask again: do you plan to review Tru-Fi?

Rafe Arnott's picture
Hi buckchester,

Yes, I was planning to use it and compare it with several headphones once I've received part of what made up the WoF. I've still had no cans shipped to me yet.

GimmeCans's picture

One reason some may have issues with the target curve being discussed here is variations in measurement setups. If you saw some of the same headphones measured by both Innerfi and Headroom, for example, you would often see fairly large discrepancies in the FR plots depicted. So which is right? Maybe they both are in certain ways; but we end users can't always tell, and True-Fi has to contend with the same measurement challenges as everybody else. They finally support a headphone I own (Shure SE535) so I may give the trial a go; but I have custom tips on mine, so does THAT throw the measurements off? Oh what a tangled web we weave, when we first practice to perceive...

RandomContext's picture

I'm not a fan of EQing either but this really improves my headphone experience with all I've tested it with. Their Android app right now isn't great and if they could make it systemwide it would improve things so much.
I look forward to them additing more headphones in the future.

Peter Puck's picture

I buy headphones that are pleasing to me I don't spend money on cans I don't like that need to be EQ'ed to be tolerated I'm a huge fan of straight wire and gain devices. If the signature is not to my liking unEQed then the money stays in my wallet. Flat response is the goal as close to it is the reality. If the cans colour the sound don't buy them unless your a kid that loves the fart cannon sound of Skull candy or beats .

RandomContext's picture

A headphone purchase is more than sound signature though. I think the software should make it easier to buy a pair by reducing one of the many checkboxes to get through.

I have a pair of Sennheiser on-ear headphones that were cheap but sound cheap. Muddy. Just a terrible listen. The software makes them quite enjoyable. Without the comfort issue I would continue use them.

If you have a few of the supported headphones and you want a flatter sound it's something to consider.

Stropdas's picture

Tried the bèta app for Android. While it improves the Akg k550 and Beyerdynamic DT 770 Pro 80 Ohm I wonder why it would better for me than the free Viper4android that does sort of the same but with far more headphones, options to finetune and system wide integration. I'll wait to see what the future brings from this company. I'm a big fan of EQ'ing for many years. A few very cheap earphones are also supported with a lot of improvement. Got a lot of them because they have a lot of abuse by me on the road.

GimmeCans's picture

Doing the 10-day trial and like it- it alleviates the somewhat 'dark' voicing of the Shure and seems to open up the 'headstage'. However, only one headphone I own (Shure SE535) is supported. Therefore I would be paying $79 for, essentially, one equalizer preset.

Also, I did a Google search for 'SE535 frequency response'. The graphs I found are so different from one another that if you removed the labeling, you would never suspect that they were from the same headphone. Which one is 'right'? Does Sonarworks use the 'right' one? Technically, someone with sufficient digital audio 'chops' could 'reverse engineer' the preset, but I'm not going to disrespect Sonarworks or Innerfi by posting how I think that would work here.

Bottom line, it does what it claims to do, so if your cans are supported, at least give the trial version a whirl and decide for yourself.

bennemann's picture

I was excited to try this software. What could be better than hearing exactly what the artist intended, right?

Well, I've been doing critical listening for a couple hours with my Beyerdynamic DT 770 Pro 80 Ohm, listening to several different genres, enabling and disabling True-Fi. With it on, the music loses a lot of clarity, becoming more muddy (and I don't mean this in the audiophile sense of a lack or emphasis of a certain frequency range, but in the subjective feeling like the music is now trapped in a mud swamp). It's a comparable effect to listening to a really low quality MP3 (i.e 96 kbps). Vocals sound considerably more distant and "ethereal". It also adds a sort of faint background-static-like noise that's noticeable even in simple instrumental songs. Finally, it somehow manages to add a bass "boom" even in songs that don't normally have it (like orchestral pieces).

The premise of standardising the sound of all headphones is quite interesting, but for that to be worth it, music has to sound better with it turned on, which is not the case at all for me.

To be fair, I did find one song where True-Fi sounds better to my ears (Running To The Sea by Chicane). But all the other songs sounded worse.