This is a review, detailed measurements and equalization of Superlux HD668B semi-open headphone. It was kindly purchased by a member new and drop shipped to me. It costs as little as $40 including free shipping on Amazon.
Needless to say, you are looking at all plastic headphone here:
It looks like you can adjust the headband but you can't. Trust me, I tried at first.
the little pads on top are supposed to be a self-adjusting mechanism. They kind of work but then put pressure on two small spots on your head.
The cable terminates in a male connector which is unusual:
The cups are on the larger side which helps a lot with comfort given the stiff spring action. I would give it an 80 to 90% comfort after you wear them a few minutes.
Superlux HD668B Measurements
Here is the HD668B frequency response:
We see good adherence to our target curve from 70 Hz to 2.5 kHz which is good. But then we have a trough and two peaks. We also have our usual bass drop off. Subtracting the frequency response from our target we get this kind of deviation:
As you see, we have huge variation in span of just a few Kilohertz as noted on the graph. That first peak interestingly enough, correlates precisely with distortion at that frequency:
That's good news in that we need to take down that peak anyway and with it, we will be also reducing distortion as a bonus.
Unfortunately we are not helped with the high distortion in bass. Boosting that may be a compromise between enough sub-bass response and more distortion. More on this later in the equalization section.
Here is our log display of distortion showing what we already know:
Group delay is not very informative:
By now you should know not to try to fix any large dips in the high frequencies as they are results of two waves cancelling each other. Any EQ will change both resulting in the same cancellation. This shows up as large group delay as indicated by the spike I have marked.
Impedance vs frequency graph is a constant so no worry about the amplifier impedance being high (although you should avoid such amps anyway):
Let me know if you like this kind of non-zoomed display better. I think it is obscuring some info of interest in the graph but can be confusing to see it highly zoomed at first glance.
Sensitivity is good (lower, better):
Superlux HD668B Headphone Listening Test and Equalization
Quick listen showed somewhat exaggerated highs with grittiness added for good measure. At the same time, there was a bit of dullness which I know is an oxymoron in that context. Overall it was not a pleasant experience. So let's try to fix the issues:
Wow, what a transformation. If blindfolded you and put these on your ears, you would swear you were listening to very expensive headphones. The large drivers provide good bit of spatial information (separation of instruments), creating a soundstage that dangles from your ears down and to the back with the mono content behind your forehead. This added effect combined with now good tonality and much deeper bass provided a very enjoyable experience. Love the collaboration of Nick Cave and Nicholas Noorenbergh on this track for example:
Let me indicate though that the sound can get a bit boomy (likely due to harmonic distortion of sub-pass) and the highs persisted to be a bit bright at times especially on female vocals. A high frequency shelving filter would fix the latter but I wanted to stick to my philosophy of not piling on too many filters. As I indicated, what is there already provides excellent performance.
OK, I lied a bit. There was something bugging me still. When I played the track, Sever The Ego by Jewel:
I kept hearing her voice to the left of center. At first I ignored but then I realized same was happening in other tracks. I looked back at the frequency response and realized that one channel is lower level across the board. So I pulled up the Roon level controls and boosted the weak channel using pink noise:
This nicely anchored the mono aspects of the track in the center, removing the hanging nail effect of channels not being balanced.
Conclusions
We have yet another example of an out of box objective and subjective performance that is disappointing. You could try to forgive it given the cost but why? With equalization this headphone (as with most) can outperform speakers costing thousands of dollars in tonality! Fit and comfort is good enough and spatial effects very satisfying. Wonderful experience is being had for $40 and a bit of software and signal processing.
You can predict my summary then: no good as is but recommended with equalization. And a bargain.
Measurement data included, this time as .csv file. Let me know if you like it better.
Needless to say, you are looking at all plastic headphone here:
It looks like you can adjust the headband but you can't. Trust me, I tried at first.
the little pads on top are supposed to be a self-adjusting mechanism. They kind of work but then put pressure on two small spots on your head.The cable terminates in a male connector which is unusual:
The cups are on the larger side which helps a lot with comfort given the stiff spring action. I would give it an 80 to 90% comfort after you wear them a few minutes.
Superlux HD668B Measurements
Here is the HD668B frequency response:
We see good adherence to our target curve from 70 Hz to 2.5 kHz which is good. But then we have a trough and two peaks. We also have our usual bass drop off. Subtracting the frequency response from our target we get this kind of deviation:
As you see, we have huge variation in span of just a few Kilohertz as noted on the graph. That first peak interestingly enough, correlates precisely with distortion at that frequency:
That's good news in that we need to take down that peak anyway and with it, we will be also reducing distortion as a bonus.
Unfortunately we are not helped with the high distortion in bass. Boosting that may be a compromise between enough sub-bass response and more distortion. More on this later in the equalization section.
Here is our log display of distortion showing what we already know:
Group delay is not very informative:
By now you should know not to try to fix any large dips in the high frequencies as they are results of two waves cancelling each other. Any EQ will change both resulting in the same cancellation. This shows up as large group delay as indicated by the spike I have marked.
Impedance vs frequency graph is a constant so no worry about the amplifier impedance being high (although you should avoid such amps anyway):
Let me know if you like this kind of non-zoomed display better. I think it is obscuring some info of interest in the graph but can be confusing to see it highly zoomed at first glance.
Sensitivity is good (lower, better):
Superlux HD668B Headphone Listening Test and Equalization
Quick listen showed somewhat exaggerated highs with grittiness added for good measure. At the same time, there was a bit of dullness which I know is an oxymoron in that context. Overall it was not a pleasant experience. So let's try to fix the issues:
Wow, what a transformation. If blindfolded you and put these on your ears, you would swear you were listening to very expensive headphones. The large drivers provide good bit of spatial information (separation of instruments), creating a soundstage that dangles from your ears down and to the back with the mono content behind your forehead. This added effect combined with now good tonality and much deeper bass provided a very enjoyable experience. Love the collaboration of Nick Cave and Nicholas Noorenbergh on this track for example:
Let me indicate though that the sound can get a bit boomy (likely due to harmonic distortion of sub-pass) and the highs persisted to be a bit bright at times especially on female vocals. A high frequency shelving filter would fix the latter but I wanted to stick to my philosophy of not piling on too many filters. As I indicated, what is there already provides excellent performance.
OK, I lied a bit. There was something bugging me still. When I played the track, Sever The Ego by Jewel:
I kept hearing her voice to the left of center. At first I ignored but then I realized same was happening in other tracks. I looked back at the frequency response and realized that one channel is lower level across the board. So I pulled up the Roon level controls and boosted the weak channel using pink noise:
This nicely anchored the mono aspects of the track in the center, removing the hanging nail effect of channels not being balanced.
Conclusions
We have yet another example of an out of box objective and subjective performance that is disappointing. You could try to forgive it given the cost but why? With equalization this headphone (as with most) can outperform speakers costing thousands of dollars in tonality! Fit and comfort is good enough and spatial effects very satisfying. Wonderful experience is being had for $40 and a bit of software and signal processing.
You can predict my summary then: no good as is but recommended with equalization. And a bargain.
Measurement data included, this time as .csv file. Let me know if you like it better.
