Why do my headphones sound worse than my speakers?

[SimpleTheater]

I think the issue is the BOSE speakers, with this beautiful response by BOSE themselves to an amazon customer "Bose does not publish specifications such as output wattage or frequency response for our products as we do not feel they accurately describe how the products actually sound in the real world" SOURCE: https://www.amazon.com/ask/questions/TxA5JC0ILGKAP7/ref=ask_ql_ql_al_hza

The company does a lot of eq'ing their products, not to be accurate, but to be tuned to what people perceive as better sounding. When I was much younger I had a 32 band equalizer in my room, and ran some pink noise to get a flatter in-room response. Most of the bands were moved slightly, if at all. When my friends came over they would just crank the bass and highs (making a smiley face) and say "Now that sounds better". That isn't exactly what BOSE does, as their product seldom produce anything resembling bass, but they attenuate the frequencies you hear better.

 

[STC]

What I am saying is that Blumlein stereo *needs* the crosstalk in order to produce its time-of-arrival separation between the channels - and it is accurate in the sense of reproducing the correct angle. And then Gerzon's quote applies to that.

Without it (i.e. in headphones or crosstalk cancellation - if that worked fully), it would produce only a vague interaural level difference which is far less distinct than the time-of-arrival difference, and probably not even 'correct' in terms of its apparent angle of origin.

This may not be true for all recordings (where headphones or crosstalk cancellation may improve the separation), but it will be true for panpotted stereo or recordings that use genuine Blumlein pair mic techniques.

This might be what the OP is experiencing.

I am not sure of in what contexts the passage was attributed to Gerzon. We need the actual paper. I am aware of the experiment by Blauert using similar angle but let’s not speculate based on someone quoting Gerzon in Stereophile.

 

[andreasmaaan]

@Cosmik do you have a good source of information for the role of crosstalk in Blumlein pair mic'ing? Or perhaps you'd care to explain it in more detail? I'm curious to learn more.

 

[daftcombo]

Read this for a hint of Bose frequency response:
http://noaudiophile.com/Bose_901/

 

[Shadrach]

I was listening to one of my records, and I noticed it sounds way more detailed on my relatively cheap speakers (https://www.amazon.com/Bose-SoundTouch-wireless-speaker-works/dp/B011IH685E) than they do on my Sennheisers (https://www.amazon.com/Sennheiser-HD-599-Open-Headphone/dp/B01L1IICR2, which are plugged into this headphone amp https://www.amazon.com/Schiit-Magni-Headphone-Amp-Preamp/dp/B07KWJ9QMN). Specifically, there are horns on a particular build during a track that sound way more distinguished on my speakers, and I really had to watch out for them in order to hear them on my headphones. I'm curious if anything can be done to make my headphones sound as "clear" as my speakers are.

I'm using an AT-LP60X turntable (https://www.amazon.com/Audio-Technica-AT-LP60X-BK-Automatic-Belt-Drive-Turntable/dp/B07N3X7KPX), and I'm using its built-in preamp. The speakers are plugged into the turntable via a 3.5mm male to RCA female adapter cable.

Interesting. I prefer the sound of my speakers to that of both my headphones (Senn 595 and Beyer DT880)
I came to the conclusion, rightly or wrongly, that it has to do with the movement of air and I've become accustom to the added reflections and reverberations of my listening environment. Strangely, if I listen to a track on headphones, I suddenly hear what I believe to be new 'sounds' in a recording and naturally enough I believe the headphones are showing more detail. However, when I play that recording again through my loudspeakers that detail that I hadn't heard before becomes obvious.
Strange how the brain and ear work.

 

[Cosmik]

@Cosmik do you have a good source of information for the role of crosstalk in Blumlein pair mic'ing? Or perhaps you'd care to explain it in more detail? I'm curious to learn more.

I modelled it. I assumed that interaural delay would be measurable using cross-correlation at the ears i.e. by what amount you must slide one relative to the other to get the highest peak match between them - simply in the time domain. I may be wrong there, but that's what seemed like common sense.

And then if you create a simulation of sources, Blumlein microphones, speakers and ears and you look for the highest similarity between signals correlated at the ears from a real source, versus reproduced over speakers (with crosstalk), a remarkable result emerges: the speakers produce the correct direction for the source based on time-of-arrival. Even more remarkable: the listener can move and turn their head and the image remains stable. I didn't include head shadowing, reflections, ear profiles etc.; just the most basic model. And I didn't try with many types of sound; just plucked notes.

In a sense, I'm not interested much beyond that because I can't do anything more than create two speakers as neutral as possible. It may not be as perfect as my model in practice, but there's not much I can do about it.

But I was overjoyed to confirm that crosstalk from speakers a la Blumlein is not the defect that most people think it is, but is, in fact, most likely the key to that beautifully sharp, clear imaging that you get from speakers. So I won't be bothering with crosstalk elimination - that I might at one time have been intrigued by.

(And isn't stereo from speakers 'comfortable' to listen to? In a way that crosstalk elimination isn't?)

 

[MRC01]

...
But I was overjoyed to confirm that crosstalk from speakers a la Blumlein is not the defect that most people think it is, but is, in fact, most likely the key to that beautifully sharp, clear imaging that you get from speakers. So I won't be bothering with crosstalk elimination - that I might at one time have been intrigued by.
...

This stereo magic happens only with certain recordings that are miced in a certain way. With my own 2-mic recordings, I've tried both Blumlein and ORTF, and both have this same effect when played back on speakers. I find that interesting since ORTF is not coincident, but spaced about 6.5" apart, which is about the same spacing as most people's ears.
However, most acoustic music multi-miced for each individual instrument, so there is no natural image at all. It sounds OK, can have some depth and consistent spacing, but it is not the magic type of 3-D sounding image that 2-mic methods provide.

 

[Cosmik]

This stereo magic happens only with certain recordings that are miced in a certain way.

Or panpotted.

 

[andreasmaaan]

I modelled it. I assumed that interaural delay would be measurable using cross-correlation at the ears i.e. by what amount you must slide one relative to the other to get the highest peak match between them - simply in the time domain. I may be wrong there, but that's what seemed like common sense.

And then if you create a simulation of sources, Blumlein microphones, speakers and ears and you look for the highest similarity between signals correlated at the ears from a real source, versus reproduced over speakers (with crosstalk), a remarkable result emerges: the speakers produce the correct direction for the source based on time-of-arrival. Even more remarkable: the listener can move and turn their head and the image remains stable. I didn't include head shadowing, reflections, ear profiles etc.; just the most basic model. And I didn't try with many types of sound; just plucked notes.

In a sense, I'm not interested much beyond that because I can't do anything more than create two speakers as neutral as possible. It may not be as perfect as my model in practice, but there's not much I can do about it.

But I was overjoyed to confirm that crosstalk from speakers a la Blumlein is not the defect that most people think it is, but is, in fact, most likely the key to that beautifully sharp, clear imaging that you get from speakers. So I won't be bothering with crosstalk elimination - that I might at one time have been intrigued by.

(And isn't stereo from speakers 'comfortable' to listen to? In a way that crosstalk elimination isn't?)

That’s fascinating. Surely worth a research paper (or at least a post on your website)?

I’m still thinking it through here, but would the same not be true of pan-pot stereo? In an ideal Blumlein configuration, the mics are coincident, with only the mics’ polar response coupled with their 90-degree angle from each other creating the interchannel level difference that enables stereo reproduction to work. In other words, would you not have made the same finding using any recording that relied on interchannel level differences to produce a stereo image?

Or am I missing something?

 

[Cosmik]

That’s fascinating. Surely worth a research paper (or at least a post on your website)?

I’m still thinking it through here, but would the same not be true of pan-pot stereo? In an ideal Blumlein configuration, the mics are coincident, with only the mics’ polar response coupled with their 90-degree angle from each other creating the interchannel level difference that enables stereo reproduction to work. In other words, would you not have made the same finding using any recording that relied on interchannel level differences to produce a stereo image?

Or am I missing something?

I think you are correct (I did mention it earlier). Again, it's ironic. I always thought that the panpot was a pathetic 'blunt' way to produce stereo - but that it somehow worked better than expected. In fact, it's probably the 'correct' way to do it with speakers.

 

[MRC01]

Is the idea behind this, that the interaural delay will be reproduced by speakers on playback, so it need not be captured in the recording. And if the recording does capture it, then it's being artificially exaggerated when played back on speakers?
If so, then why does ORTF also create such a great stereo image?

 

[Cosmik]

Is the idea behind this, that the interaural delay will be reproduced by speakers on playback, so it need not be captured in the recording. And if the recording does capture it, then it's being artificially exaggerated when played back on speakers?

That's my take on it.

If so, then why does ORTF also create such a great stereo image?

I think that any microphone technique can produce a stereo image (it just needs to result in interaural timing differences at the ears), but it may not be 'correct' in the sense of reproducing the angles of the sources to the listener accurately. The purist Blumlein system appears to do that.

I also think that non-Blumlein mics may possibly make the image less stable with listener movement, head turning etc.

But if a particular technique results in something that sounds great, then why not? There are other aspects to recording that mean we can't simply plonk two mics down in the audience and expect it to sound perfect, anyway.

 

[daftcombo]

I have a CD called "Répons" by Pierre Boulez.
In the booklet, it is written that people who would want to listen to it with headphones should send en email and ask for a copy of the CD version made for headphones. Perhaps the latter is a binaural recording.

I don't know if there are other examples of releases made in two versions: one for speakers, one for headphones.

Sadly, I sent a message a few years after my purchase and they replied me that the headphones version CD was out of stock.

 

[Blumlein 88]

Is the idea behind this, that the interaural delay will be reproduced by speakers on playback, so it need not be captured in the recording. And if the recording does capture it, then it's being artificially exaggerated when played back on speakers?
If so, then why does ORTF also create such a great stereo image?

One thing using ORTF or similar does is enlarge the sweet spot where stereo works pretty well. The reason it doesn't mess it up is our hearing can work okay with channel differences about twice what our ear spacing produces. Why? Who knows. So the effect Cosmik is describing creates an interchannel time difference equal to the width of our ears. You can use a microphone spacing like ORTF for about that much more and our hearing still works okay with it. Plus it gives you an enlarged space to move your head without corrupting optimum results.

http://www.richardbrice.net/blumlein_patent.htm

You can read the old patents Alan Blumlein submitted. All of it is described there though maybe not so clearly.

 

[iopgh]

I want to elaborate on exactly what I meant in the OP, since my terminology admittedly isn't the strongest yet.

When I said that my speakers sound more detailed, I meant that I can hear sounds that are otherwise nearly imperceivable on my headphones. The horns that I referenced are present in my headphones, but they're "muddied" to the extent that I wouldn't notice them were I not specifically paying attention for them, whereas I can hear those sounds very clearly on my speakers.

If you're running Windows, I can make an EQ profile for you in Equalizer APO & Peace GUI.

I appreciate the gesture, but atm I don't have a means of connecting my audio setup to my PC.

What happens across a wide array of different recordings?

The case that I described applies to my records generally, in that certain noises are very subtle on my headphones but are easily noticeable on my speakers. I'd have to listen further and make more comparisons to go into greater detail and to offer more examples

What you observed doesn't sound reasonable, all else being equal, in that assuming all things are functioning correctly.

This occurred to me as well, so I felt it necessary to describe my problems to those with better intuition in case something obvious was wrong.

 

[solderdude]

The HD599 is 'darker' than 'neutral' and sounds a bit 'muddy' compared to better headphones.
One could say the mids and treble are somewhat subdued compared to the bass and lower mids.

I have no idea how the FR of the speaker is nor how it sounds nor how and where it is placed in a room.

Grab any other headphone and speaker and the sound differs yet again.
A fact of life... speakers sound different and headphones sound different.

 

[Blumlein 88]

I want to elaborate on exactly what I meant in the OP, since my terminology admittedly isn't the strongest yet.

When I said that my speakers sound more detailed, I meant that I can hear sounds that are otherwise nearly imperceivable on my headphones. The horns that I referenced are present in my headphones, but they're "muddied" to the extent that I wouldn't notice them were I not specifically paying attention for them, whereas I can hear those sounds very clearly on my speakers.

I appreciate the gesture, but atm I don't have a means of connecting my audio setup to my PC.

The case that I described applies to my records generally, in that certain noises are very subtle on my headphones but are easily noticeable on my speakers. I'd have to listen further and make more comparisons to go into greater detail and to offer more examples
This occurred to me as well, so I felt it necessary to describe my problems to those with better intuition in case something obvious was wrong.

My apologies for going off on a tangent. Something not uncommon on forums.

If it is just those horns, or only highly noticeable on horns it could be one of two or both things. Maybe your speakers are too hot in the frequency of the horns accentuating them more than should be the case in that recording. Or maybe the headphones have a dip where the frequency of the horns lie making the horns hard to hear. Or it could be a little of both.

Most of the time, you'd expect details like that to be heard more easily on headphones even if your speakers image better and are better in most ways. Mainly because the headphones take out most of the sound of your room. Headphones are a little bit like a sonic microscope for such little details or dimly heard sounds even if they lack general fidelity in other ways.

So are your speakers radically different sounding much lighter and brighter while the headphones sound heavy, muddy, with little treble? That is possible. I don't know either of your pieces of gear. Try your headphone with something else. Even a cellphone just to see if they sound about the same as they do on your headphone amp. If you know someone with different phones, even cheap earbuds, try those and see if the horns are heard better.

 

[STC]

I modelled it. I assumed that interaural delay would be measurable using cross-correlation at the ears i.e. by what amount you must slide one relative to the other to get the highest peak match between them - simply in the time domain. I may be wrong there, but that's what seemed like common sense.

Please explain more. Why are you measuring interaural delay when it can be calculated?

 

[Cosmik]

Please explain more. Why are you measuring interaural delay when it can be calculated?

Remember that with Blumlein stereo there is no delay between the channels. If there is an effective interaural delay, it has been produced by summing the two (or more if I were to expand the model) speakers at each ear. Hence the idea of assessing similarity to 'reality' by correlation.

I don't directly measure or calculate interaural delay; I look for the similarity between correlation peak with speakers compared to reality.

 

[urfaust]

Im sure many elaborated answers here are valid, but it could be a simplier explanation.

-Your room, maybe has suprisingly good "room mode", like adding pleasing reverbs to a mix.
-Your speakers are emphasizing different frequencies, that you deem as more details contrary to the response of your headphones.
-Levels most of the time trump everything, you can feel your speakers in your chest.

(Now set them at very low levels and listen at 10cm, see how it still sounds awesome and more detailed, or not)