Building a Notch Filter for the Beyerdynamic DT 770 Pro Headphones with Measurements

In this article I’ll go over how I designed, built and tested a passive notch filter for the Beyerdynamic DT 770 Pro 250-ohm closed-back headphones and include measurements using REW of the different filters responses. I bought these headphones over Black Friday weekend last year from Beyerdynamic’s website where they were doing a deal where you got a free pair of black Velour EDT 770 VB ear pads, a $40 value, plus the headphones were another $10 off their usual price of $169 and they did free shipping. I was in the market for some new headphones to replace my trusty old Audio-Technica ATH-M30’s that I’ve had for several years and I thought these would do quite nicely. That same weekend my son and I had stopped by Guitar Center and they had about 8 pairs of headphones on display that you could listen to with sample music of varying styles. We checked each of them out and after a quick listen it was pretty obvious the Beyerdynamic DT 770’s were the clear standout. Good bass, nice treble, clean and clear midrange, not to mention they were the most comfortable of the bunch. Before this I hadn’t even considered these headphones, and up until this point I had already picked up a pair of Sennheisers that were also on sale over Black Friday. Though I wasn’t able to A/B these two headphones directly, I decided to return the Sennheisers and picked up the Beyerdynamic DT 770 Pros instead. A decision that would later send me down this rabbit hole of researching filtering and EQ options in order to make these headphones actually listenable for the long term.

While these headphones do sound great overall, they suffer from one major flaw that could be a deal-breaker for most – they are extremely bright! Doing a direct comparison to some of my other headphones, these suckers have got some highs that seriously sparkle, and it’s not subtle. People say that this is just that classic Beyerdynamic sound and I will say that for some tracks, this does sound fantastic and, in short durations, is manageable. But in general, most songs sounds bright and harsh, making long listening expeditions not that enjoyable at all. In all other regards however these headphones sound amazing and they are super comfortable (even while wearing glasses, like I do) so after giving the DT 770s a fair shot straight up, I decided I had to figure out how to EQ these things into submission. Fortunately there’s a ton of info for these headphones on the internet, and I quickly realized I wasn’t alone in my perception of these headphones. And while there are plenty of reviews that will talk about how these headphones sound, I found three major sources to be the most valuable in providing raw measured data that enabled me to tackle this treble problem a little more scientifically.

Naturally my first stop was Rtings.com. Without question, one of the greatest resources for objective reviews of headphones and TVs. They give the DT 770’s an overall Neutral Sound score of 8.3 which is quite high and favorable. Raw frequency response measurements confirm what everyone can hear, that these headphones are extremely neutral (in terms of matching the Harmon target response) from 20 Hz up to almost 6 kHz and therefore have excellent lower-bass response and mid-bass response up through the midrange, but above 6 kHz is where things get a little wonky. The treble climbs well above the target by over 10 dB and peaks at around 8.2 kHz. Even though they provide frequency responses for all their headphones, they don’t offer filter or EQ suggestions on how to improve them. So you’re kind of on your own to figure out how to fix anything that looks amiss. But this is a good start.

That’s when I ran across another incredibly valuable resource for headphones, plus a second set of measurements for the DT 770s, and that was from oratory1990 over on Reddit. oratory1990 has measured response plots for the DT 770s including EQ options to bring them into alignment with the Harmon target. This is where for the first time I downloaded and played around with Equalizer APO and got my first taste of the possibilities for active EQ in Windows. I entered in the EQ settings he recommended and WOW, what a difference! For the first time these headphones sounded absolutely amazing all the way through, that overly bright treble, the sibilance, the shrill, was all gone. I couldn’t believe it was that easy and with that I had figured I had my solution. While I wasn’t a super fan of having to use a software-based EQ, it was powerful enough, and seamless enough, so, why not?

Well, I quickly learned why not, at least for me, that this solution was not a long-term fix. I planned on using these headphones in my mini-budget home recording studio and I quickly realized that Equalizer APO does not work with ASIO drivers and it does not work with live line-in sources, like a piano, microphone or guitar. Basically it doesn’t work with Reaper (my go-to DAW), as it does not EQ anything recorded in it or played through it or anything live via the line in. The EQ only applies only to pre-recorded music played directly in Windows, like with Spotify or YouTube. I thought I could build a similar filter in Reaper using ReaEQ, but quickly learned how much of pain that was as well. Not mention, the biggest gripe about using a software-based EQ was having to turn it off any time I switched to using speakers. So with the Equalizer APO not working for most of my intended applications, I ended up searching for another solution. Now if your setup is more dedicated to just playing back music, or you don’t have speakers, then Equalizer APO is fantastic and I found it to be super useful.

So then last but not least, I came across DIY-Audio-Heaven and as the name suggests, I was in heaven! I have to give mad props to Solderdude who basically inspired this entire project by being the first person (as far as I could tell) to offer a fully passive solution to fixing the Beyerdynamic treble peak. And it’s so simple! I don’t know why I didn’t think of this before. As someone who has designed and built so many freaking speakers, with crossovers, and filters, even notch filters, I don’t know why the idea to passively correct the DT 770s didn’t come to me sooner. But Solderdude absolutely nailed it with his notch filter design, and not to jump ahead too much, this one fix literally solved the treble peak issue with these headphones and has made them sound just incredible. Amazingly enough, he provides passive filter recommendations for hundreds of headphones, not to mention provides measurements and reviews of hundreds of headphones, for your reading pleasure. As it turns out, a lot of headphones that are bright, or sound bad for a myriad of other reasons, can be improved significantly with a single passive in-line filter. And he’s got options for how to fix a lot of them. Not only does he provide this as a service, where you can buy them from him, he basically gives you the design, so if you’re a little more into DIY, then you can tackle this project on your own pretty easily. This was all I needed to start down this path of how to passively EQ the DT 770 Pros to make them sound better.

Now I took the idea a bit further and did A LOT of modeling and simulations of different notch filters, including predicting what the DT 770s would sound like with tons of other filter options without ever building a single one of them. I then built a handful of prototypes and got some real measurements, and finally landed on a single design, that I went and bought parts for and then built the final filter. All the while, I was taking measurements, running simulations, and finalizing the overall design. But I will say that you don’t have to do all that to get great sound out of these headphones. Solderdude’s original filter design is pretty much all you need to build to get these headphones in check. While I tweaked the design by making the filter attenuation adjustable with a switch, and did a bunch of measurements, including measuring my own DT 770s, you really don’t have to do all that. I’ll provide the parts list for the inductors, caps and resistors I selected, since I did spend some time making sure I bought the right type of parts per Solderdude’s recommendations. I had a ton of fun designing, building an measuring different filters in REW and just seeing the simulations match up to the measured performance is always a good time.

So we know that there are at least 3 different sources for measured FR plots of the DT 770’s, so from there I pulled each measurement into FPGraphTracer and digitized each measurement so I could pull them into REW and actually compare them, and most importantly, start to manipulate them with filters from a simulation perspective. One thing I noticed was that while in general the measurements tracked each other, they sort of disagreed on where that treble peak actually occurred and how high that peak really was. Either this could be attributed to measurement error or variation in the headphone’s manufacturing or both. I started to go down the long and windy path of understanding headphone measurement techniques, and all the rife that exists in doing so, and while most people agree that any kind of DIY headphone setup is going to be garbage, I thought I’d give it a try anyway. So in a lot of these plots you’ll see what I called Dan’s Flat Board measurements which were taken with just a small board with an ECM8000 microphone stuck in the middle of it and one headphone cup centered directly over it. I make zero claims as to the accuracy of this measurement method and was only using this technique to compare the headphones to themselves through different filters. So in effect, when compared against the same setup, the response is actually quite useful for dialing in the filter for the right frequency, to target the peak, and the right amplitude to see how to attenuate the peak. Not to mention as a means to ensure the final filter was built correctly and performing as expected. To my surprise, I think I got pretty close with my initial measurements to what everybody else was also measuring. So that was kind of nice. It was sensitive to ear cup placement however and a week later after I tore down and built back up the measurement gear I couldn’t repeat the results exactly as I had before, especially in the bass. So you will see that in the measurements. This had nothing to do with the filter as the baseline measurements without the filter show the same response. I went back again and redid the headphone measurements with the final filter using my JDS Labs Atom Amp+ and I got the bass back. That’s when I realized that the bass response was sensitive to how far apart the ear cups were during the measurement, or how much pressure the ear cup being measured was applying to board. Suffice to say, the bigger your head, the more bass you are likely to experience.

Now with measurements of my exact DT 770s in the bag, taken with my super crude mic-in-a-board, I could see the treble peak that was causing so much trouble. Turns out to be more like 7.25 kHz, which aligned with Solderdude’s measurements much more than it did with Rtings. oratory1990’s measurement had a lot of smoothing so it was hard to tell where the real peak occurred but it was somewhere in the ballpark. I didn’t stress too much the massive suckout in all my measurements at 4 kHz. I suspect this was an artifact of my setup as oratory1990 and Rtings do not show this. However Solderdude did measure something similar, and he believed also it was not a real suckout. Audibly music does not appear to have this deep a depression, so I’m fine with saying it’s probably not real, and for the most part am just going to focus on cleaning up the 7.25 kHz peak. Now this is where I combined the techniques of oratory1990 and Solderdude by building a simple notch filter in Equalizer APO to match what could be built passively. This was so I could get an idea of how much improvement I could get out of the DT 770s by forgoing the elaborate and complex EQ and making just one notch filter right at the area of concern. While oratory1990’s EQ fixes other issues with the DT 770s, in terms of matching the Harmon target, it would be impossible to build a passive filter that achieved the same results. So I disabled every filter except for one, and set it for 7.25 kHz with a Q of about 5 and an attenuation of 8 dB. And that was it! While it didn’t sound quite as good as having all of the individual PEQ’s enabled, it solved what I would call the most offensive non-compliance to the target. In fact, it was pretty difficult to tell the difference, in the highs at least, between going from the full oratory1990 recommended EQ to just a single 8 dB notch at 7.25 kHz. Thus began the pursuit to build the perfect single passive notch filter for the DT 770 Pros.

One of the biggest advantages of Equalizer APO is it allows you to make instantaneous tweaks to the filter real-time while listening to the changes. I spent some time listening to variety of tracks while tweaking the dB attenuation, and kept record of what I thought sounded best based on and what could I actually hear. I kept the notch centered around 7.25 kHz since I knew by my own measurements that that was the area that needed the most attention and just adjusted the attenuation. From that I pretty much unanimously thought a 10 dB drop sounded the best. But, for some tracks, it took away too much and 6 dB sounded better. Sometimes 6 dB wasn’t enough and around 8 dB was better. So this is when I came up with the idea that it would be much more cool to build a passive filter with a switch that would allow up to 3 different attenuator options while simultaneously not complicating the design much. Basically the way an LRC notch filter works is the L and C components set the notch frequency and Q while the R component sets the attenuation. So while it would be much harder to design a passive filter with an adjustable center frequency and Q, it’s much easier to make one with an adjustable, or selectable, attenuation by simply changing the R value. Which is exactly what I did here.

So I have a schematic and parts layout diagram for how to make a filter with a selectable attenuation using a DP3T switch. The idea here is when the switch is in the center position, only the one resistor is in the circuit, which is the largest value resistor, and provides the most attenuation, or the deepest notch (and consequently the highest Q). The other two options add either of the other two resistors in parallel to the main resistor which drops the resistance by the formula Rt = 1/(1/R1+1/R2). In this case you can choose two other values to provide two other unique attenuation settings for the filter. You can simulate this effect with various online modeling tools and even do it in Passive Crossover Designer 8.0. While the models provide a good representation of the frequency of the notch, I’ve found them to less accurate in terms of modeling how deep the notch actually ends up by a couple of dB. This could be in part due to additional resistances in the real circuit, such as your headphone amp, that you need to consider in the models. Either way, I ended up building and measuring the prototype filters with a whole bunch of different resistors ranging from 220 ohms to 2,000 ohms to completely open (or basically just an LC circuit) to see what the real effect is in practice. You can then simulate these different filters’ effect on the DT 770 headphones by exporting .frd or .txt files from your simulation tool, or the measured plots from REW, and bringing them into REW. In REW, using the math function, you can apply these filter to the raw headphone’s measured response and see what the overall effect will be. I did this with a bunch of filter designs and applied them to all of the different measurements I had scrubbed from online including the one I took myself. I snagged the 2018 Harman target curve as well, so you’ll see that graph plotted here for comparison.

Without going into all of the options one could explore (I created this massive Excel spreadsheet that let me quickly tweak with different LCR options, see here to the left), I landed on using a 0.22 uF cap with a 2.2 mH inductor (hitting that 7.25 kHz peak on the nose) and then went with three different final resistor values of 1 kohms, 500 ohms and 280 ohms. This resulted in a notch that could be selected between approximately -10 dB, -7.5 dB and -5 dB. This matches pretty closely to Solderdude’s recommendation for the same value L and C but with a single 330 ohm resistor. I went a little more aggressive with the attenuation options, but left one option pretty close to his. I think a 330 ohm resistor will land you right in the 6 dB range which realistically is also pretty much spot on to bring that peak down to the Harmon target. While I could hear differences between these attenuation settings by playing around in Equalization APO, once I finished the filter and switched between the different settings, they all sounded the same to me. In fact, upon first listen, I thought maybe I had built the filter wrong. Subsequent measurements proved the filter to have been built right. But oddly enough, being able to pick out the difference in depth of that null across a 5 dB gap was harder than I thought it should have been. I suspect this is in part because the Q of the filter is pretty high, it ranges from 2.8 to 10 depending on which resistor is selected, and thus its effect is over such a small bandwidth, that it becomes harder to hear even a 5 dB change in amplitude. For now I’ve just left it at the max attenuation setting, or 1 kohms, and have been very happy with it and it definitely makes an audible difference over no filter at all. Note that if you wanted you could use the DP3T switch to provide two attenuation options, like 5 dB and 10 dB with one option being 0 dB, which would remove the filter all together. Just select one of the resistors in switch to be a 0-ohm. That will effectively bypass the LC circuit completely, bringing back that classic Beyerdynamic treble peak. That option always exists however, just by unplugging the filter and plugging the headphones straight into the amp.

Anyway, check all all the pictures and graphs from REW that I created while working on this project. I tried to grab just the useful stuff and make some comparisons that seemed logical. I annotated a lot of the pictures and graphs as well so please click through and read my notes for more information on the data. I ended up with hundreds of measurements overall and ran simulations of way more filters than I showcase here. I’ve included links throughout this post to some of the various resources and tools I used so feel free to click on those to check out more. And if you’re curious about whether or not this filter decreases the sound quality in some way, or adds distortion, check out the distortion plots below taken with and without the filter. The % THD shown in the plots were set at 7 kHz, where the distortion is the highest, and yes, it does get worse with the filter in place, however, it is predominantly 3rd harmonic only and when compared to the distortion of the headphones themselves, it is quite negligible. Other than suppressing the treble peak, I hear no other differences in sound or sound quality with the filter in place. I think the trick here too is picking high-quality part for the filter, like the ceramic C0G caps, a low-DCR inductor with high saturation current and metal film resistors. Oh and one more thing, for the 1/4″ headphone M/F cable I grabbed this one from Amazon and for the little project box I bought a 5-pack of these. The fit was a bit tight, you might want to go a size bigger. So far though I am super happy with this little LCR passive filter for the Beyerdynamic DT 770 Pro headphones. I was seriously about to return them until I found this sweet trick. So shoutout to Solderdude at DIY-Audio-Heaven for giving me the inspiration to take on this project. It’s made all the difference and now I can enjoy listening to music and creating music on these otherwise awesome headphones. Thanks for stopping by and please leave a comment below if you enjoyed this article. Do you own a pair of DT 770 Pros? Would you build this filter?

Schematic and Design Layout

Exports from REW – Measurements and Simulation

Distortion Plots (not with headphones)

Assembly Pictures

DIY Headphone Measurement Rig

Final Pictures of My Setup

About Dan

For a complete biography, stop by www.danmarx.org and click around for a while.
This entry was posted in DIY Audio, General and tagged , , , , , , , , , , , , , , , , , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *