stellarelephant
100+ Head-Fier
Hi, Folks. I've spent the past month refining these custom EQ settings for the Sennheiser HD 598, and I hope they will give you all as much enjoyment as they now give me. IMO, they take this already excellent headphone up to the next level in terms of neutrality, bass impact, and top-end clarity. A reference quality listening experience was my goal, and I believe I've attained it. Like, in the ballpark of +/-2 dB response all the way through the useful range of these sweet 40mm drivers...the holy grail. But you be the judge!
These EQ curves were designed to precisely flatten the 598's frequency response and deliver extremely accurate sound. I used Logic Pro's native parametric EQ plugin, Channel EQ, to set initial EQ values. These values were then entered into two Rockboxed DAPs for testing--an iPod Video 5.5 gen, and an xDuoo X3. I've listened to the settings for 100 hours or more, fine tuning by ear along the way, for the most natural sound possible.
If you use Rockbox on your player, you can download my EQ settings here and copy them onto your device.
https://www.dropbox.com/sh/5as9jxl5gi19h39/AAAX6x7p4jYkcnrgYql4WnZ1a?dl=0
You can also simply enter the values listed way down below into any good parametric EQ plugin on your computer.
Here's the lowdown:
The above picture shows HD 598 measurements as published by Innerfidelity.com and Headphone.com, with my compensatory EQ curve sandwiched in the middle. Big thanks to these two awesome websites for their very accurate measurements of the 598. You guys rock. As you can see, my EQ curve is essentially an inverse of the other two. My goal was to attain a razor-flat response, which yields a realism far beyond what I have ever heard from a headphone before.
STARTING POINT
The HD 598 is perhaps the best candidate for corrective EQ that I have come across. Spend some time on headphone.com comparing the 598's frequency response to other headphones, and you'll surely see why (click Learn -> Build a Graph). While the 598's natural response does not give as much HF and LF extension as some pricier models, it is exceptionally smooth overall, more so than models selling for five times the price. There are very few sharp peaks, which is a huge advantage when using EQ. Making sharp cuts and boosts in the signal using high Q values can cause audible phasing issues in my experience. EQing a headphone from the Grado lineup, for instance, which has sharp breakup modes throughout the treble region, is a nightmare. Owing to it's smooth response, EQing the 598 is a dream! Add to that, the 598 actually responds very well to a bass boost via EQ. Some lesser headphone drivers simply cannot push much air without distortion, and lose all composure when low bass is boosted. I was pleased to find that the 598's drivers are up to the task, and can indeed produce respectable low bass with impact. The 598 has amazing potential.
The 598 isn't perfect, however. It suffers from the following sonic deficiencies when played without EQ:
1. Poor low bass extension below 60 Hz. This results in barely any perceptible bass impact whatsoever.
2. Upper bass / low midrange bloat from 100 to 400 Hz. This buries the low bass in the mix even further, and makes the bass sound wooly and slow.
3. Rolled-off treble above 1000 Hz, a.k.a. the famous Sennheiser veil. This pushes vocals back in the mix and softens the overall presentation unnaturally.
4. A resonant mode around 9900 Hz. There's only one, and its not a bad one, but the peak is there, and it causes listening fatigue.
5. Drastically rolled off high treble starting at 10 kHz. Where's the air?
PARAMETRIC EQ
Well, I set out to correct all of these issues with a parametric EQ. Parametric is the key word here. A simple graphic EQ will not give you the control you need to sculpt the signal with precision, although it could possibly yield some benefits for some the issues listed above. Certainly not for number 4. With a parametric EQ, each boost and cut may be performed with a variable bandwidth, called Q. I spent a long while in Logic Pro sculpting the EQ curve to perfectly mirror the 598's measured response curves.
TUNING BY EAR
The original EQ values were loaded onto both of my Rockboxed DAPs for testing--an iPod Video 5.5 gen, and an xDuoo X3. I needed to determine a few things, including how much bass the 598 could handle from a low shelf filter, and the exact location of the driver resonance mentioned earlier. I played a huge variety of music through both devices while tweaking with the EQ bands.
To dial in the bass, I listened mostly to Random Access Memories by Daft Punk, Junto by Basement Jaxx, and Voodoo by D'Angelo. All of these albums go boom. In the end, I settled on a 3.2 dB Low Shelf boost, centered at 60 Hz. Hello bass impact! I find it impressive the that 598 can handle a boost this large without distortion. The bass is still a tad reserved, but it digs much deeper now.
To find the exact frequency of the resonant band, I listened to high-pitched vocals with a lot of ambience. The self-titled album by Pentatonix was helpful, along with So Familiar by Steve Martin and Edie Brickell. For me, the resonant band turned out to be at 9950 Hz. I notched it out with a narrow 3 dB cut. Here's the catch: I believe that this frequency varies from driver to driver. I noticed some discrepancy between the two response graphs I studied, and it looks like the pair of cans measured by Inner Fidelity actually varied between left and right sides! I recommend that anyone who tries these EQ settings play some music with some good treble content while sweeping the notch between 8500 and 10k. When it sounds most natural, you've got it. Alternatively, you can sweep with a ridiculous 15 dB boost through the same frequencies. Whenever it sounds worst, you've got it. Now just set it back to -3.
EQ VALUES
name boost/cut frequency Q
LS: 3.2 60 0.7
PK1: -2.0 220 0.7
PK2: -0.5 400 1.0
PK3: 1.8 1800 2.0
PK4: 3.0 3600 0.4
PK5: -3.0 9950 4.5
HS: 10.0 10000 0.7
Precut: -10.0 dB
CONCLUSION / LISTENING IMPRESSIONS
Realism! The veil has been lifted. Additional midrange and treble detail is immediately audible. Kick drums now have respectable slam, while bass lines are delivered with better tonal clarity and more consistent amplitude as the notes descend. Bass roll-off is now on par with my closed AKG K-55 headphones, with better bass resolution. Bass speed is still a tad slow for my taste, but really the 598 is not a fast headphone to begin with, so it blends quite well. Finally, the wool/mud is gone from the upper bass and low mids. Vocals sound far more present, open and natural. There is a pleasant hint of air on the top end that was completely absent before. Above 10kHz the 598's breakup modes are in effect, but luckily human hearing isn't very sensitive in this area, and a 10dB high shelf boost brings just the right amount of spaciousness without tizz. All in all, this is what I've been after. Headphones that sound more like reference loudspeakers. Yes.
I'm very interested in others' impressions of these settings!
Rockbox users:
https://www.dropbox.com/sh/5as9jxl5gi19h39/AAAX6x7p4jYkcnrgYql4WnZ1a?dl=0
Here you can download two different .cfg files for easy A/B testing. "TWEAK" has the EQ applied and "STOCK" is flat. Both files will reduce the overall volume of your device, because I have applied a precut to prevent clipping. (FYI, you must always use a precut with the same amplitude as the maximum EQ boost.) This way, rather than turning EQ on and off, and having to adjust the volume manually each time, you can simply toggle back and forth between the two .cfg files, which both have the same overall volume. This yields a true comparison.
Happy listening!
These EQ curves were designed to precisely flatten the 598's frequency response and deliver extremely accurate sound. I used Logic Pro's native parametric EQ plugin, Channel EQ, to set initial EQ values. These values were then entered into two Rockboxed DAPs for testing--an iPod Video 5.5 gen, and an xDuoo X3. I've listened to the settings for 100 hours or more, fine tuning by ear along the way, for the most natural sound possible.
If you use Rockbox on your player, you can download my EQ settings here and copy them onto your device.
https://www.dropbox.com/sh/5as9jxl5gi19h39/AAAX6x7p4jYkcnrgYql4WnZ1a?dl=0
You can also simply enter the values listed way down below into any good parametric EQ plugin on your computer.
Here's the lowdown:
The above picture shows HD 598 measurements as published by Innerfidelity.com and Headphone.com, with my compensatory EQ curve sandwiched in the middle. Big thanks to these two awesome websites for their very accurate measurements of the 598. You guys rock. As you can see, my EQ curve is essentially an inverse of the other two. My goal was to attain a razor-flat response, which yields a realism far beyond what I have ever heard from a headphone before.
STARTING POINT
The HD 598 is perhaps the best candidate for corrective EQ that I have come across. Spend some time on headphone.com comparing the 598's frequency response to other headphones, and you'll surely see why (click Learn -> Build a Graph). While the 598's natural response does not give as much HF and LF extension as some pricier models, it is exceptionally smooth overall, more so than models selling for five times the price. There are very few sharp peaks, which is a huge advantage when using EQ. Making sharp cuts and boosts in the signal using high Q values can cause audible phasing issues in my experience. EQing a headphone from the Grado lineup, for instance, which has sharp breakup modes throughout the treble region, is a nightmare. Owing to it's smooth response, EQing the 598 is a dream! Add to that, the 598 actually responds very well to a bass boost via EQ. Some lesser headphone drivers simply cannot push much air without distortion, and lose all composure when low bass is boosted. I was pleased to find that the 598's drivers are up to the task, and can indeed produce respectable low bass with impact. The 598 has amazing potential.
The 598 isn't perfect, however. It suffers from the following sonic deficiencies when played without EQ:
1. Poor low bass extension below 60 Hz. This results in barely any perceptible bass impact whatsoever.
2. Upper bass / low midrange bloat from 100 to 400 Hz. This buries the low bass in the mix even further, and makes the bass sound wooly and slow.
3. Rolled-off treble above 1000 Hz, a.k.a. the famous Sennheiser veil. This pushes vocals back in the mix and softens the overall presentation unnaturally.
4. A resonant mode around 9900 Hz. There's only one, and its not a bad one, but the peak is there, and it causes listening fatigue.
5. Drastically rolled off high treble starting at 10 kHz. Where's the air?
PARAMETRIC EQ
Well, I set out to correct all of these issues with a parametric EQ. Parametric is the key word here. A simple graphic EQ will not give you the control you need to sculpt the signal with precision, although it could possibly yield some benefits for some the issues listed above. Certainly not for number 4. With a parametric EQ, each boost and cut may be performed with a variable bandwidth, called Q. I spent a long while in Logic Pro sculpting the EQ curve to perfectly mirror the 598's measured response curves.
TUNING BY EAR
The original EQ values were loaded onto both of my Rockboxed DAPs for testing--an iPod Video 5.5 gen, and an xDuoo X3. I needed to determine a few things, including how much bass the 598 could handle from a low shelf filter, and the exact location of the driver resonance mentioned earlier. I played a huge variety of music through both devices while tweaking with the EQ bands.
To dial in the bass, I listened mostly to Random Access Memories by Daft Punk, Junto by Basement Jaxx, and Voodoo by D'Angelo. All of these albums go boom. In the end, I settled on a 3.2 dB Low Shelf boost, centered at 60 Hz. Hello bass impact! I find it impressive the that 598 can handle a boost this large without distortion. The bass is still a tad reserved, but it digs much deeper now.
To find the exact frequency of the resonant band, I listened to high-pitched vocals with a lot of ambience. The self-titled album by Pentatonix was helpful, along with So Familiar by Steve Martin and Edie Brickell. For me, the resonant band turned out to be at 9950 Hz. I notched it out with a narrow 3 dB cut. Here's the catch: I believe that this frequency varies from driver to driver. I noticed some discrepancy between the two response graphs I studied, and it looks like the pair of cans measured by Inner Fidelity actually varied between left and right sides! I recommend that anyone who tries these EQ settings play some music with some good treble content while sweeping the notch between 8500 and 10k. When it sounds most natural, you've got it. Alternatively, you can sweep with a ridiculous 15 dB boost through the same frequencies. Whenever it sounds worst, you've got it. Now just set it back to -3.
EQ VALUES
name boost/cut frequency Q
LS: 3.2 60 0.7
PK1: -2.0 220 0.7
PK2: -0.5 400 1.0
PK3: 1.8 1800 2.0
PK4: 3.0 3600 0.4
PK5: -3.0 9950 4.5
HS: 10.0 10000 0.7
Precut: -10.0 dB
CONCLUSION / LISTENING IMPRESSIONS
Realism! The veil has been lifted. Additional midrange and treble detail is immediately audible. Kick drums now have respectable slam, while bass lines are delivered with better tonal clarity and more consistent amplitude as the notes descend. Bass roll-off is now on par with my closed AKG K-55 headphones, with better bass resolution. Bass speed is still a tad slow for my taste, but really the 598 is not a fast headphone to begin with, so it blends quite well. Finally, the wool/mud is gone from the upper bass and low mids. Vocals sound far more present, open and natural. There is a pleasant hint of air on the top end that was completely absent before. Above 10kHz the 598's breakup modes are in effect, but luckily human hearing isn't very sensitive in this area, and a 10dB high shelf boost brings just the right amount of spaciousness without tizz. All in all, this is what I've been after. Headphones that sound more like reference loudspeakers. Yes.
I'm very interested in others' impressions of these settings!
Rockbox users:
https://www.dropbox.com/sh/5as9jxl5gi19h39/AAAX6x7p4jYkcnrgYql4WnZ1a?dl=0
Here you can download two different .cfg files for easy A/B testing. "TWEAK" has the EQ applied and "STOCK" is flat. Both files will reduce the overall volume of your device, because I have applied a precut to prevent clipping. (FYI, you must always use a precut with the same amplitude as the maximum EQ boost.) This way, rather than turning EQ on and off, and having to adjust the volume manually each time, you can simply toggle back and forth between the two .cfg files, which both have the same overall volume. This yields a true comparison.
Happy listening!
