[71 dB]

I'm with you, Dawg!
Crossfeed incoming! in 3... 2... 1...

I tried to talk about crossfeed and the thread got locked. I'm also frustrated about the rules on politics and religion over here. That's why I'm not very active anymore. Why bother?

 

[bigshot]

For my part, I'd also like to see better digital EQs in our phones and computers. The EQ in my phone is terrible, not worth using.

I agree. I think the future of good sound is signal processing.

 

[Whazzzup]

How can headphone audio really improve? What hasn’t ? It either will or be replaced.

 

[castleofargh]

having them not sound like headphones. I know a few people who really seem to enjoy the sound of headphones more than speakers, but within the people I know, it's fair to say they're an extreme minority. I'd personally take average speakers over the best headphone gears anytime if it wasn't for external factors like mobility or neighbors with brooms(who admittedly make a compelling case for headphones at night). don't get me wrong, I'd be miserable without headphones and IEMs, but they feel most unnatural with typical albums. and yes crossfeed can help a little for that.
but as I said, something good requires DSP and extensive customization(through options that people won't know how to set up, or through measuring the HRTF of the user and having the DSPs use that instead of standards approaches).

ultimately headphones can become superior to speakers, but not while we play the wrong stereo signal through them and don't care about the listener's body.

 

[sine_wave]

Are there any companies actively researching this kind of digital signal processing? Harmon? Sony?

 

[castleofargh]

since the starting trend of virtual googles, many people have been working on those stuff. but I don't imagine many aim at full customization, because that's a bother and not where money resides.
we have seen various work in progress at shows over the years, a few headphone companies have made partial attempts on some Bluetooth headphones. a few try to measure something about our ears(with sound and mics or pictures), some have tried to associate stuff with head tracking, I remember one checking the distance between the cups while on the listener's head to help with the head tracking model. there has been a few kickstarter projects too, the Ossic X doesn't look like it's going anywhere, I'm still waiting for the Realizer A16 to come out, but there has been several projects with various ideas over the last years. you have a few related products linked in this post https://www.head-fi.org/threads/hearing-test-based-equalization.873077/ I remember the AKG N90Q being discussed a lot when it came out(the price isn't a joke), and other stuff but I don't recall most of the names.

if you want to torture your brain a little about some technologies and the principles behind them, you can look at this post https://www.head-fi.org/threads/a-l...-minded-immersive-audio-and-holophony.865973/ where @jgazal tries to bring up anything relevant to 3D sound. which we don't necessarily need for stereo albums and headphones, but the basics about HRTF and how the headphone bypasses part of the body, are relevant for us.

so yes I'd say it's a fairly active domain, even though gaming is probably the main driving force as audiophiles are still for the most part thinking that digital processing equals witchcraft and is "bad for sound". hopefully we'll all wake up from that stupid myth soon.

 

[bigshot]

The thing that has held back multichannel music is the lack of a one-size-fits-all solution. People don't like variables. They like a killer product that they can just buy off the shelf and take home and play with right away. I suspect the need for head measurements and mapping the tubes in our ears will hobble virtual reality sound just like the need for glasses has hobbled 3D.

 

[buzzy]

Maybe not what you're looking for, but a really good wireless connection and the audio chain to support that (standards, hardware), with wide availability rather than just in niche products.

I recently got some decent Bluetooth headphones. They support AAC, and do in fact sound better than just about any of the other Bluetooth devices I've tried. But when I'm paying attention I can easily hear ways in which the audio isn't quite up to par.

But what these Bluetooth headphones did teach me is

1) sometimes it's nice to be able to move around with decent headphones on; the ability to move more than offsets the audio quality difference, especially when I'm not giving the music my full attention; and

2) when cord noise is factored into the experience, some of my wired headphones are not necessarily more enjoyable than the Bluetooth headphones​

The first one isn't at all surprising. The second one was surprising to me.

 

[colonelkernel8]

Maybe not what you're looking for, but a really good wireless connection and the audio chain to support that (standards, hardware), with wide availability rather than just in niche products.

I recently got some decent Bluetooth headphones. They support AAC, and do in fact sound better than just about any of the other Bluetooth devices I've tried. But when I'm paying attention I can easily hear ways in which the audio isn't quite up to par.

But what these Bluetooth headphones did teach me is

1) sometimes it's nice to be able to move around with decent headphones on; the ability to move more than offsets the audio quality difference, especially when I'm not giving the music my full attention; and

2) when cord noise is factored into the experience, some of my wired headphones are not necessarily more enjoyable than the Bluetooth headphones​

The first one isn't at all surprising. The second one was surprising to me.

I really think we're close to this. LDAC already does it and with Apple's strong push towards wireless, it's only a matter of time until we're getting essentially an uncompressed audio signal wirelessly into portable, battery powered devices. With chips like THX's AAA amplifier, we can expect the amps in these wireless headphones to get smaller, quieter, measurably better, and use less power too.

 

[Zapp_Fan]

As a "member of the trade", here's some major technological changes / progress that would improve headphones and IMO will take years to become common:

1. Universal lossless audio over Bluetooth (BT5.1 fingers crossed)
2. Breakup modes of drivers all outside of audible range (e.g. driver cones are commonly made with beryllium or other exotic materials)
3. IMD and THD reduced to inaudible levels
4. Sensitivity of voice coils / planar drivers increased so that impulse response (dynamics) is inaudibly distorted (i.e. extremely accurate) even at low wattages
5. Unwanted resonance / ringing reduced to inaudible levels
6. Noise isolation / cancellation improved to >40dB across entire audible frequency range (may be physically impossible)

Needless to say, none of these have been achieved in the < $100 range, so there is still a great deal of work to be done before headphones are a solved problem.

 

[colonelkernel8]

As a "member of the trade", here's some major technological changes / progress that would improve headphones and IMO will take years to become common:

1. Universal lossless audio over Bluetooth (BT5.1 fingers crossed)
2. Breakup modes of drivers all outside of audible range (e.g. driver cones are commonly made with beryllium or other exotic materials)
3. IMD and THD reduced to inaudible levels
4. Sensitivity of voice coils / planar drivers increased so that impulse response (dynamics) is inaudibly distorted (i.e. extremely accurate) even at low wattages
5. Unwanted resonance / ringing reduced to inaudible levels
6. Noise isolation / cancellation improved to >40dB across entire audible frequency range (may be physically impossible)

Needless to say, none of these have been achieved in the < $100 range, so there is still a great deal of work to be done before headphones are a solved problem.

What about a closed loop feedback mechanism on the driver itself so that the amplifier is controlling the motion of the "cone" and not merely the current through the voice coil? I think some higher end subwoofers do this because it's easier at low frequencies, but I bet new DSPs are fast enough to do this on the full audible range.

 

[castleofargh]

As a "member of the trade", here's some major technological changes / progress that would improve headphones and IMO will take years to become common:

1. Universal lossless audio over Bluetooth (BT5.1 fingers crossed)
2. Breakup modes of drivers all outside of audible range (e.g. driver cones are commonly made with beryllium or other exotic materials)
3. IMD and THD reduced to inaudible levels
4. Sensitivity of voice coils / planar drivers increased so that impulse response (dynamics) is inaudibly distorted (i.e. extremely accurate) even at low wattages
5. Unwanted resonance / ringing reduced to inaudible levels
6. Noise isolation / cancellation improved to >40dB across entire audible frequency range (may be physically impossible)

Needless to say, none of these have been achieved in the < $100 range, so there is still a great deal of work to be done before headphones are a solved problem.

some of those wishes are more realistic than others. but if we can go a little wild, while within the realm of actually doable, I'd like to start raising clones of great singers so that we can force them to sing new albums when the guy dies. I believe it would in the long run improve headphone listening a lot.

 

[Zapp_Fan]

some of those wishes are more realistic than others. but if we can go a little wild, while within the realm of actually doable, I'd like to start raising clones of great singers so that we can force them to sing new albums when the guy dies. I believe it would in the long run improve headphone listening a lot.

You laugh, but it's now possible to do voice resynthesis based on past recordings of voices. So this is actually probably not far off. Basically a combination of Vocaloid and Adobe's "photoshop for audio" will bring this to bear in the next 10 years. Autotune 2020.

 

[bigshot]

Uncanny valley

Maybe we could reprocess all the Beatles albums so instead of the singing being done by John and Paul, it's done by Yoko Ono and Linda McCartney.

 

[jaakkopasanen]

I think castleofargh nailed it before the jitter debate. Digital components in our signal paths are right up there where they need to be, transducers are orders of magnitude worse. We won't get rid of problems with transducers before get rid of the transducers themselves. Feeding audio signal directly into the auditory nerve is the endgame and hopefully it's not as sci-fi as one might think. There is already muscle nerve stimulations done in academia but of course audio signal is a quite bit more complex. We'll get there eventually, don't worry.

Bigger problems with headphones are soundstage and imaging. Nobody in their right mind would be able to convince themselves that they actually believe sounds coming out of headphones are coming from a real world. Crossfeed has been around for a long time now but is quite limited in it's capabilities. Better alternatives are the convolution based approaches where speakers are simulated with headphones. Smyth Realiser and Fong audio's Out of your Head are examples of this technology. Problem with these is the requirement for high level of personalisation. Humans have only two ears and that is not enough to localize sounds in 3d space with just delay and phase shifts. Brains guess the sound direction based on transformations made to the sound by torso, head and ears along with phase, delay and amplitude differences in both ears. Unfortunately everyone's torso, head and ears change the sounds differently meaning that the impulse responses used in convolution approaches need to be tailored for each listener individually. This is not so much a technical problem as it is a practical and a business problem. It's a lot of work trying to get people visit a dedicated laboratory for these measurements. I guess headphone stores could have a listening room and a measurement rig for this purpose...

I think it might be possible to get quite good results by generating the personal impulse responses with recurrent neural networks based on a couple of minutes of comparisons of existing impulse responses. I'm working on this project on my free time. Biggest obstacle is gathering enough data to train the networkt.

There is also different kind of approach studied in Finland. IDA audio is developing a method to generate complete head related transfer function (HRTF) from a 3D scan of the listener's head and ears. They are even looking into making said 3D scans with a smartphone camera. We will have to see if they solve personal audio for good or if it turns out to be just a dud.