Lunatique
1000+ Head-Fier
- Joined
- Mar 7, 2008
- Posts
- 1,481
- Likes
- 384
(UPDATED on April 16th, 2020, with new links to missing test tones and YouTube videos.)
I've been wanting to write a comprehensive guide on how to do surgical EQ corrections to achieve the best sound possible, in a way that is the easiest and most reliable, without any guesswork and constantly wondering if you got it right.
There are other tutorials on how to EQ headphones, but I haven't seen one that teaches people how to do it in the most reliable and easiest way possible. The other tutorials usually tell people to use pink noise and sinewave generator, and while those are useful tools, they are not the easiest or most reliable for people who are inexperience in critical audio assessment at the professional level.
I want to state at the beginning that experienced, informed, educated professional audio and advanced audiophiles understand that the most ideal sound is within that small range of acceptable accuracy/neutrality in sonic signature. This is the golden standard used by the professional audio industry and everything you've ever listened to that's produced by professional musicians, producers, mixers, mastering engineers, etc., adhere to that golden standard. Whatever slight subjective coloration among them all fall within that acceptable range of accuracy/neutrality. Before we begin, let me try to settle once and for all why achieving the most accurate and neutral sound should be the golden standard that all music lovers, audiophiles, and audio professionals, and why intentional coloration is not a good thing for most people.
Why would you want neutral/accurate sound? Isn't that a bad thing for musical enjoyment? Doesn't neutral mean boring, cold, thin, sterile, clinical, etc.?
One of the most popular but grossly mistaken misconceptions in the hobbyist headphone/audiophile world, is that neutral/accurate sound means "clinical," "sterile, "cold," "bright," "sharp," "boring," etc. I have tried to to do my part to educate others about this for years in forums but it's such a pervasive misinformation that it gets passed around and regurgitated at frightening frequency--I just can't seem do anything to turn the tide (neither can other informed people who are also trying to combat this type of pervasive ignorance).
True neutral/accurate sound has no coloration, just like if you look out a very clean and clear glass pane window and all you see is reality, as if the glass panes don't even exist as you accidentally walk into it and bump your head. If you have a television or computer display that is high fidelity with high dynamic range, accurate and wide color gamut, optimal brightness and contrast, refined detail, high refresh rate, etc., you would never describe that display as being "clinical" or "sterile" or "cold" or "boring," right? It simply looks very natural and life-like. You would also never intentionally push the color balance of your display so that the flesh tones looks too green and with exaggerated saturation that looks neon, or the shadows are completely crushed into flat blacks with no details left, or the highlights are burned out so it's just solid patches of white and no micro details, etc. What you want is a display that is perfectly accurate/neutral, so everything appears natural and life-like. So if you wouldn't want significant deviation from accuracy in your display, why would you want it in your audio gear?
A truly neutral/accurate sound system will have authoritative and deeply extended powerful bass that is very well controlled, articulate, detailed, punchy, and full. It will have mids that are natural, smooth, detailed, and rich. Its treble will be airy, refined, and clear. There should be no hint of muddy, bloated or anemic bass, no recessed, sibilant, or nasal mids, and no hard-edged, peaky, splashy, or dull treble. Your audio system should simply disappear and what you hear is very natural life-like, as if the audio gear has disappeared.
The fact is, outside of the small minority of higher-end or well-informed/experienced audio professionals and audiophiles, the vast majority of audio hobbyists and musicians have never even heard truly neutral/accurate sound before (and this is taking into consideration the amount of acceptable variations within the threshold of what we can objectively call neutral/accurate that's measurable), and this is exactly why this misconception about neutral/accurate sound is so popular. It really takes having heard it at least once in your life to understand why it is and should be the golden standard to judge everything else by.
Some of you might want to jump in and say that there's no true neutrality/accuracy for speakers and headphones since our individual physiological differences (such as HRTF, ear canal shapes) means we all hear differently from each other, regardless if the manufacturers produce audio devices that measure flat in anechoic chambers or a testing dummy head. There's also the matter of room acoustics that will significantly alter the sound of speakers in a room. I'm going to address that right now.
I can appreciate that individual HRTF can alter what each of us hear, and a carefully tuned product that measures flat might not be flat sounding to some people due to that fact. However, there is still individual objectivity based on what you actually hear in relation to a measurement that is neutral. I can use test tones as an invaluable tool to perform surgical EQ with parametric EQ, adjusting frequency bands that are both very narrow and very wide and in-between, and I can patiently adjust until the entire range from 20Hz to 20KHz sounds very smooth when playing back a logarithmic sweep, or very even amplitude when playing sequential sinewave test tones at regular intervals. That is what you can do for yourself individually, adjusting for your own idiosyncratic physiology so that you are getting a neutral frequency response. And if you are an audio professional, you can compare your own physiological biases against a measured neutral response to understand where you differ from it, so you can make critical mixing/mastering decisions more accurately.
As for the argument that the consumers are listening on all types of devices anyway, so what's the point for mixing/mastering engineers to care about neutrality/accuracy--it's very simple. If you don't start with the most neutral/accurate point possible, you will make unwanted deviations far worse because you're already starting with something that's inaccurate and the consumer audio devices will only exacerbate it. If you start with neutral/accurate, then at least the starting point is ideal, and whatever coloration the individuals add to it with their audio devices will be starting with a source that's accurate/neutral, so the deviation won't be as severe.
If I can't convince some of you that neutrality/accuracy should be the golden standard we judge audio by, then that's fine--you can believe whatever you want and enjoy audio however you want. I'm writing this for people who do care and want to learn how to achieve neutral/accurate sound, so let's just live and let live and not get into a debate about this. If you really want to debate this, I already have a long thread about this I posted years ago with many pages of raging debates on this topic--you can simply add your 2 cents to that thread, which is located here:
http://www.head-fi.org/t/564465/misconception-of-neutral-accurate
I'll say this though--if you suspect that you probably have never heard truly neutral/accurate audio reproduction before at the level of a professional mastering sound system in a professional level audio production studio with proper studio design schematics and acoustic treatment (or the headphone equivalent of it--which is extremely rare and likely does not exist as of yet--especially if you don't know how to do proper surgical EQ corrections), then you should at least accept the fact that maybe you have no authoritative opinion on this matter and you need to experience it at least once in your life before you start arguing against neutral/accurate audio reproduction. For many people, finally hearing what it sounds like is a profound revelation that forever changes how they feel and think about audio and music reproduction.
Also, I want to make it clear that some levels of intentional coloration (such as using tube amps to achieve a warmer sound) still falls within the general acceptable range of neutral/accurate sound, because the deviation is much more broad and subtle and not in isolated frequency bands that causes really uneven frequency response with erratic dips and spikes, or silly bombastic bloated bass that completely overwhelms other frequency ranges and sounds uncontrolled and muddy and lacking in articulation and detail. So those of you who like some subtle level of intentional coloration--you are safely within the general range of acceptable neutrality/accuracy.
Okay, I now know I want neutral/accurate sound. Show me how to achieve it in the easiest and most reliable way possible.
Now that I've established that for some people, truly accurate/neutral sound means aural bliss and is the one and only standard we should have/need, let me show you how to actually achieve it in the easiest, most reliable way possible.
I mentioned that using pink noise and sinewave generators are not the easiest or most reliable way, and here's the reason why: It's actually harder to judge frequency balance when the entire audible frequency range is playing back at the same time (pink noise), because unless you are an experienced audio professional (or equivalent of one), it will be very difficult for you to assess just which frequency bands are actually out of balance, and by exactly how much. We have a natural imbalance in our human hearing known as the Fletcher-Munson curve, where extreme high and and low frequencies sound quieter to us when playing at quiet levels, despite being at the same amplitude as other frequencies, and that imbalance changes as the amplitude gets louder and louder. We also get confused and overwhelmed easily when too much sensory information if presented to us all at once.
Sinewave generators are also not the easiet to use to do critical audio assessment, because you need to input the number for the frequency and then generate one frequency at a time, and there's no way to instantly switch between various different frequency bands in split seconds to judge the differences in their amplitude.
So while pink noise and sinewave generators are helpful tools, they shouldn't be your only tools. Instead, what you want to use as a more reliable and easier way to assess frequency response is to use log (logarithmic) sweep and pre-rendered sinewave test tones at regular intervals from 20Hz to 20KHz.
Here are the tools you'll need (all free or have trial versions, or very cheap):
A music player that has a good parametric EQ (don't use graphic EQs--they are not suitable for surgical EQ corrections). My recommendations are:
J River Media Center (desktop/laptop computers) - This media librarian/player has the best implementation of DSP plugins management I've tried (and I tried many), and is vastly more ergonomic and intuitive than something like foobar2000. It is also one of the best in terms of sound quality, and very speedy in response even with extremely large music libraries. You can also easily host VST plugins in Media Center without having to jump through hoops.
In Media Center, I use a freeware parametric EQ called EasyQ, from RS-MET:
http://www.rs-met.com/freebies.html
It's extremely easy to use, has no additional coloration (some EQs apply intentional coloration such as analog emulations), and I get great results from it. If you are concerned about phase distortions, then use a linear phase EQ (just search for one online). It's really not necessary most of the time, unless you are actually hearing weird distortions (this has never happened to me in all the years I've been doing this, and when there were distortions, it had to do with gain staging, not phase issues). There are lots of free and commercial parametric EQs out there, and you'll find them easily with a quick search, particularly with KVR Audio's plugin database:
http://www.kvraudio.com/plugins/newest
To install a VST plugin in Media Center, just open up the DSP Studio in Media Center and click on "Manage Plug-ins," then "Add JRiver, VST, or Winamp plugin," then navigate to the dll file and open it.
Neutron (Android) - Neutron is the best audio player on Android, period. It is the only one that has extensive professional features such as advanced processing algorithms, a useful parametric EQ, a proper crossfeed (some other player apps claim they have crossfeed but they have no idea what a proper crossfeed is and theirs sound horrible and will harm the sound quality far too much), ReplayGain, etc.
Equalizer (iOS) - I switched to Android around 2013, so I haven't kept up with iOS apps, but this was the parametic EQ/audio player I used. There are probably other alternatives that are just as good or better by now, but as long as it's got a good parametric EQ, it should do the job.
When you are doing comparisons of how the headphone sounds with and without the EQ, you can use the bypass button that most EQs have, or there might be a activate/deactive button or checkbox in your player of choice.
Audio Frequency Sprectrum Analyzer - This is a very important tool for assessing audio. It shows you visually in real-time the frequency of the audio content being played. You need this to identify frequencies during a log sweep or in musical materials where the prominent audible energies are.
J River Media Center already has a built-in analyzer, so you don't need to go find one. If you need an analyzer, just google for "Audio Frequency Sprectrum Analyzer " and you'll find many. If you want to search only for freeware, you can use KVR Audio's plugin search database to specify you want freeware only.
Various test tones - I've uploaded all the test tones you really need to assess frequency response balance. It includes a log (logarithmic) sweep, pre-rendered sinewave test tones at regular intervals from 20Hz to 20KHz, and pink noise. You can download them here:
https://1drv.ms/u/s!AuJvJJ14eSzGge14GMaT9CPbob8z-w?e=mqLIA0
You can also search YouTube for "sine wave test tones" and "sine wave sweep."
Choose a sweep that doesn't take too long. Keep it around 30 seconds for the entire sweep. Here's an example of a good one:
Here's one with sine wave tones at regular intervals. You can just drag the playback seek bar to jump around to various frequencies (or just keyboard shortcuts J and L for big jumps, or left and right arrow keys for smaller jumps):
A quick side-note about speaker/room correction: For those of you who are interested in correcting your speakers and the room modes caused by acoustic issues in your listening area, I highly recommend IK Multimedia's ARC System 2:
http://www.ikmultimedia.com/products/arc/
I use it and swear by it. I get perfectly neutral/accurate frequency response at my listening position. My Klein+Hummel O 300Ds and Neumann KH805 sounds amazing with the ARC System. There are other similar products on the market (both hardware and software), and you can find them by searching for speaker/room correction software/hardware. ARC System is the best IMO because of how much it improved from the first version (the first version wasn't nearly as good, which is why some people used to poopoo it), how easy it is to use, and the result speaks for itself.
Here's how to use the tools:
Audio Frequency Sprectrum Analyzer - You need this to see exactly what frequency is playing during the log sweep. You need to put this first in your DSP signal chain, before the EQ, otherwise the EQ (or any othe DSP plugin) will skew its reading. So for example if you are using J River Media Center, you want to open up the DSP Studio and CTRL+Drag the Analyzer plugin all the way to the top of the plugins list, as far as it'll be allowed to go. Make sure the EQ plugin is lower than the Analyzer plugin in the signal chain.
Some analyzer plugins will by default use an algorithm that's not ideal for easy visual assessment, since it does not take into consideration human hearing sensitivity shift in high frequency range. This means if you look at the analyer during a log sweep, it'll seem like the amplitude gets lower and lower the higher the frequency gets. But don't worry about it, because we're only using the analyer to identify the frequency, not its amplitude. If it bothers you, you can switch to a different algorithm that compensates for it and will appear flat across the frequency range when playing back a log sweep.
This is how it should look like when playing a log sweep:
Log sweep - When you play a logarithmic sweep, what you'll hear is a sinewave tone that sweeps the entire audible frequency range from 20Hz to 20KHz. This is the fastest and most reliable way to hear any imbalances in your headphone, with the exception of the lowest sub-bass or highest treble frequency ranges (and we'll use pre-rendered sinewave test tones for that). When you listen to the sweep, if there are dips and spikes, you will hear them very clearly. A perfectly neutral/accurate headphone will sound very smooth during the sweep, without any significant dips or peaks. Most headphones will have significant peaks and dips somewhere in the upper-mids / sibilance range (3KHz to 8KHz) and the lower treble range (9KHz to 12KHz). You might also hear bloated mid to upper bass around the 125Hz to 250Hz range. Some headphones also have a dip right before the upper-mids in the range of 2KHz and 3KHz.
Note that it's natural to hear a gradual drop in amplitude when going from 6KHz and up. This is because highest frequencies will have far less energy to our human hearing. It's also natural to hear that from 50Hz to 40Hz, the bass energy feels stronger than the other bass frequencies even when they are in fact not louder. It'll sound like your skull is vibrating, and this is normal. In fact, if you don't hear them with slightly strong energy, that means there's a roll-off/dip. From 20Hz to 30Hz, it'll seem like the energy gets lower, and this is normal too. At such low frequency, they are more felt than heard. Also, many headphones cannot handle frequencies that low and are completely rolled off at such low frequencies. Just because the manufacturer lists 20Hz-20KHz in the specifications does not mean much because without specifying how much deviation, that headphone could very well be playing 30Hz at -12 dB, which is essentially useless. This is why the standard for specifying acceptable deviation in the professional audio industry is to use ± 3 dB (plus or minus 3 dB of threshold in acceptable accuracy). Consumer audio industry is far less strict with standards, which is why there is so much bull$hit and snake oil and outright lies and false advertising in consumer audio industry.
When you have identified the dips and peaks, write them down. Don't worry about how many dBs of difference to write down--just write down at what exact frequencies you're hearing audible differences.
Pre-rendered sinewave test tones - These are much better than sinewave generators because you can instantly switch from one exact frequency to another without having to waste time entering any numbers, and can jump between any frequencies instantly and repeatedly.
When you playback the sinewave tones, make sure you do not have any kind of normalization /smart volume/RepayGain turned on. This means any settings that will automatically make different songs playback at the same volume. Although the sineware test tones should be already normalized to the same volume, it's good to take this extra precaution just in case.
What I usually do when using pre-rendered sinewaves, is to first assess the bass region. start at 200Hz and sequentially play each interval all the way down to 20Hz. If the low sub-bass is rolled off or bloated, you will hear it.
Next, I would start from about 80Hz and start going up sequentially all the way to about 1KHz. If there's bloated upper to mid-bass I'll hear it. Headphones that sound overtly warm tend to have a thickness in the 200Hz range, and you should hear it.
Then from 1KHz, you go up sequentially all the way to 20KHz. Listen for noticeable dips and spikes. Some headphones will dip around 2~3KHz, and many tend to have significant dips or peaks between 4KHz and 8KHz. For IEMs, the resonance peak at around 7KHz is a big problem for some people due to the ear canal shape, and can spike as much as 12 dB or more.
Many headphone manufacturers tune their headphones to have a peak at around 10KHz to 12KHz (either narrow or wide bandwidth), to create the illusion of detail. This is not a good thing because it's not natural detail--it sounds artificial. Frequencies above 10KHz tend to sound rolled off, and the headphone itself will often in fact be rolled off in the highest frequencies. For those of you who are past teenage years, the older you are, the more you will lose hearing in the highest frequencies. Most older adults might only be able to barely hear 16KHz.
Understanding headphone measurements and ideal target response before we do corrections:
Now that you've listend and noted all the areas of imbalance in your headphone, we'll start correcting those problems.
Before we start with the EQing, it's important to learn about the Harman Target Response Curve. Do not skip this step--it is crucial to your understanding of how headphones are supposed to sound different from speakers, and why a neutral pair of headphone shouldn't measure totally flat. Read this two-part explanation of headphone frequency response measurements before moving on:
http://www.innerfidelity.com/conten...equency-response-part-one#PxVkSf8KrXBSEM6I.97
http://www.innerfidelity.com/conten...equency-response-part-two#W6HDd66v1yfLMk0R.97
You might also want look at the photo I posted of the Audez'e frequency response graph and superimposed correction in one of my old threads about EQing heasphones here (and read what I wrote about it just below the photo):
http://www.head-fi.org/t/551426/my-eq-curves-for-lcd-2-hd650-m50-and-007mk2
See that gently sloping white line of -10 dB I drew from 1KHz to 20KHz? That is what Tyll (of InnerFidelity) used to recommend as the ideal frequency response curve. As you probably already noticed, the Harman Target Reponse Curve is similar, but with a few differences. I currently prefer the Harman Target Response Curve (or Tyll's slightly modified version of it) since it's more updated and closer to what a full-range speaker system sounds like.
Now, lets start correcting the imbalance:
You start with any frequency ranges--it doesn't really matter. The key is to constantly assess your correction with the pre-rendered sinewave test tones and the log sweep (and you don't have to playback the entire log sweep when only listening for a specific requency range. For example, if you are working on the 3 to 8KHz range, you can jump ahead to around halfway point to start the sweep at 1KHz).
I usually start with the narrowest bandwidth (using peak/dip mode, NOT low/high pass/shelf mode) when applying EQ at a specific frequency. I will adjust the gain and then use the pre-rendered sinewave test tones in that range to listen for balance, until I get perfect balance. Once I have achieved that, I'll use the log sweep and listen of any dips or peaks right before and after the exact frequency I'm correcting. If the sweep is not perfectly smooth, I'll adjust the bandwidth until it smooths out, and if necessary continue to adjust the gain too. This takes patience because you need to get the bandwidth just right so the correction is countering the imbalance precisely. You do this for every single frequency imbalane that is audible.
Sometimes, a headphone might have the entire treble or bass region recessed or bloated, and you can switch to high or low pass/shelf mode instead of peak mode. Adjust so that visually, you can see the correction applied to the entire affected region. Then within the shelved area, if there are frequencies that needs to be corrected, you can do that with more dip/peak bands.
Keep in mind that sometimes you need to have EQ bands very close to each other (for example, one at 6KHz and one at 7KHz). when corrections happen in frequencies that are in close proximity to each other, they will also counter or amplify each other, so this is why you cannot do corrections by numbers alone and must visually look at the curve itself to make sure you are getting the exact correction you need.
Ideally, once you have finished all corrections, the log sweep will sound far smoother and balanced compared to no correction. And when you play the pre-rendered sinewave test tones sequentially through the entire frequency range, you'll hear no significant imbalances in amplitude (other than the natural Fletcher-Munson curve).
EDIT: It appears some people were confused by exactly what my method involves, especially regarding whether the Fletcher-Munson equal loudness curve and Harman Target Response Curve are involved in the EQ process. I'll summarize my process here again:
And here's an article/video showing Bob Katz and Tyll of InnerFidelity talking about EQing headphones to match the Harman Target Response Curve as a great starting point: http://www.innerfidelity.com/conten...es-harman-target-response#lgR1tm6s3SeYdDEO.97
Testing with musical material:
While it's important to also use musical tracks to test the results of your EQ correction, I have noticed that many people don't know how to select musical material that makes it very easy to hear problems in their audio gear.
The trick with easy audio assessment isn't simply to use music that's mastered perfectly--it's often the opposite (though perfect mastering is useful too, but in a different way). Music that is mastered too bright is great for revealing problems in the sibilance range, such as assessing headphones that are too bright and fatiguing. These masterings that are the bright side should still be within tolerable threshold when playing back on audio gear that is neutral/accurate, but if the gear is too bright, these bright masterings will become unbearable, and that's a very easy way to test/correct overly bright headphones.
You also want music that contains ample low frequency information, so that you can easily hear when the headphone's sub is rolled off and sounds too anemic, or when it's excessively bloated and muddy, or if it lacks proper amount of punch, articulation, and speed.
For the mids, you want music that has enough focus on the mids, such as vocals and instruments that are in similar frequency range as vocals (such as string instruments, woodwind instruments, guitars, etc.) If the headphone lacks presence or bite in the mids, you will hear it, and if it's too aggressive and nasal, you'll hear it too.
And of course, music that has a great balance in its arrangement and mastering is good for assessing overall balance, imaging, depth, etc.
Over the years, I've put together a playlist that I use specifically for testing audio gear, and I'm now going to share them with you and show you how to use them. If you really like any of the tracks and want more information to hunt down the composer/artist so you can purchase their music, don't hesitate to ask to find out more. Here's the link to download the tracks:
edited:sorry no link to copyrighted musics(signed evil modo)
(With the link to the MP3s removed, I've tried to find the songs on Youtube to link to instead. I couldn't find all of them, but found enough of them to be very helpful.)
I'll group these tracks into categories based on how I use them. (Don't think about the actual genre/language of these tracks because you're using these tracks for testing audio gear, not for leisurely listening. If you happen to enjoy the music, that's just icing on the cake.)
And don't get hung up on the fact that these are MP3/YouTube audio. They are perfectly fine for the kind of tests you will be doing with them, and using uncompressed versions will not change anything. You'd have to focus intensely and strain really hard to hear any differences compared to the original CD quality (and you often won't hear any differences reliably if the YouTube video is HD) and at that point the differences are too subtle to be relevant in this specific context.
When you do comparisons of musical material, it's important to not switch between the EQ curve and the bypassed/unEQd sound very quickly without allowing your brain to first adapt to the sonic signature. That is not the correct way to do audio assessment. What you need to do is to listen for at least 30 seconds so your brain can adjust itself to that sonic signature. When you do switch, stay with it for at least 30 seconds before you switch again.
Testing for bass:
04 - Love For Sale - The upright bass in this track will sound too muddy and bloated with headphones that have too much bass. The bassline will overwhelm the rest of the frequencies too much and you won't hear the articulation of each bass note clearly. On neutral sounding gear, the bass notes will sound balanced with the rest of the arrangement.
12.sub-ID%20Mi%20Woof - This track can test just how deep and powerful the bass frequency response is. If your headphone's bass it too rolled-off, you will not hear/feel that really powerful and deeply resonant monster Roland 808 drum machine kick drum. It is one of the most famous kick drum sounds in the world (at least in hip-hop and electronic music), and you definitely want to be able to play it with enough authority and power. If the gear is too exaggerated in the bass region, the 808 kick drum will sound ridiculously monstrous and way out of balance. Also, if your headphones cannot handle the powerful bass of the 808 kick drum, it might distort.
02 - klendathu drop - At 0:49, you should hear a powerful sustained bass note centered at around 45Hz. If it doesn't sound powerful and majestic, then your headphone's sub-bass is too rolled off. If it sounds too overwhelming, then your headphone's sub-bass is too exaggerated.
lucite-dry - This short clip has bass kick that should sound punchy at with deep and powerful resonance (though not as low in frequency as the track with the 808 kick drum). If the bass kick sounds limp and weak, then your headphone's bass response is too anemic. If the bass kick sounds too overwhelming, then your headphone's bass is too exaggerated.
03 - Mahalle - Starting at 0:16, you should hear that deep sub-bass thud that accompanies the strumming instrument each time it repeats. It should be very obvious and accents the start of each repeated strumming pattern. If it's not strong enough, the track will sound too thin.
02 - L.F.O. - At 1:38, you will hear just the sub-bass notes by itself. Like the 808 kick drum, if your headphones cannot handle the sub-bass energy, it could distort. On neutral gear those bass notes should sound clean and deep.
Halo Theme - The intro section of the Halo theme is great for testing sub-bas because it has prominent low sub-bass energy going all the way down to 20Hz. If your speakers or headphones cannot handle that much low sub-bass frequency, it will likely distort, or sound too anemic/wimply instead of majestic and epic.
11 - Dogfighter - Although this track is a hybrid of jazz fusion and orchestral, the bass kick should have a solid and beefy thud that's fairly strong, but not overwhelming.
02 - 戦いの達命 (Battle of our lives) - The big bass drum hits between 0:17-0:18 and 0:19-0:20 needs to sound strong and deep and really punctuate the orchestration with a sense power.
bigbeat01 - Here's another deep sub-bass track. When the bassline starts at 0:24, it should be very full and deep and powerful, as if your entire head is rumbling. If you are not hearing that, then your headphone's sub-bass is not deep/powerful enough. If it's way too overwhelming and sounds like a muddy mess, then the sub-bass of your headphone is too exaggerated.
track 06 - cats on mars (dmx krew remix) - A fun track for sub-bass. The bass kick should should sound full and deep and with ample body. If it sounds weak then the sub-bass is too rolled off. If it's too overwhelming, the sub-bass is too exaggerated.
Testing for the mids:
02 - Rivers Of Love - This track's vocals need to sound clear, warm, and rich. At 0:54 when she sings "through a scene," the word "through" will sound too wooly if the headphone has problems with the mids such as being too thick sounding. Also, there are some spots in the vocals that can have a bit of sibilance if the headphone's too bright.
156608_rp350_11 - This electric guitar track needs to have enough bite in the upper-mids in order to retain its sense of raw power. I distinctly remember when I first got the Aucez'e LCD-2 and tested it, I was disappointed by how recessed it's upper-mids were and how wimpy those guitars sounded.
06 Rain (I Want A Divorce) - This string arrangement needs to sound rich and full, and not thin and brittle. The bass notes that punctuates throughout the track needs to be deep and rich and really fill out the bass frequencies and provide an authoritative anchor for the arrangement.
02 - 戦いの達命 (Battle of our lives) - This track was already mentioned for the big bass drum hits, but the brass-heavy orchestration is also very good for testing the mids. Brass intruments, like distorted guitars, needs to have enough bite and presence to convey that sense of raw power.
[BARBEE BOYS]BARBEE BOYS 2(07)‚Í‚¿‚ ‚킹‚̃ƒbƒJ - A J-Rock song from the 80's that can sound too aggressively bright if the headphone has too much emphasis in the upper-mids.
Testing for upper-mids/brightness problems:
15 - Fade - The bright sounding percussive chords that starts at 0:38 will sound fine on neutral sounding audio gear. It's bright but not grating. On gear that's too bright, they can sound piercing and painful.
09 - 遠雷 (Distant Thunder) / 06- MORNING CALL - Both of these tracks have intros that have bright shimmering sounds that repeat in patterns, and on headphones that are too bright, they can sound too sharp and fatiguing.
05 - 少年は天使を殺す (The Boy Killed the Angel) - This track is mastered quite bright and on neutral audio gear, the backing vocal chant that comes right after the opening drum beat will have sibilance that's very prominent, but it's still within tolerance and should not sound painful. But on audio gear that is too bright, it will become very annoying and even painful. This continues all the way through the song with both the main vocals and the backing vocals. Also, that snare drum can sound fatiguing on audio gear that's too bright.
02 - Here's Where The Story Ends - This track isn't overtly bright, but when the chorus starts at 1:10, the sibilance can become too hot and annoying on headphones that are too bright.
01 - Already Yours - This wall-of-distorted guitars track can get fatiguing fast if the headphone is too bright. On a neutral sounding audio gear, it's tolerable to listen all the way through, but any skew towards excessive brightness it'll become unbearable.
03 PARTY - This track's chorus starting at around 0:48 has some sibilance issues, and on neutral audio gear, it'll sound a bit hot in the sibilance but still tolerable. On gear that's too bright, those distinct sibilances will become painful.
05 - 淋しいから言えないから (Because I Was Lonely, I couldn't Say) - Morikawa Miho's voice during the chorus starting at 1:07 can sound overly sharp on audio gear that's too bright, but if the gear is neutral, she'll sound bright but not painful or grating.
Testing for treble:
09 第七感（セッティエーム サンス) (Seventh Sense) - This track has really tizzy and splashy treble. It's still within tolerance on neutral audio gear, but on gear that's too bright in the treble, it'll become really tizzy and annoying, and maybe even painful.
10 William, It Was Really Nothing - This track also has a lot of treble energy, but it's not quite as drastic as Seventh Sense. On headphones with exaggerated treble, it will sound like a splashy mess instead of shimmering.
For overall balance, clarity, and richness in complex arrangements:
Mojo_Madness / ST_Theme_v1 / 03 Come in 007, Your time is up / 07 BROTHER - These action/adventure/drama oriented score cues have rich orchestrations and I like to use them to listen for overall richness, balance, clarity, dynamic range, imaging, depth, etc.
The forum only allows 30 media links per post, so I'm moving the ones past #30 to another post. Here's the link to that post:
https://www.head-fi.org/threads/the...nd-for-non-professionals.796791/post-15562131
Wrapping up:
Okay, I hope this has been helpful to those of you who always wanted to try EQing your headphones but don't know how, or felt confused/lost when using pink noise and sinewave generators. My method of using logarithmic sweep tone and pre-rendered sinewave test tones is much more straightforward and intuitive, and anyone should be able to do very precise EQ curves using my method and achieve the most neutral/accurate frequency response for any headphone (unless that headphone's frequency response and physical drivers are so bad that it just can't be corrected properly).
I've been wanting to write a comprehensive guide on how to do surgical EQ corrections to achieve the best sound possible, in a way that is the easiest and most reliable, without any guesswork and constantly wondering if you got it right.
There are other tutorials on how to EQ headphones, but I haven't seen one that teaches people how to do it in the most reliable and easiest way possible. The other tutorials usually tell people to use pink noise and sinewave generator, and while those are useful tools, they are not the easiest or most reliable for people who are inexperience in critical audio assessment at the professional level.
I want to state at the beginning that experienced, informed, educated professional audio and advanced audiophiles understand that the most ideal sound is within that small range of acceptable accuracy/neutrality in sonic signature. This is the golden standard used by the professional audio industry and everything you've ever listened to that's produced by professional musicians, producers, mixers, mastering engineers, etc., adhere to that golden standard. Whatever slight subjective coloration among them all fall within that acceptable range of accuracy/neutrality. Before we begin, let me try to settle once and for all why achieving the most accurate and neutral sound should be the golden standard that all music lovers, audiophiles, and audio professionals, and why intentional coloration is not a good thing for most people.
Why would you want neutral/accurate sound? Isn't that a bad thing for musical enjoyment? Doesn't neutral mean boring, cold, thin, sterile, clinical, etc.?
One of the most popular but grossly mistaken misconceptions in the hobbyist headphone/audiophile world, is that neutral/accurate sound means "clinical," "sterile, "cold," "bright," "sharp," "boring," etc. I have tried to to do my part to educate others about this for years in forums but it's such a pervasive misinformation that it gets passed around and regurgitated at frightening frequency--I just can't seem do anything to turn the tide (neither can other informed people who are also trying to combat this type of pervasive ignorance).
True neutral/accurate sound has no coloration, just like if you look out a very clean and clear glass pane window and all you see is reality, as if the glass panes don't even exist as you accidentally walk into it and bump your head. If you have a television or computer display that is high fidelity with high dynamic range, accurate and wide color gamut, optimal brightness and contrast, refined detail, high refresh rate, etc., you would never describe that display as being "clinical" or "sterile" or "cold" or "boring," right? It simply looks very natural and life-like. You would also never intentionally push the color balance of your display so that the flesh tones looks too green and with exaggerated saturation that looks neon, or the shadows are completely crushed into flat blacks with no details left, or the highlights are burned out so it's just solid patches of white and no micro details, etc. What you want is a display that is perfectly accurate/neutral, so everything appears natural and life-like. So if you wouldn't want significant deviation from accuracy in your display, why would you want it in your audio gear?
A truly neutral/accurate sound system will have authoritative and deeply extended powerful bass that is very well controlled, articulate, detailed, punchy, and full. It will have mids that are natural, smooth, detailed, and rich. Its treble will be airy, refined, and clear. There should be no hint of muddy, bloated or anemic bass, no recessed, sibilant, or nasal mids, and no hard-edged, peaky, splashy, or dull treble. Your audio system should simply disappear and what you hear is very natural life-like, as if the audio gear has disappeared.
The fact is, outside of the small minority of higher-end or well-informed/experienced audio professionals and audiophiles, the vast majority of audio hobbyists and musicians have never even heard truly neutral/accurate sound before (and this is taking into consideration the amount of acceptable variations within the threshold of what we can objectively call neutral/accurate that's measurable), and this is exactly why this misconception about neutral/accurate sound is so popular. It really takes having heard it at least once in your life to understand why it is and should be the golden standard to judge everything else by.
Some of you might want to jump in and say that there's no true neutrality/accuracy for speakers and headphones since our individual physiological differences (such as HRTF, ear canal shapes) means we all hear differently from each other, regardless if the manufacturers produce audio devices that measure flat in anechoic chambers or a testing dummy head. There's also the matter of room acoustics that will significantly alter the sound of speakers in a room. I'm going to address that right now.
I can appreciate that individual HRTF can alter what each of us hear, and a carefully tuned product that measures flat might not be flat sounding to some people due to that fact. However, there is still individual objectivity based on what you actually hear in relation to a measurement that is neutral. I can use test tones as an invaluable tool to perform surgical EQ with parametric EQ, adjusting frequency bands that are both very narrow and very wide and in-between, and I can patiently adjust until the entire range from 20Hz to 20KHz sounds very smooth when playing back a logarithmic sweep, or very even amplitude when playing sequential sinewave test tones at regular intervals. That is what you can do for yourself individually, adjusting for your own idiosyncratic physiology so that you are getting a neutral frequency response. And if you are an audio professional, you can compare your own physiological biases against a measured neutral response to understand where you differ from it, so you can make critical mixing/mastering decisions more accurately.
As for the argument that the consumers are listening on all types of devices anyway, so what's the point for mixing/mastering engineers to care about neutrality/accuracy--it's very simple. If you don't start with the most neutral/accurate point possible, you will make unwanted deviations far worse because you're already starting with something that's inaccurate and the consumer audio devices will only exacerbate it. If you start with neutral/accurate, then at least the starting point is ideal, and whatever coloration the individuals add to it with their audio devices will be starting with a source that's accurate/neutral, so the deviation won't be as severe.
If I can't convince some of you that neutrality/accuracy should be the golden standard we judge audio by, then that's fine--you can believe whatever you want and enjoy audio however you want. I'm writing this for people who do care and want to learn how to achieve neutral/accurate sound, so let's just live and let live and not get into a debate about this. If you really want to debate this, I already have a long thread about this I posted years ago with many pages of raging debates on this topic--you can simply add your 2 cents to that thread, which is located here:
http://www.head-fi.org/t/564465/misconception-of-neutral-accurate
I'll say this though--if you suspect that you probably have never heard truly neutral/accurate audio reproduction before at the level of a professional mastering sound system in a professional level audio production studio with proper studio design schematics and acoustic treatment (or the headphone equivalent of it--which is extremely rare and likely does not exist as of yet--especially if you don't know how to do proper surgical EQ corrections), then you should at least accept the fact that maybe you have no authoritative opinion on this matter and you need to experience it at least once in your life before you start arguing against neutral/accurate audio reproduction. For many people, finally hearing what it sounds like is a profound revelation that forever changes how they feel and think about audio and music reproduction.
Also, I want to make it clear that some levels of intentional coloration (such as using tube amps to achieve a warmer sound) still falls within the general acceptable range of neutral/accurate sound, because the deviation is much more broad and subtle and not in isolated frequency bands that causes really uneven frequency response with erratic dips and spikes, or silly bombastic bloated bass that completely overwhelms other frequency ranges and sounds uncontrolled and muddy and lacking in articulation and detail. So those of you who like some subtle level of intentional coloration--you are safely within the general range of acceptable neutrality/accuracy.
Okay, I now know I want neutral/accurate sound. Show me how to achieve it in the easiest and most reliable way possible.
Now that I've established that for some people, truly accurate/neutral sound means aural bliss and is the one and only standard we should have/need, let me show you how to actually achieve it in the easiest, most reliable way possible.
I mentioned that using pink noise and sinewave generators are not the easiest or most reliable way, and here's the reason why: It's actually harder to judge frequency balance when the entire audible frequency range is playing back at the same time (pink noise), because unless you are an experienced audio professional (or equivalent of one), it will be very difficult for you to assess just which frequency bands are actually out of balance, and by exactly how much. We have a natural imbalance in our human hearing known as the Fletcher-Munson curve, where extreme high and and low frequencies sound quieter to us when playing at quiet levels, despite being at the same amplitude as other frequencies, and that imbalance changes as the amplitude gets louder and louder. We also get confused and overwhelmed easily when too much sensory information if presented to us all at once.
Sinewave generators are also not the easiet to use to do critical audio assessment, because you need to input the number for the frequency and then generate one frequency at a time, and there's no way to instantly switch between various different frequency bands in split seconds to judge the differences in their amplitude.
So while pink noise and sinewave generators are helpful tools, they shouldn't be your only tools. Instead, what you want to use as a more reliable and easier way to assess frequency response is to use log (logarithmic) sweep and pre-rendered sinewave test tones at regular intervals from 20Hz to 20KHz.
Here are the tools you'll need (all free or have trial versions, or very cheap):
A music player that has a good parametric EQ (don't use graphic EQs--they are not suitable for surgical EQ corrections). My recommendations are:
J River Media Center (desktop/laptop computers) - This media librarian/player has the best implementation of DSP plugins management I've tried (and I tried many), and is vastly more ergonomic and intuitive than something like foobar2000. It is also one of the best in terms of sound quality, and very speedy in response even with extremely large music libraries. You can also easily host VST plugins in Media Center without having to jump through hoops.
In Media Center, I use a freeware parametric EQ called EasyQ, from RS-MET:
http://www.rs-met.com/freebies.html
It's extremely easy to use, has no additional coloration (some EQs apply intentional coloration such as analog emulations), and I get great results from it. If you are concerned about phase distortions, then use a linear phase EQ (just search for one online). It's really not necessary most of the time, unless you are actually hearing weird distortions (this has never happened to me in all the years I've been doing this, and when there were distortions, it had to do with gain staging, not phase issues). There are lots of free and commercial parametric EQs out there, and you'll find them easily with a quick search, particularly with KVR Audio's plugin database:
http://www.kvraudio.com/plugins/newest
To install a VST plugin in Media Center, just open up the DSP Studio in Media Center and click on "Manage Plug-ins," then "Add JRiver, VST, or Winamp plugin," then navigate to the dll file and open it.
Neutron (Android) - Neutron is the best audio player on Android, period. It is the only one that has extensive professional features such as advanced processing algorithms, a useful parametric EQ, a proper crossfeed (some other player apps claim they have crossfeed but they have no idea what a proper crossfeed is and theirs sound horrible and will harm the sound quality far too much), ReplayGain, etc.
Equalizer (iOS) - I switched to Android around 2013, so I haven't kept up with iOS apps, but this was the parametic EQ/audio player I used. There are probably other alternatives that are just as good or better by now, but as long as it's got a good parametric EQ, it should do the job.
When you are doing comparisons of how the headphone sounds with and without the EQ, you can use the bypass button that most EQs have, or there might be a activate/deactive button or checkbox in your player of choice.
Audio Frequency Sprectrum Analyzer - This is a very important tool for assessing audio. It shows you visually in real-time the frequency of the audio content being played. You need this to identify frequencies during a log sweep or in musical materials where the prominent audible energies are.
J River Media Center already has a built-in analyzer, so you don't need to go find one. If you need an analyzer, just google for "Audio Frequency Sprectrum Analyzer " and you'll find many. If you want to search only for freeware, you can use KVR Audio's plugin search database to specify you want freeware only.
Various test tones - I've uploaded all the test tones you really need to assess frequency response balance. It includes a log (logarithmic) sweep, pre-rendered sinewave test tones at regular intervals from 20Hz to 20KHz, and pink noise. You can download them here:
https://1drv.ms/u/s!AuJvJJ14eSzGge14GMaT9CPbob8z-w?e=mqLIA0
You can also search YouTube for "sine wave test tones" and "sine wave sweep."
Choose a sweep that doesn't take too long. Keep it around 30 seconds for the entire sweep. Here's an example of a good one:
Here's one with sine wave tones at regular intervals. You can just drag the playback seek bar to jump around to various frequencies (or just keyboard shortcuts J and L for big jumps, or left and right arrow keys for smaller jumps):
A quick side-note about speaker/room correction: For those of you who are interested in correcting your speakers and the room modes caused by acoustic issues in your listening area, I highly recommend IK Multimedia's ARC System 2:
http://www.ikmultimedia.com/products/arc/
I use it and swear by it. I get perfectly neutral/accurate frequency response at my listening position. My Klein+Hummel O 300Ds and Neumann KH805 sounds amazing with the ARC System. There are other similar products on the market (both hardware and software), and you can find them by searching for speaker/room correction software/hardware. ARC System is the best IMO because of how much it improved from the first version (the first version wasn't nearly as good, which is why some people used to poopoo it), how easy it is to use, and the result speaks for itself.
Here's how to use the tools:
Audio Frequency Sprectrum Analyzer - You need this to see exactly what frequency is playing during the log sweep. You need to put this first in your DSP signal chain, before the EQ, otherwise the EQ (or any othe DSP plugin) will skew its reading. So for example if you are using J River Media Center, you want to open up the DSP Studio and CTRL+Drag the Analyzer plugin all the way to the top of the plugins list, as far as it'll be allowed to go. Make sure the EQ plugin is lower than the Analyzer plugin in the signal chain.
Some analyzer plugins will by default use an algorithm that's not ideal for easy visual assessment, since it does not take into consideration human hearing sensitivity shift in high frequency range. This means if you look at the analyer during a log sweep, it'll seem like the amplitude gets lower and lower the higher the frequency gets. But don't worry about it, because we're only using the analyer to identify the frequency, not its amplitude. If it bothers you, you can switch to a different algorithm that compensates for it and will appear flat across the frequency range when playing back a log sweep.
This is how it should look like when playing a log sweep:
Log sweep - When you play a logarithmic sweep, what you'll hear is a sinewave tone that sweeps the entire audible frequency range from 20Hz to 20KHz. This is the fastest and most reliable way to hear any imbalances in your headphone, with the exception of the lowest sub-bass or highest treble frequency ranges (and we'll use pre-rendered sinewave test tones for that). When you listen to the sweep, if there are dips and spikes, you will hear them very clearly. A perfectly neutral/accurate headphone will sound very smooth during the sweep, without any significant dips or peaks. Most headphones will have significant peaks and dips somewhere in the upper-mids / sibilance range (3KHz to 8KHz) and the lower treble range (9KHz to 12KHz). You might also hear bloated mid to upper bass around the 125Hz to 250Hz range. Some headphones also have a dip right before the upper-mids in the range of 2KHz and 3KHz.
Note that it's natural to hear a gradual drop in amplitude when going from 6KHz and up. This is because highest frequencies will have far less energy to our human hearing. It's also natural to hear that from 50Hz to 40Hz, the bass energy feels stronger than the other bass frequencies even when they are in fact not louder. It'll sound like your skull is vibrating, and this is normal. In fact, if you don't hear them with slightly strong energy, that means there's a roll-off/dip. From 20Hz to 30Hz, it'll seem like the energy gets lower, and this is normal too. At such low frequency, they are more felt than heard. Also, many headphones cannot handle frequencies that low and are completely rolled off at such low frequencies. Just because the manufacturer lists 20Hz-20KHz in the specifications does not mean much because without specifying how much deviation, that headphone could very well be playing 30Hz at -12 dB, which is essentially useless. This is why the standard for specifying acceptable deviation in the professional audio industry is to use ± 3 dB (plus or minus 3 dB of threshold in acceptable accuracy). Consumer audio industry is far less strict with standards, which is why there is so much bull$hit and snake oil and outright lies and false advertising in consumer audio industry.
When you have identified the dips and peaks, write them down. Don't worry about how many dBs of difference to write down--just write down at what exact frequencies you're hearing audible differences.
Pre-rendered sinewave test tones - These are much better than sinewave generators because you can instantly switch from one exact frequency to another without having to waste time entering any numbers, and can jump between any frequencies instantly and repeatedly.
When you playback the sinewave tones, make sure you do not have any kind of normalization /smart volume/RepayGain turned on. This means any settings that will automatically make different songs playback at the same volume. Although the sineware test tones should be already normalized to the same volume, it's good to take this extra precaution just in case.
What I usually do when using pre-rendered sinewaves, is to first assess the bass region. start at 200Hz and sequentially play each interval all the way down to 20Hz. If the low sub-bass is rolled off or bloated, you will hear it.
Next, I would start from about 80Hz and start going up sequentially all the way to about 1KHz. If there's bloated upper to mid-bass I'll hear it. Headphones that sound overtly warm tend to have a thickness in the 200Hz range, and you should hear it.
Then from 1KHz, you go up sequentially all the way to 20KHz. Listen for noticeable dips and spikes. Some headphones will dip around 2~3KHz, and many tend to have significant dips or peaks between 4KHz and 8KHz. For IEMs, the resonance peak at around 7KHz is a big problem for some people due to the ear canal shape, and can spike as much as 12 dB or more.
Many headphone manufacturers tune their headphones to have a peak at around 10KHz to 12KHz (either narrow or wide bandwidth), to create the illusion of detail. This is not a good thing because it's not natural detail--it sounds artificial. Frequencies above 10KHz tend to sound rolled off, and the headphone itself will often in fact be rolled off in the highest frequencies. For those of you who are past teenage years, the older you are, the more you will lose hearing in the highest frequencies. Most older adults might only be able to barely hear 16KHz.
Understanding headphone measurements and ideal target response before we do corrections:
Now that you've listend and noted all the areas of imbalance in your headphone, we'll start correcting those problems.
Before we start with the EQing, it's important to learn about the Harman Target Response Curve. Do not skip this step--it is crucial to your understanding of how headphones are supposed to sound different from speakers, and why a neutral pair of headphone shouldn't measure totally flat. Read this two-part explanation of headphone frequency response measurements before moving on:
http://www.innerfidelity.com/conten...equency-response-part-one#PxVkSf8KrXBSEM6I.97
http://www.innerfidelity.com/conten...equency-response-part-two#W6HDd66v1yfLMk0R.97
You might also want look at the photo I posted of the Audez'e frequency response graph and superimposed correction in one of my old threads about EQing heasphones here (and read what I wrote about it just below the photo):
http://www.head-fi.org/t/551426/my-eq-curves-for-lcd-2-hd650-m50-and-007mk2
See that gently sloping white line of -10 dB I drew from 1KHz to 20KHz? That is what Tyll (of InnerFidelity) used to recommend as the ideal frequency response curve. As you probably already noticed, the Harman Target Reponse Curve is similar, but with a few differences. I currently prefer the Harman Target Response Curve (or Tyll's slightly modified version of it) since it's more updated and closer to what a full-range speaker system sounds like.
Now, lets start correcting the imbalance:
You start with any frequency ranges--it doesn't really matter. The key is to constantly assess your correction with the pre-rendered sinewave test tones and the log sweep (and you don't have to playback the entire log sweep when only listening for a specific requency range. For example, if you are working on the 3 to 8KHz range, you can jump ahead to around halfway point to start the sweep at 1KHz).
I usually start with the narrowest bandwidth (using peak/dip mode, NOT low/high pass/shelf mode) when applying EQ at a specific frequency. I will adjust the gain and then use the pre-rendered sinewave test tones in that range to listen for balance, until I get perfect balance. Once I have achieved that, I'll use the log sweep and listen of any dips or peaks right before and after the exact frequency I'm correcting. If the sweep is not perfectly smooth, I'll adjust the bandwidth until it smooths out, and if necessary continue to adjust the gain too. This takes patience because you need to get the bandwidth just right so the correction is countering the imbalance precisely. You do this for every single frequency imbalane that is audible.
Sometimes, a headphone might have the entire treble or bass region recessed or bloated, and you can switch to high or low pass/shelf mode instead of peak mode. Adjust so that visually, you can see the correction applied to the entire affected region. Then within the shelved area, if there are frequencies that needs to be corrected, you can do that with more dip/peak bands.
Keep in mind that sometimes you need to have EQ bands very close to each other (for example, one at 6KHz and one at 7KHz). when corrections happen in frequencies that are in close proximity to each other, they will also counter or amplify each other, so this is why you cannot do corrections by numbers alone and must visually look at the curve itself to make sure you are getting the exact correction you need.
Ideally, once you have finished all corrections, the log sweep will sound far smoother and balanced compared to no correction. And when you play the pre-rendered sinewave test tones sequentially through the entire frequency range, you'll hear no significant imbalances in amplitude (other than the natural Fletcher-Munson curve).
EDIT: It appears some people were confused by exactly what my method involves, especially regarding whether the Fletcher-Munson equal loudness curve and Harman Target Response Curve are involved in the EQ process. I'll summarize my process here again:
And here's an article/video showing Bob Katz and Tyll of InnerFidelity talking about EQing headphones to match the Harman Target Response Curve as a great starting point: http://www.innerfidelity.com/conten...es-harman-target-response#lgR1tm6s3SeYdDEO.97
Testing with musical material:
While it's important to also use musical tracks to test the results of your EQ correction, I have noticed that many people don't know how to select musical material that makes it very easy to hear problems in their audio gear.
The trick with easy audio assessment isn't simply to use music that's mastered perfectly--it's often the opposite (though perfect mastering is useful too, but in a different way). Music that is mastered too bright is great for revealing problems in the sibilance range, such as assessing headphones that are too bright and fatiguing. These masterings that are the bright side should still be within tolerable threshold when playing back on audio gear that is neutral/accurate, but if the gear is too bright, these bright masterings will become unbearable, and that's a very easy way to test/correct overly bright headphones.
You also want music that contains ample low frequency information, so that you can easily hear when the headphone's sub is rolled off and sounds too anemic, or when it's excessively bloated and muddy, or if it lacks proper amount of punch, articulation, and speed.
For the mids, you want music that has enough focus on the mids, such as vocals and instruments that are in similar frequency range as vocals (such as string instruments, woodwind instruments, guitars, etc.) If the headphone lacks presence or bite in the mids, you will hear it, and if it's too aggressive and nasal, you'll hear it too.
And of course, music that has a great balance in its arrangement and mastering is good for assessing overall balance, imaging, depth, etc.
Over the years, I've put together a playlist that I use specifically for testing audio gear, and I'm now going to share them with you and show you how to use them. If you really like any of the tracks and want more information to hunt down the composer/artist so you can purchase their music, don't hesitate to ask to find out more. Here's the link to download the tracks:
edited:sorry no link to copyrighted musics(signed evil modo)
(With the link to the MP3s removed, I've tried to find the songs on Youtube to link to instead. I couldn't find all of them, but found enough of them to be very helpful.)
I'll group these tracks into categories based on how I use them. (Don't think about the actual genre/language of these tracks because you're using these tracks for testing audio gear, not for leisurely listening. If you happen to enjoy the music, that's just icing on the cake.)
And don't get hung up on the fact that these are MP3/YouTube audio. They are perfectly fine for the kind of tests you will be doing with them, and using uncompressed versions will not change anything. You'd have to focus intensely and strain really hard to hear any differences compared to the original CD quality (and you often won't hear any differences reliably if the YouTube video is HD) and at that point the differences are too subtle to be relevant in this specific context.
When you do comparisons of musical material, it's important to not switch between the EQ curve and the bypassed/unEQd sound very quickly without allowing your brain to first adapt to the sonic signature. That is not the correct way to do audio assessment. What you need to do is to listen for at least 30 seconds so your brain can adjust itself to that sonic signature. When you do switch, stay with it for at least 30 seconds before you switch again.
Testing for bass:
04 - Love For Sale - The upright bass in this track will sound too muddy and bloated with headphones that have too much bass. The bassline will overwhelm the rest of the frequencies too much and you won't hear the articulation of each bass note clearly. On neutral sounding gear, the bass notes will sound balanced with the rest of the arrangement.
12.sub-ID%20Mi%20Woof - This track can test just how deep and powerful the bass frequency response is. If your headphone's bass it too rolled-off, you will not hear/feel that really powerful and deeply resonant monster Roland 808 drum machine kick drum. It is one of the most famous kick drum sounds in the world (at least in hip-hop and electronic music), and you definitely want to be able to play it with enough authority and power. If the gear is too exaggerated in the bass region, the 808 kick drum will sound ridiculously monstrous and way out of balance. Also, if your headphones cannot handle the powerful bass of the 808 kick drum, it might distort.
02 - klendathu drop - At 0:49, you should hear a powerful sustained bass note centered at around 45Hz. If it doesn't sound powerful and majestic, then your headphone's sub-bass is too rolled off. If it sounds too overwhelming, then your headphone's sub-bass is too exaggerated.
lucite-dry - This short clip has bass kick that should sound punchy at with deep and powerful resonance (though not as low in frequency as the track with the 808 kick drum). If the bass kick sounds limp and weak, then your headphone's bass response is too anemic. If the bass kick sounds too overwhelming, then your headphone's bass is too exaggerated.
03 - Mahalle - Starting at 0:16, you should hear that deep sub-bass thud that accompanies the strumming instrument each time it repeats. It should be very obvious and accents the start of each repeated strumming pattern. If it's not strong enough, the track will sound too thin.
02 - L.F.O. - At 1:38, you will hear just the sub-bass notes by itself. Like the 808 kick drum, if your headphones cannot handle the sub-bass energy, it could distort. On neutral gear those bass notes should sound clean and deep.
Halo Theme - The intro section of the Halo theme is great for testing sub-bas because it has prominent low sub-bass energy going all the way down to 20Hz. If your speakers or headphones cannot handle that much low sub-bass frequency, it will likely distort, or sound too anemic/wimply instead of majestic and epic.
11 - Dogfighter - Although this track is a hybrid of jazz fusion and orchestral, the bass kick should have a solid and beefy thud that's fairly strong, but not overwhelming.
02 - 戦いの達命 (Battle of our lives) - The big bass drum hits between 0:17-0:18 and 0:19-0:20 needs to sound strong and deep and really punctuate the orchestration with a sense power.
bigbeat01 - Here's another deep sub-bass track. When the bassline starts at 0:24, it should be very full and deep and powerful, as if your entire head is rumbling. If you are not hearing that, then your headphone's sub-bass is not deep/powerful enough. If it's way too overwhelming and sounds like a muddy mess, then the sub-bass of your headphone is too exaggerated.
track 06 - cats on mars (dmx krew remix) - A fun track for sub-bass. The bass kick should should sound full and deep and with ample body. If it sounds weak then the sub-bass is too rolled off. If it's too overwhelming, the sub-bass is too exaggerated.
Testing for the mids:
02 - Rivers Of Love - This track's vocals need to sound clear, warm, and rich. At 0:54 when she sings "through a scene," the word "through" will sound too wooly if the headphone has problems with the mids such as being too thick sounding. Also, there are some spots in the vocals that can have a bit of sibilance if the headphone's too bright.
156608_rp350_11 - This electric guitar track needs to have enough bite in the upper-mids in order to retain its sense of raw power. I distinctly remember when I first got the Aucez'e LCD-2 and tested it, I was disappointed by how recessed it's upper-mids were and how wimpy those guitars sounded.
06 Rain (I Want A Divorce) - This string arrangement needs to sound rich and full, and not thin and brittle. The bass notes that punctuates throughout the track needs to be deep and rich and really fill out the bass frequencies and provide an authoritative anchor for the arrangement.
02 - 戦いの達命 (Battle of our lives) - This track was already mentioned for the big bass drum hits, but the brass-heavy orchestration is also very good for testing the mids. Brass intruments, like distorted guitars, needs to have enough bite and presence to convey that sense of raw power.
[BARBEE BOYS]BARBEE BOYS 2(07)‚Í‚¿‚ ‚킹‚̃ƒbƒJ - A J-Rock song from the 80's that can sound too aggressively bright if the headphone has too much emphasis in the upper-mids.
Testing for upper-mids/brightness problems:
15 - Fade - The bright sounding percussive chords that starts at 0:38 will sound fine on neutral sounding audio gear. It's bright but not grating. On gear that's too bright, they can sound piercing and painful.
09 - 遠雷 (Distant Thunder) / 06- MORNING CALL - Both of these tracks have intros that have bright shimmering sounds that repeat in patterns, and on headphones that are too bright, they can sound too sharp and fatiguing.
05 - 少年は天使を殺す (The Boy Killed the Angel) - This track is mastered quite bright and on neutral audio gear, the backing vocal chant that comes right after the opening drum beat will have sibilance that's very prominent, but it's still within tolerance and should not sound painful. But on audio gear that is too bright, it will become very annoying and even painful. This continues all the way through the song with both the main vocals and the backing vocals. Also, that snare drum can sound fatiguing on audio gear that's too bright.
02 - Here's Where The Story Ends - This track isn't overtly bright, but when the chorus starts at 1:10, the sibilance can become too hot and annoying on headphones that are too bright.
01 - Already Yours - This wall-of-distorted guitars track can get fatiguing fast if the headphone is too bright. On a neutral sounding audio gear, it's tolerable to listen all the way through, but any skew towards excessive brightness it'll become unbearable.
03 PARTY - This track's chorus starting at around 0:48 has some sibilance issues, and on neutral audio gear, it'll sound a bit hot in the sibilance but still tolerable. On gear that's too bright, those distinct sibilances will become painful.
05 - 淋しいから言えないから (Because I Was Lonely, I couldn't Say) - Morikawa Miho's voice during the chorus starting at 1:07 can sound overly sharp on audio gear that's too bright, but if the gear is neutral, she'll sound bright but not painful or grating.
Testing for treble:
09 第七感（セッティエーム サンス) (Seventh Sense) - This track has really tizzy and splashy treble. It's still within tolerance on neutral audio gear, but on gear that's too bright in the treble, it'll become really tizzy and annoying, and maybe even painful.
10 William, It Was Really Nothing - This track also has a lot of treble energy, but it's not quite as drastic as Seventh Sense. On headphones with exaggerated treble, it will sound like a splashy mess instead of shimmering.
For overall balance, clarity, and richness in complex arrangements:
Mojo_Madness / ST_Theme_v1 / 03 Come in 007, Your time is up / 07 BROTHER - These action/adventure/drama oriented score cues have rich orchestrations and I like to use them to listen for overall richness, balance, clarity, dynamic range, imaging, depth, etc.
The forum only allows 30 media links per post, so I'm moving the ones past #30 to another post. Here's the link to that post:
https://www.head-fi.org/threads/the...nd-for-non-professionals.796791/post-15562131
Wrapping up:
Okay, I hope this has been helpful to those of you who always wanted to try EQing your headphones but don't know how, or felt confused/lost when using pink noise and sinewave generators. My method of using logarithmic sweep tone and pre-rendered sinewave test tones is much more straightforward and intuitive, and anyone should be able to do very precise EQ curves using my method and achieve the most neutral/accurate frequency response for any headphone (unless that headphone's frequency response and physical drivers are so bad that it just can't be corrected properly).
Last edited:
