How We Created Our House Curve

What Is a Frequency Response Curve?

To judge how faithfully earbuds or headphone speakers reproduce the music, podcast, or game you love, it's common practice to measure their frequency response. A frequency response shows you how loudly the speakers play each individual frequency, or pitch, in relation to the rest. It's usually charted as a curve, with amplitude displayed in decibels (dB) on the x-axis and frequency displayed in Hertz (aka Hz or sound waves per second) on the y-axis.

Spikes in the curve mean that those frequencies have relatively higher decibel levels and sound louder than the rest. Dips mean they're softer, with lower decibels. It sounds simple enough. But, if you're not familiar with frequency response curves, they can be tricky to read. How do you know if it's a good curve or a bad one? This is a trick question because a listener's preference, ear health, and expertise all play a role. That said, while most of us have strong opinions about the best place to get a slice of pizza, a truly great pizzeria makes a slice that works for almost everyone. The same is true of headphones and earbud speakers. Most expert testers agree on the best sets, and those sets tend to have very similar frequency response curves. They are so similar, in fact, that you can combine the best examples to create an idealized target response curve. We've done this. We call it the GearLab House Curve.

Brief History of the Harman Curve

While creating a target response curve may seem like a no-brainer to judge the accuracy and quality of headphones and earbuds, we owe this idea to an innovative group of researchers at Harman International. Here's the story.

Walkmans popularized headphones in the 1980s, iPods and wired earbuds took over around the turn of the century, and modern, high-quality headphones gained traction in the late 80's. By 2012, sound engineers were trying to reproduce balanced, accurate sounds that sound as though you're in the room with a band through tiny speakers placed over or in your ears. It was a herculean challenge, and headphone quality was all over the place. Many listeners were missing the polished products created by hard-working artists, composers, and sound engineers.

In 2012, Dr. Sean Olive and Todd Welti of Harman International recognized the issue and designed a double-blind experiment to compare listener feedback to the frequency response curves of six headphones. They found the headphones that had a gradual and neutral curve sounded best, then used those examples to define the first-ever target curve for the industry.

That first test polled 10 trained listeners. Subsequent research surveyed 283 people with more age, gender, and critical audio experience diversity from Canada, China, Germany, and the United States. Most were Harman employees.

Harman has since expanded the number of headphones and earbuds they've tested and the number of listeners who've judged them, creating refined iterations of the Harman Curve along the way. With that many data points, the Harman Curve is considered an industry standard. It both helps you judge how well a set of headphones works and helps sound engineers create accurate, high-quality options.

Knowledge Gaps

The Harman Curve isn't perfect, though. In their tests, Dr. Olive found that their target curve matched the musical tastes of 64% of users. That's a strong majority, but it may not be you.

Finding the best target response curve is still a moving target, particularly for in-ear headphones or earbuds. The original Harman Curve was developed for over and around-the-ear headphones. In a 2017 presentation, Dr. Olive shared that those target curves may not apply to in-ear headphones, i.e., earbuds. That's because, around the ear, headphones interact with the shape and acoustics of your outer ear. Earbuds do not. Earbuds also produce unique acoustics in the ear canal and are highly dependent on a good seal, which impacts frequencies below 500 Hz. That's why it's so important to use the right size when your earbuds come with a number of tip options.

Target response curves also need to be specific to the type of device you are using to measure your headphones. These test devices replicate the natural shape and acoustics of a human ear and record and analyze sound using a microphone where our eardrums would be. The original Harman research used the GRAS 45 CA. We use the B&K 5128.

The GearLab House Curve

In short, the gauntlet has been thrown. What's the best target curve when using the latest and greatest equipment in the audio world? We answered it by creating the GearLab House Curve. It emphasizes a bit more treble and bass than the Harman Curve and reflects our expert listener team's tastes. We think you'll like it too.

Think of the House Curve as a starting point, a truly great margarita pizza. Then, if you want to add some extra treble (toppings) or bass (thick crust), you do. Simply compare the measured frequency response curve of the earbuds you're interested in to our House Curve to help you find the right option for your musical tastes. Here's how we did it.

We started with dozens of earbuds, two expert listeners,SoundCheck software from Listen Inc, and our very own B&K 5128 High-frequency Head and Torso Simulator (HF HATS). We call it Darwin. Darwin performs a range of tasks, like testing microphones and noise-cancelling features, but the frequency response is among the most important. To do this, it sends a consistent tone, a soundwave with a consistent amplitude, across all audible frequencies and measures any points where the decibels increase or decrease.

Our experts, Gena Bradshaw and Dave Berkman, started by critically listening to and rating each headphone's bass, middle, and treble response using a carefully curated playlist. Though their musical tastes differ — Berkman prefers jazz and heavy metal while Gena loves classic piano pieces and pop — both testers arrived at similar results.

We then had Darwin listen in (triple-checking to ensure a good seal) and measure each earbud's frequency response. Using additive synthesis, we combined the top five response curves for each frequency range, meaning that if a pair of earbuds or headphones ranked number four in base tones but number 25 in treble, only its low frequencies factor into the target curve.

We peer-reviewed the process and the results with multiple professional sound engineers from SoundInc and beyond and got the green light. Our results are similar to the Harman Curve except, as we mentioned, for slightly more emphasis on the treble and bass ranges.

How the House Curve Informs Our Test Results

After creating the House Curve, we carefully compared it to each earbud's frequency response to help us visualize what we're hearing. For instance, if the treble sounds too bright, we would expect to see a spike above the House Curve in the high frequencies. A hollow or weak-sounding bass line should appear as a dip in the chart below 200 Hz. We found that the House Curve closely aligns with our expert listeners' favorite pairs, giving us a great deal of confidence in our results.

Comparing the ratings and opinions of our test panel to each earbud's curve allows us to double-check their subjective opinions using objective data. Combining our expert listener's preferences with the data contained in our House Curve helps us identify which earbuds will work best for people with various styles and listening habits.

If you love podcasts, for example, it's important that high bass and low middle frequencies are neutral and not overpowered by the bass. We can see that in the response curve and have our testers verify that it works in practice. If you have to have the most powerful bass that pocket-sized earbuds can possibly produce, we make sure our expert listeners confirm those results.

The more you learn about what you like to listen to, the more useful a target curve will be. It's also important to keep in mind that most earbuds can adjust the relative levels of their base, middle, and treble frequencies. We test their baseline settings, but note if you can tweak them a bit for yourself. Keep in mind that we've never had luck making terrible earbuds sound better using these minor adjustments. They're best used to make a great set sound just right for you.

Our Panel of Audio Experts

Meet the team who contributed to the creation of our House Curve: Our expert audio team encompasses a range of tastes and audio expertise, from a classical music pro to a bass guitarist and DJ with an extensive music catalog to a Salt Institute-trained podcast producer.

Genaveve Bradshaw holds a bachelor's degree in music from the University of New Mexico, where she studied classical singing, music theory, and composition. Gena plays piano, guitar, and flute and has performed in musicals, operas, songwriter showcases, plays, recitals, charity fundraisers, sports events, and variety shows. Each of these performances presents unique audio challenges and must be sculpted to accommodate the space and the crowd. To complement her ear training and theory background, Gena has a degree in Music from the University of New Mexico, and continues to study music production and sound engineering. She's a regular Sound Gym.

Dave Berkman is the owner and head DJ at Ascent Productions, as well as a bass guitarist. Berkman pushed every set of earbuds to their vibrational breaking point to test the limits of their bass resonance.

Lead researcher Graham Faulknor holds a master's degree in mechanical engineering from the University of Colorado, Boulder. In his 15-year engineering career, he designed, built, and tested a wide array of devices, from satellites to medical devices to race cars. He's now putting his considerable powers behind our sound team's test development and analysis.

Conclusion