Sorry, I already lost the original post of this when I accidentally pressed backspace, so it's just the basic points this time...

I realised that the only time I've ever been 'convinced' by reverb algorithms and IRs is through headphones. Binaural recordings can also be scarily convincing.

So why do we tend to compromise the mix through headphones and favour a loudspeaker mix, especially in a time when more and more people are listening through earphones?

In fact, aren't headphones the modern equivalent of the 50's/60's in-car radio speaker?

Certainly, I intend to offer dedicated headphone and loudspeaker mixes (and perhaps even surround mixes too) but I suspect the ideal experience (in terms of realism) would involve headphones and close loudspeakers for low frequencies and so that we feel the sound vibration also.

Also, I'm not much of a 'communal' listener or watcher of films -I'm happy to do both on my own most of the time and really get into it- so it seems silly to go to extremes with surround sound speaker systems when a binaural headphone track could offer much more realism* -if offered!

What are your thoughts on this?

Cheers,

José

(*P.S. I also suspect that the future of 3D film may involve direct projection to the retina of ideally calibrated (via some computer modelling) images rather than trying to separate a single image from a screen. How can a single screen ever offer a 100% convincing 3D experience?)

Combining headphones and speakers is fraught with phase problems. Just wander about a room with speakers and cans running and you will find spots where the bass dips to nothing!

Headphone or "binaural" listening is a quite different mechanism from even very close free field loudspeakers.

I understand that many have struggled trying to unify the math for the two experiences (I understand the struggle NOT the math!) Google M.Gerzon. It would seem that the unified field theory is a doddle by comparison! There are of course a few converter systems now on the market but opinions seem very undecided as yet as to their usefulness?

Just my 2pee'oth!

Dave.

I work 95% of the time in headphones and complete my mixes there. The only time I use speakers is to check that my mix translates well to what I have available - the car stereo is as good as anything TBH.

(Though I do have my old method of feeling the speaker cone to help me evaluate the balance of low-frequency content!)

However you monitor, as long as you know what's going on you can make a mix that translates. This isn't about 'more realism', but simply knowing your monitoring to a point where you can trust both it and your perception. I have never made dedicated speaker/headphone mixes - a good mix is a good mix whatever you listen on!

--------------------
An Eagle for an Emperor, A Kestrel for a Knave.

Quote J.A.S:

---

I realised that the only time I've ever been 'convinced' by reverb algorithms and IRs is through headphones. Binaural recordings can also be scarily convincing.

---

The reason is the same in both cases. Using headphones removes the physical listening room's acoustics from the equation. When listening via speakers, the real or virtual acoustics are effectively layered on top of those of the listening room, and the brain is clever enough to figure that out. Listening via speakers in an anechoic chamber is scarily convincing because there are no room acoustics! Been there, done that!

Quote:

---

So why do we tend to compromise the mix through headphones and favour a loudspeaker mix, especially in a time when more and more people are listening through earphones?

---

Historically, the majority of listening was done via speakers and so optimising the mix for that format was the logical norm. These days, for some significant audience groups, headphone listening has become more dominant, and for that audience it makes sense to offer a mix optimised for them.... but ensuring some reasonable compatibility between the two is still common sense.

Quote:

---

I suspect the ideal experience (in terms of realism) would involve headphones and close loudspeakers for low frequencies and so that we feel the sound vibration also.

---

As Dave has said, this is fraught with technical problems, not to mention the annoyance of others in the vicinity who aren't sharing the headphone music! You can get chairs with 'but-kicker's built in to provide some visceral LF sensations for headphone users. But that's not much help for people wanting to listen on the move!

Quote:

---

... it seems silly to go to extremes with surround sound speaker systems when a binaural headphone track could offer much more realism

---

The idea of 'more realism' is difficult to quantify. With physical speakers the ambiguity of location can be minimised very easily, and precisely placed spatial sounds can be used to gerat dramatic effect. The trouble with binaural systems is that there is a huge amount of ambiguity in the perceived directions of sound, dependent the different physiologies of different listeners, as well as their expectations. Binaural headphone systems that incorporate motion sensors and can process the sound field to maintain a sense of a stable sound stage work considerably better... but they are very expensive and, again, impractical for 'on the move' applications.

hugh

--------------------
Technical Editor, Sound On Sound

Quote The Elf:

---

I do have my old method of feeling the speaker cone to help me evaluate the balance of low-frequency content!

---

Most monitor speakers dispense with 'speaker cloth' in favour of bare drive units, so you can also keep an eye on large bass excursions by eye as well as feel

Quote:

---

I have never made dedicated speaker/headphone mixes - a good mix is a good mix whatever you listen on!

---

I totally agree with this approach - if a mix doesn't work well on both headphones and loudspeakers then it's not a good mix


Martin

--------------------
YewTreeMagic

Thanks for your thoughts. I'm not sure I agree with...

"if a mix doesn't work well on both headphones and loudspeakers then it's not a good mix"

...though!

Reverb, especially, can sound over-the-top through headphones yet perfectly fine through speakers. Any compromise between headphones and speakers would inevitably degrade one or the other, or both to some extent. Also I've noticed that some (depth and stereo) spatial effects designed for speakers I've used sound strange or ineffective through headphones and vice-versa.

I can't see what would be wrong (in a downloadable music culture) to offer different mixes for both, that take advantage of the different listening experiences. (I don't get why 24-bit versions aren't also bundled in for the hell of it!)

Quote J.A.S:

---

Any compromise between headphones and speakers would inevitably degrade one or the other

---

I would agree with the previous claim. A 'good' mix does work acceptably well on both speakers and headphones -- and it must because you can't prescribe what format your listeners will choose to use. However, it is also reasonable to optimise the mix for one format if you decide to favour one part of your potential audience, accepting that doing so may also compromise slightly the mix when auditioned on the other format.

Quote:

---

Also I've noticed that some (depth and stereo) spatial effects designed for speakers I've used sound strange or ineffective through headphones and vice-versa.

---

This is inevitable, because time (and phase) differences between channnels and comb filtering effects can provide powerful spatial clues on headphones, but are less effective on speakers, while level differences tend to be more effective on speakers than in headphones. Again, it's an inherent aspect of headphone vs speaker listening, and is due to the fact that both ears hear both speakers, but each ear only hears its own headphone!

Quote:

---

I can't see what would be wrong (in a downloadable music culture) to offer different mixes for both, that take advantage of the different listening experiences.

---

None at all. Although as a commercial venture you'd need to justify the additional time spent optimising the mix for each audience with higher sales returns, and I'm not sure you could...

Quote:

---

I don't get why 24-bit versions aren't also bundled in for the hell of it!

---

Because there is absolutely no point. the vast majority of consumers do not have suitable listening environments (or interest) that would be able to distinguish between 24 and 16 bit files. Hell... with most material and in most locations they'd be hard pressed to tell the difference with 14 or 12 bit files!

Hugh

--------------------
Technical Editor, Sound On Sound

RE: "the vast majority of consumers do not have suitable listening environments (or interest) that would be able to distinguish between 24 and 16 bit files. Hell... with most material and in most locations they'd be hard pressed to tell the difference with 14 or 12 bit files!"

Classical music listeners tend to be the worst in my experience. Never that surprised to see an Alba stereo system, but I think that's a generational thing as much as anything.

I suppose I do care more about what other music enthusiasts think listening in 'closer to ideal' conditions. I suspect a lot of differences of opinion are really about what is practical for consumers verses what is ideal for the music.

One of the problems of "mixing for headphones" would I should think be people like me.
I am a bit of a lost cause now but ten years ago my hearing was fairly good but I still had at least 20dB loss past 2kHz in my right ear. Now my brain (yes even my tiny one!) sorted that out for "real world" sound and stereo because it creeps up on you and you adapt but slap cans on and all goes to spacial hell in a handcart!

24bits for domesticals?? Do the numbers and 1kW per channel assuming 90dB/W/mtr speakers goes nowhere!

Dave.

Quote J.A.S:

---

I suppose I do care more about what other music enthusiasts think listening in 'closer to ideal' conditions.

---

Maybe... but are you really telling me that your 'closer to ideal' listening conditions and monitor speakers enable to you achieve in excess of 120dB dynamic range? I'd be utterly amazed if that were true. Typical ambient noise floor in a quiet domestic room will be about 20-30dB SPL. So if you want the noise floor of a practical 24-bit D-A to sit at the same level as the room's noise floor (so that you can hear the dying embers of your digital dynamic range), then your monitors will need to be able to generate 140-150dB SPL at the listening position. That just ain't gonna happen with any conventional 'monitor speaker' let alone a compact budget two-way speaker!

It's hard enough to achieve realistic 90dB dynamic range in most semi-pro situations, which makes 16 bit replay systems more than adequate since the ambient room noise completely overwhelms the digital system noise. Sony got that bit completely right when they conceived the CD format!

Added to which, there are precious few recordings that have anything like 90dB dynamic range -- the ambient noise of the recording dominating the digital system noise.

In other words, 24 bits for consumer replay is a total waste of time and effort, and it provides no pracical or aesthetic advantages whatsoever.

24 bit working has a valid place in recording and production, because it affords an essential headroom margin. But a finished mix requires no headroom margin -- as the current CD market ably demonstrates -- in which case the replay wordlength needs only to be sufficient to ensure the digital system's signal-noise ratio is comfortably lower than that of the recording itself.

Hence 16, 14, 12 or even fewer bits, depending on the material are entirely adequate. ASk any lossy-codec designer I've played genuine 3-bit recordings of music to students before now and thay have been unable to tell the difference (or be affected by it) from the original 16 bit CD version!

Quote:

---

I suspect a lot of differences of opinion are really about what is practical for consumers verses what is ideal for the music.

---

Perhaps... but I suspect it's more to do with a lack of undertanding of the science, combined with daft number envy!

Hugh

--------------------
Technical Editor, Sound On Sound

Back when vinyl was king people with high quality systems would set the listening level such that the surface/tape noise was just perceptable when close to the kit and thus just inaudible at the listening position. This gave the best dynamic range possible in any given envoirenment (in many cases though, even here the peak levels would be too loud for 'er indoors, even in the kitchen!).

When CD came in people still cranked the VC seeking a noise of some sort and the result was some very angry comments from the distaff side! And this was at a time when 20-30 watts per channel was considered fairly top of the range. Only mad people had in excess of 50 watts!

Dave.

My personal reasons for wanting to offer separate headphone mix/es are really because I want to offer binaural (or 'stereo surround') versions that actually might differ quite a lot from the prefered loudspeaker mix. Also, there are differences in reverb and spatial effects when listening through headphones compared to loudspeakers, but it depends how adamant you are about keeping these in some exact way. I don't think these points have been negated, except for the practical/commercial concerns that are fair enough but not such a problem for me offering downloads from my website. (My main objection would have been the likelihood of people listening to the wrong version, but there you go.)

As for the bit-depth thing, I agree that the improvement in audible quality between a 16-bit and the higher bit-depth versions (of a finalised mix) is probably nil. I'm not always sure what somebody can recognise ('tell') always matches what people are actually affected by, but in this case, I reckon most people would be more affected by knowing that it's in 24-bit than any actual added audible advantage. Still, iTunes don't seem to mind.

But my thoughts were more about how DVDs come in a higher bit-depths, so people might prefer it when using your music in a film production. Would a higher bit-depth have an advantage for any effects/processing that might be necessary for the soundtrack? Also, do DVDs offer 20-bit and 24-bit perhaps because of the different room/ambience amp settings that are available?

(A 24-bit master is also supposed to be preferable for making vinyl prints, but that doesn't really make it necessary to include it in a general release. That said, I've heard there are enthusiasts who like to make prints (compilations?) of other people's stuff.)

Quote:

---

ASk any lossy-codec designer I've played genuine 3-bit recordings of music to students before now and thay have been unable to tell the difference (or be affected by it) from the original 16 bit CD version!

---

I know that the first ever recording (phonautograph) was released in MP3, and people were asking for higher quality! They (First Sounds?) said it would offer no improvement.

Quote J.A.S:

---

But my thoughts were more about how DVDs come in a higher bit-depths, so people might prefer it when using your music in a film production. Would a higher bit-depth have an advantage for any effects/processing that might be necessary for the soundtrack? Also, do DVDs offer 20-bit and 24-bit perhaps because of the different room/ambience amp settings that are available?

---

DVDs were invented to provide a seamless route from the film industry to the home viewer, omutting most of the remixing and repurposing that had previously been required in the days of the VHS. Dolby Digital and DTS were employed as the sound carriers so that the film soundtrack didn't have to be remixed for the reduced dynamic range capability of the domestic audience. Instead, metadata embedded within the sound track would automatically decrease the dynamic range and provide downmixing according to the wishes of the production company.

As DD and DTS are data-reduced formats, they don't strictly adhere to the 16/20/24 bit concept because they don't use linear PCM coding. The same is true of MPEG encoded material. However, the foramt was also specified to accept linera PCM and so it made sense to offer 20 and 24 bit support for professional and semi-pro applications.

Significantly, the film industry creates mixes with headroom (unlike the music industry). The default average dialogue level is -27dBFS, leaving well over 20dB of headroom for the explosions. In a cinema context, where the potential peak level could well be 120db SPL, it is necessary therefore necessary to work with a wide dynamic range format (20-24 bits) to ensure the system noise floor is close to the acoustic noise floor.

Quote:

---

A 24-bit master is also supposed to be preferable for making vinyl prints

---

Yes, but only because the mastering engineer wants his source material to retain the original mix headroom margin so that he has space to work!

Hugh

--------------------
Technical Editor, Sound On Sound

And vitually all DTVs and DVD players* have compression options in their setup menues.

How nice it would be if we could have full dynamic range on all CDs and the same option in amps and players for ICE for example.

*And even better would be a stonking great 30dB "ducker" that sensed adverts!

Dave.

Thanks for the responses.

RE: "And vitually all DTVs and DVD players* have compression options in their setup menues. How nice it would be if we could have full dynamic range on all CDs and the same option in amps and players..."

This is what I meant ("...perhaps because of the different room/ambience amp settings that are available?"). I wondered if a higher (effective) resolution might benefit this post production effects/processing.

RE: Significantly, the film industry creates mixes with headroom (unlike the music industry).

A lot of the stuff I'm trying to do has properties of a cinematic/soundtrack ('field recording' style) and my hope is that animators or film makers might take an interest in some of it. Sounds may need to be added to the 'background' layer (for any story requirements). It's difficult to describe in short.

RE: "Yes, but only because the mastering engineer wants his source material to retain the original mix headroom margin so that he has space to work!"

So if people were making prints of compilations (of other people's) music it would be better to use a 24-bit file then.

Cheers

Quote J.A.S:

---

So if people were making prints of compilations (of other people's) music it would be better to use a 24-bit file then.

---

It would be sensible to compile the 16 bit source tracks into a 24 bit master to provide plenty of headroom and an insignificant noise floor within which the source material can then be rebalanced and equalised to make them sound like they belong together -- all the things a mastering engineer normally does -- and then produce a 16 master of the final compilation for the end consumer. But there would be little if any practical advantage in starting with 24 bit source files.

hugh

--------------------
Technical Editor, Sound On Sound

Quote J.A.S:

---

My personal reasons for wanting to offer separate headphone mix/es are really because I want to offer binaural (or 'stereo surround') versions that actually might differ quite a lot from the prefered loudspeaker mix. Also, there are differences in reverb and spatial effects when listening through headphones compared to loudspeakers, but it depends how adamant you are about keeping these in some exact way. I don't think these points have been negated, except for the practical/commercial concerns that are fair enough but not such a problem for me offering downloads from my website. (My main objection would have been the likelihood of people listening to the wrong version, but there you go.)

---

That's a fair enough approach for your own downloads, although I suspect even then offering binaural mixes for headphone users and a different mix for loudspeaker users is going to confuse a lot of people

However, for commercial releases I stand by my previous comment that a good mix is one that works well (albeit slightly differently) on both loudspeakers and headphones. It doesn't take much effort to avoid extreme pan setting that make some headphone listeners feel sick, without compromising the mix when listened through loudspeakers


Martin

--------------------
YewTreeMagic