The Owliophile EQ tool provides four different audio generators and players, a fully equipped parametric equalizer, and further tools to help find the right EQ settings. Here's a description of the tools at the top:
The Single Tone Tool: This tool lets you play a sine wave at a frequency of your choice. The recommended way to change the frequency is to drag the line through the graph. Use it to find peaks and dips in your frequency response. These will usually show up in the upper mids and treble region (~3kHz and above). The tool also lets you play pink and white noise -- in those cases the frequency selection has no effect.
The Sweep Tool: The sweep tool will take over the job of dragging the line through the graph. You define the start and end frequency as well as the total duration of the sweep and click play. Some unevenness in the frequency response may be easier to detect in a downward sweep -- you can use the "reverse" option for that.
The Alternating Tones Tool: This tool can be used to match the volume between two tones or frequency regions. When the noise option is activated, you will hear bandpass-filtered pink noise (Q=30) at the selected frequencies instead of clean sine waves. This de-emphasizes the volume of the center frequency and is more appropriate to judge wider regions in the frequency spectrum for low-Q filter adjustments.
The Music Player: The Music Player lets you load in audio files to test your EQ settings in practice. We recommend making only low-Q adjustments based on music listening impressions.
The Marker Tool: The Marker Tool (in the Display section) lets you draw dashed vertical lines into the plot to mark and remember important frequencies such as the locations of peaks, dips, or channel imbalances. We recommend using it in combination with the Single Tone Tool and the Sweep Tool.
The Delay Compensation Tool: The Delay Compensation Tool can be used to synchronize your audio and the vertical frequency line when using Bluetooth audio devices. Bluetooth devices typically delay the audio by 100-300ms, which can noticeably de-synchronize the sweep from the moving bar, and make it difficult to mark the right frequencies with the Marker Tool during a sweep. The delay compensation is not active for the Single Tone Tool and should be left at 0 when using wired audio devices.
The parametric equalizer: The equalizer available in the EQ Tool is fully parametric, i.e., you can adjust each band's frequency, Q ("narrowness"), and gain (volume boost / decrease) for peak filters, and frequency and gain for shelf filters. This is equivalent to popular system-wide EQ tools like Equalizer APO or eqMac, and the settings can be transferred 1:1 to those tools. In fact, this is the point of the EQ Tool: Help you find the right settings for your preferred EQ, then copy them over. Note that new filters are added at the frequency selected in the Single Tone Tool, and banks can be used to keep things organized and for easy A/B testing.
First, let's be clear that we don't want to force anyone to use any specific method to equalize their headphones -- feel free to use the tools as you see fit. That said, we have developed a method that works very well for us and we recommend you give it a try. The basic idea is to even out the frequency response, first locally with narrower filters, then globally with wider ones, before adding additional filters that bring the frequency response back in line with common targets and/or your own preferences.
Note that a flat frequency response is neither "neutral" nor "correct" in headphones. If you are unsure why that is, you could look up the terms "Head-related Transfer Function (HRTF)", "Diffuse Field Curve", "Frontal Field Curve" and "Harman Target" for more information. In short, our ears, heads, and bodies aren't acoustically invisible. They change the frequency response of incoming acoustic signals on their way to our ear drums significantly, as described by the HRTF, and our brains are accustomed to this non-flat frequency response. Consequently, speakers that measure flat with a microphone in an empty room do not measure flat with microphones placed at a person's ear drums. In order to sound "neutral", "natural", or "correct", headphones should therefore roughly follow a frequency response curve close to the listener's HRTF evaluated for frontal speaker placement.
Peaks, dips, and channel imbalances: The first step in the process is to address narrow peaks and dips in the frequency response, as well as correcting channel imbalances. We recommend using filter bank 1 for this step. The frequency range where these issues occur is usually north of 3 kHz. Drag the vertical line (Single Tone Tool) through the graph to find the peaks and dips, and apply for each one an EQ filter in the opposite direction until the sweep sounds smooth/even. Often, multiple filters are required to even out complex frequency response shapes. Channel imbalances can occur in narrow or wide frequency bands and result in a shift of the perceived location of the tone to one side. Select "left" or "right" for the EQ filter's channel and pull the gain up or down until the tone sits center again. Use sweeps (both fast and slow) to confirm if a region is smooth, and to find remaining uneven areas. Use as few filters as possible but as many as necessary. This step can take anything between 5 minutes and an hour for experienced listeners, depending on the complexity of the frequency response that is to be corrected.
Flat tonal balance: Once the frequency response is locally smooth we move on to unevenness on larger scales (high bandwidth, low-Q filters); making it flat (to the ear) between about 200 Hz and 10 kHz. We recommend using a new bank (2) for this to keep a good overview. Use the Alternating Tones Tool in noise mode, and set up a number of filters in 1/3, 1/2 or full octaves with wide Qs (e.g., Q=1 for filters in full octave steps, Q=2 for half octaves, etc.). Start from a "base tone" (e.g., 500 Hz) and move in multiples from there in both directions. Adjust the filters' gain values until all frequency bands sound equally loud. Note that what we perceive as equally loud between different frequency regions changes based on the absolute volume we're listening at (search for Fletcher-Munson curves or Equal Loudness Contours (ISO 226-2003) for more). This means you will likely come to different results in this step depending on the volume level at which you perform it. In order for any "preference presets" to fully work as intended in step 3, and your preferred "step 3 settings" to be transferable between sessions, headphones, and maybe even listeners, the flattening of the tonal balance should always happen at the same volume. If you own a measurement rig, you can use it to adjust the volume to a fixed level (e.g., 80 dBA). We are still looking into other options to achieve this -- drop us a message in the contact section if you have an idea!
Shaping the frequency response: The prior two steps served to make the perceived frequency response as smooth and flat as possible across the spectrum, but a perceived-flat frequency response will not sound "correct", and we also have yet to address the bass and upper treble areas. We are thus looking for a set of filters that turn the perceived-flat frequency response into one based on your anatomy and preferences. There are multiple ways to approach this:
(1) Speaker-matching: If you have a speaker or set of speakers that measure flat (or whose tonal balance you aim to reproduce in your headphones), set up a new bank (3) with filters the same way as in step 2 (low-Q filters spaced 1/3 to full octaves apart), and perform the same correction you performed with your headphones in step 2 now with your speakers (make sure to deactivate any headphone-related filters, i.e., banks 1 and 2, before you start working with the speakers). Once you have equalized your speakers to sound completely flat to your ear using the Alternating Tones Tool, you invert the gain settings on all of these filters (bank 3) and reactivate filter banks 1 and 2. This should give you at least an approximation of your speaker's tonal balance on your headphones.
(2) Preset-matching: Instead of using speakers, you can use your headphone with, e.g., a Harman Target preset as reference for tonal balance and follow the steps outlined for speaker matching: Deactivate the filters from the first two steps (banks 1 & 2), set up the AutoEQ or oratory1990 EQ profile for your headphone (bank 3), create another bank and EQ the headphone's tonal balance to "perceived-flat" again, then deactivate the external preset (bank 3), invert the gain of all filters in the new bank (4), and reactivate banks 1 and 2. This should transfer an approximation of the Harman Target's tonal balance into your preset.
Importantly, you should see this reference-based tonal balance setup as a starting point for further refinement and experimentation. We recommend playing with bass-shelves, narrow 200Hz dips, and wide 2 kHz "Hifiman dips", ideally in a separate bank to keep things tidy and allow for easy A/B testing. It is also perfectly fine to skip the reference-based tonal balance setup and freely explore the possibilities. Experiment and season to taste!
The three default banks in the EQ tool come pre-configured to fit the three steps outlined above. Feel free to use them accordingly, or delete and start from scratch.
Note that somewhere above 10 kHz you should perceive a fall-off of in the frequency response and equalizing for equal loudness above 10 kHz is likely to result in an unnatural and harsh upper treble region.
Owliophile is a group of people (currently 2 -- Leon and Luke) that are interested in high quality audio, audio products, and predominantly headphones. We created this website to provide information and tools around audio and headphones, and potentially conduct empirical research.
Each headphone's frequency response (the relative loudness of each audible frequency measured at the ear drum) is different on each head or measurement rig. While general tendencies are shared between different listeners and measurement rigs, details like peaks and dips in the treble region vary from head to head or even ear to ear. Moreover, the difference between headphones also differs from head to head, and there is no easy way to derive good personalized EQ settings from anybody else's settings or measurements. To reach a desirable frequency response curve at your own ear drum, or to just get rid of strange, uneven, or metallic sounding treble, equalizing your headphones by ear may be one of your best options.
This depends both on the headphone you're working with, your personal preferences, and the amount of effort you're willing to put into this. In our opinion, manual EQ settings for headphones are a game changer and go clearly beyond what can be achieved with AutoEQ or oratory1990's EQ profiles. That said, this is not a criticism of those works, and we find them incredibly valuable -- they themselves represent game changing improvements for most headphones with comparatively little effort for the user. However, for those willing to put in the extra effort to tune their headphones for their own head and ears, there's quite a bit more to gain. In our experience, headphones with manual EQ can often become competitive to headphones otherwise considered several classes above, and on occasion even entry-level products can rival high-end ones.
Fair enough. However, if you can spare 30 seconds, we recommend you switch over to the EQ Tool, put on a pair of headphones you have nearby, and run a sweep from 2000 to 14000 Hz to see if you hear any peaks or dips in the frequency response or off-center movements in the tone's location. If you do, there's a good chance that you might enjoy your headphones more with manual EQ settings applied. Setting up the EQ will take longer, but it might be worth giving it a shot before you go out and buy another pair of headphones -- the change can be quite dramatic. We outline below how we use the tool, which should give you a starting point if you are inexperienced with manual EQ adjustments. The goal would be to achieve a smoother presentation of the sound without significant peaks or dips. Once the sweep sounds smooth, load up some music, and switch the EQ bypass on and off to compare. If you're convinced the smoother frequency response sounds worse than the original, then maybe this isn't for you. If, however, you do prefer the smoother frequency response: Welcome to Owliophile!
The main purpose of the EQ tool is to help you create personalized profiles to use in other apps, such as Equalizer APO, Peace, or eqMac. These apps apply EQ settings system-wide and let you enjoy the improved sound of your audio system across music, movies, games, and anything else playing from your system. The "Save EQ" button allows you to save your EQ profile as a file on your computer. For easier transition you can save it directly in a format compatible with some of the most popular EQ apps out there. If your EQ app is not listed in the save options, you can manually transfer the settings. Note that each app is different with regard to the handling of, e.g., banks/groups, names, etc., and this information may not be saved in the app-specific save files. We therefore recommend saving your settings in the Owl EQ format to be able to load it back in and tweak it at a later time.
Yes, if both apps are supported by the tool. Note that some information may get lost in translation. However, peak filters, as well as high and low shelf filters should survive the conversion.
There are several aspects and concepts to understand when it comes to equalization in general, and personalized equalization in specific.
Disclaimer: Please note that the below reflects our current understanding of the subject, resulting from reading books, presentations and other online resources, as well as our own experience with headphones, measurements, and equalization. If you find fault with any of it, we would appreciate a message via the contact section with a link or other resource to explain the mistake. This section may be updated as our understanding evolves.
Peaks and Dips: Most headphones on most heads produce a range of peaks and dips, in particular in the upper mids and treble region. This is a result of the acoustic interactions between the head, pinna, and ear canal of the listener and the headphone. These peaks and dips show up in different places and to different extents on each head and with each headphone, and are not precisely predictable from measurements or other people's settings (or measurement rigs). In many instances, these peaks and dips in the frequency response are perceived as flaws, or put the other way around, "it sounds better when they're gone". It's these peaks and dips that often make a headphone sound "unpleasant", "scratchy", "metallic", "coarse", "papery", "sibilant", "hollow" or just "strange". The Owl EQ tool is designed to help find the precise frequencies, shapes, and magnitudes of the peaks and dips in anyone's frequency response.
Tonal imbalance: The tonal balance -- think of the ratio of bass to mids to treble but on a more fine-grained scale -- of one and the same headphone can vary significantly from listener to listener. This variability of one headphone's tonal balance between different listeners is dependent on the acoustic design of the headphone and doesn't affect all headphones to the same extent -- open-back headphones with low acoustic impedance typically suffer from it less than closed back headphones with high acoustic impedance. Thus, opinions on specific open back headphones are often more in line with each other than opinions on specific closed back headphones. However, less interpersonal variation doesn't mean zero, and the individual differences in tonality can still be significant in even the most open headphones. Manual EQ adjustments can help you find the "ideal" tonal balance for any headphone as perceived by you, correcting, e.g., for the specific differences between your head and a measurement rig with that specific headphone.
Acoustic transmission and minimum phase systems: An analogue signal transmission system is mathematically fully described by its frequency response, phase response, and distortion components. Frequency response and phase response together fully describe the linear part of the acoustic transmission, while the distortion components capture the non-linear part. Distortion components include harmonic distortions, which are frequently discussed in audio, but also intermodulation distortion, amplitude compression, and others, which are considered less relevant for the reproduction of sound in headphones.
Focusing on the linear part then, let's look into the concept of "minimum phase" systems. Phase shifts between different frequency regions occur unavoidably in any analogue signal transmission system with a non-flat frequency response. In other words, any system that changes the frequency response of the signal it transmits, inevitably also causes frequency-specific delays (i.e., phase shifts). The term "excess group delay" describes frequency-specific delays (phase shifts) in the system's phase response beyond those that follow directly and unavoidably from the frequency response. In a minimum phase system, the excess group delay is zero, which means that the phase response follows directly from the frequency response and vice versa -- the phase shifts are the minimum they could be given the system's frequency response, hence "minimum phase".
Headphones, and in particular passive HiFi-headphones are generally considered to be close to minimum phase systems, i.e., they are presumed to have inaudible or negligible excess group delay and distortion components. If this presumption holds, their acoustic properties can be viewed as being (almost) fully described by their frequency response. Further, if the transmission of an acoustic signal from headphones to ear drums fulfills the minimum phase property, then any headphones producing the same frequency response at the listener's ear drum should sound identical to that listener.
Moreover, standard EQ filters as found in EQ tools like Equalizer APO, eqMac, or any DAW, are indeed also minimum phase filters, meaning they alter the frequency response and phase response of the signal in unison, maintaining the minimum phase property. Now if neither the EQ filters, nor the physical parts on the headphone itself introduce any non-minimum phase behavior, then a change in the frequency response effected by the EQ filter is acoustically indistinguishable from a change in frequency response caused by the physical tuning of a headphone (e.g., a pad change or the addition of damping material).
In reality there are caveats to this and more aspects to consider; not every headphone can evoke the exact same psychoacoustic effects -- and we will address these below -- but this doesn't change the fact that equalization can have a massive impact on the way a headphone sounds, with little to no drawbacks.
In summary, manual equalization may be the most reliable way to achieve a "smooth" and "ideally balanced" frequency response at our ear drums. If done right, the process can completely transform the way a headphone sounds.
Yes, there are limits to the notion that "frequency response is everything" and "EQ can fix anything".
Disclaimer: Please note that the below reflects our current understanding of the subject, resulting from reading books, presentations and other online resources, as well as our own experience with headphones, measurements, and equalization. If you find fault with any of it, we would appreciate a message via the contact section with a link or other resource to explain the mistake. This section may be updated as our understanding evolves.
Driver limitations: All headphone drivers have physical limits with regard to what frequency responses they can produce at what volumes. A particular driver may, for instance, simply not be able to play a 20Hz tone at 95 dBA. It may show a significant uptick in distortion as it gets closer to its excursion limit before going into nasty sounding clipping once it reaches the limit (avoid clipping your headphone drivers!). These limits mean that the sound (i.e., the frequency response) you're looking for may be out of reach with any given pair of headphones at your preferred listening levels. However, these limitations mostly affect the bass frequencies, as most modern headphones should be able to produce adequate volume above ~100 Hz without clipping.
Distortion: Headphones produce distortion artifacts, in particular harmonic distortion, i.e., additional frequency components at multiples of any given frequency component in the signal. These fall under non-linear signal transmission, and pose a limit to the notion of headphones as minimum phase systems. While distortion in headphones is often under the threshold of audibility (in which case it can be considered negligible), or will in some instances even be perceived as pleasant, strong amplification of individual frequency bands can increase distortion in those bands to audible levels, or emphasize those that are already audible. A driver's distortion behavior can therefore pose further limits to the achievability of "any desired sound" through equalization -- headphone X may be able to reach the same frequency response as headphone Y, but not with the same (perceived) distortion profile.
Multi-modal sound perception: There's more to "sound" than just the frequency response at the ear drum. In particular, lower bass frequencies can be perceived through mechanisms such as bone and fluid conduction to the cochlea, as well as skin vibration via air and/or ear pads. These additional channels of perception challenge the idea that 'frequency response is everything,' as they contribute to the overall perception of bass (and lower mids) beyond what is measured at the eardrum. While EQ certainly does impact also these channels of perception, parity in frequency response as measured at the ear drum does not guarantee total parity in the perception of a headphone.
Non-minimum phase behavior: Non-minimum phase behavior refers to deviations in a (linear) system where the system's frequency and phase response are no longer inherently linked. The deviations from "minimum phase" in the phase response are expressed as excess group delay, which thus represents frequency-dependent delays in the time domain. In headphones, this behavior arises from phenomena such as resonances (e.g., driver breakup modes or ear cup reflections), acoustic impedance mismatches, and multi-path propagation (e.g., ports or vents causing sound to arrive at different times). Psychoacoustically, this can cause transient smearing, phase anomalies that disrupt imaging and soundstage, and uneven tonal perception, particularly in the treble and bass regions where resonances and delays are most prominent. Equalization with minimum-phase filters can correct frequency response deviations but cannot resolve time-domain artifacts like excess group delay or resonances, as these fall outside the scope of minimum-phase behavior.
Ringing: As a prominent example of time-domain artifacts (i.e., non-minimum phase behavior), "ringing" describes prolonged oscillations in narrow frequency bands that can occur after transients. In most cases, audible ringing is caused by the application of strong high-Q (i.e., narrow frequency band) boosts with EQ filters. While EQ filters are typically minimum phase and do not introduce time-domain artifacts into the signal itself, significant boosts in narrow frequency bands can trigger pre-existing resonances in the physical system, such as drivers, ear cups, or pads. These resonances occur when the system struggles to control vibrations at the boosted frequencies, resulting in audible oscillations. Additionally, cup reflections or standing waves in the ear canal may further reinforce the effect. Thus, while high-Q boosts may successfully fill gaps in the frequency response and create smoother sounding treble in headphones, they can inadvertently cause ringing, in which case it is recommended to disable them. Removing individual filters may not significantly impact the perceived sound quality of a headphone but still constitutes a limitation in the frequency responses we can achieve without introducing unwanted artifacts.
Susceptibility to positional variation: Changing the position of a headphone on a listener's head will typically also change its precise frequency response. While smaller changes often go unnoticed by the listener, some headphones are more susceptible to positional variation than others and even small positional variations can have audible effects on the sound. This can limit the usefulness of high-Q filters: A narrow 15dB dip at 6000 Hz may be easy to fix with a high-Q boost at 6000 Hz, but if a slight variation in the position of the headphone moves the dip to 6100 Hz, the high-Q boost will now miss its target and cause a spike at 6000 Hz, followed by the dip at 6100 hz that is now no longer compensated for, making the situation worse than if no filter was set. It is therefore recommended to test the "robustness" of any high-Q filter and disable the ones that turn out to be problematic when shifting the headphone around.
Soundstage and pinna interaction: The soundstage, or the space that a headphone creates within which the music or other audio content is perceived to be located is undeniably linked to the ear's pinna, and there is, to the best of our knowledge, no scientific consensus of how exactly soundstage in headphones is linked to frequency response, phase response, or distortion components. It is therefore also unclear how exactly one would equalize a headphone to perfectly match another headphone's spatial presentation, if this is even possible. Matching the frequency responses of two headphones as closely as possible via EQ will certainly bring their spatial presentations closer, but does not guarantee parity.
An interesting and, in our opinion, promising approach to characterize the spatial characteristics of a given headphone with measurements is the RTIGNS.com passive soundstage test and the "pinna related transfer function" (PRTF). In short, they have two different ears for their measurement rig -- one normal and one with the pinna cut off -- which lets them isolate the effect the pinna has on the frequency response with any given headphone. Both the magnitude and the shape of the frequency response differential (PRTF) are then used to compute a passive soundstage score. In RTINGS' test, the headphones are assessed in stock configuration without any equalization applied, and the differences in the headphones' stock frequency response may be causal for some or all of the differences observed between headphones. Presumably though, the space a headphone leaves around the ear, and if and where it is touching or even squishing the pinna may contribute to the PRTF beyond what is explained by frequency response alone.
However, to the ear and our perception it shouldn't matter which way the frequency response that reaches the ear drum came about; in fact, under the assumption that the pinna doesn't create audible excess group delay or distortion components, the pinna is, for all intents and purposes, just another minimum phase component in the system, and we should be able to mimic its effects via EQ filters. It appears reasonable to assume that the pinna's imprint on the frequency response embeds localization cues into the signal which are highly personalized, as they depend entirely on the listener's precise pinna shape and structure.
So how does this impose a limit to (manual) equalization? It is possible that by going for a "neutral", "smooth" or "target-conform" sound with equalization, and in particular tool-assisted by-ear equalization with high-Q filters, we tend to inadvertently change subtle spatial cues that help to create or shape a headphone's perceived spatial properties.
While all of the above are limits to the efficacy of manual equalization, in our experience none of these prevent you from significantly improving the quality of your headphone using manual equalization.
The first step in the Owliophile research agenda is to understand how people are using the tool and what kind of data quality we can expect. For now, we would appreciate if users used the submit button after completing a profile for their headphone, submitting the profile to us for inspection.
The main development of the tool is completed, but occasional updates may come in the future. We are happy with the functionality of the tool in its current state, but we do have a list of things that we'd like to add and/or improve over time. We keep that list private as some ideas may be discarded later. If you have ideas that you'd like to see implemented, feel free to reach out!
Yes! We constantly take in feedback about the Owliophile EQ tool for ideas and future development. Is there anything that doesn't work, that you would like to see improved, or any ideas for additional features in the tool? Let us know using the contact form. Also feel free to reach out for any other inquiries.
If you want to support this project, there are a few ways to do this:
Donation: If you feel like this tool has provided actual value to you, please consider a donation: paypal.me/owliophile
Promotion: If you find that the tool has improved your experience or understanding of audio and headphones, share it with friends, colleagues and online!
Media: If you happen to do YouTube or write blogs, make a video / article about the tool or mention it in passing, and, ideally, let us know about it!