Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards Binaural processing with headphones/monitors on DPs

The need for using binaural correction is obvious, unprocessed stereo output cannot be equally optimal for headphone and output through stereo boxes - Binaural DSP:

There are such external solutions (I use Reapers AddOns), with Pianoteq comes one (with dramatic effect), but at first place I would like to know whether DPs or at least some of them do some binaural sound enhancements when output over headphones (or their line output is practically identical to their headphone outputs, beside of some electronical differences like impedance).

Another question is how much directional studio monitors are needing binaural corrections?

Have we identified another big area for instant and dramatic improvements by the manufacturers - or is it a widely settled issue (or a non-issue in Your opinion)?

Different EQ curves and dynamics processing for headphone output is pretty common, I believe. I would be shocked to learn that DP manufacturers actually bother spending money on binaural stereo enhancement for DP headphone outs.

The limited amount of reading I've done on this subject suggests that binaural enhancement solutions for headphones can create a pleasant effect, falling short of a dramatic improvement.

At least one manufacturer of desktop audio interfaces implemented such processing for headphone outputs--it was basically intended to assist people using headphones for mixing to create mixes more compatible with open-field monitors. I believe it was Lexicon or Focusrite, can't remember which. I suppose if you really wanted to experiment with this idea, you could dig up one of those units and listen to your DP through it to see what effect the extra processing might have.

Yes, that "enhancement" is basically another audio gimmick. True binaural recording however is rather impressive. It requires recording with microphones in a dummy head positioned in a typical listening position in the actual venue of a live performance. I heard special demos many years ago that were quite convincing compared to listening to a conventional recording made at the same time and played back in a typical living room setting. No doubt the same technique could be used to sample pianos at the player's head position in typical playing venues. Not likely any piano sampling company would do that however.

I have been using VLC player for playing good quality FLAC music and found the binaural filter useful to have a comfortable background music in the background. (It's tamed at first, not dramatic.)

But it gets dramatic in Pianoteq when I go back from headphones to my monitors have to turn binaural off (to switch to stereo or recording perspective) otherwise the music quality suffers very much.

My point is, sound output cannot be equally good over headphones and boxes, they need other number of shoes.

I think also, this is an underused potential, not very popular because close recorded samples can be mixed and processed better along other components of an instrument (convolution reverb, etc.)

On their UK and Ireland website, Yamaha has a new SILENT piano called SH Silent Piano.

The webpage says that it uses binaural sampling as opposed to stereo sampling.

The man demonstrating the new piano in this video also says it is binaural sampling.



Maybe Yamaha believes binaural is the best way to record piano now. I wonder if these new samples will trickle down to their digital pianos.

Thank You, Tyruke for pointing out this.

But You don't have to buy a silent with the prohibitive prices to get the best sampled sounds.

The only problem will be, that even if binaural recorded sounds are really superior, it will take time, before we get it in new SW instrument or in affordable DPs with such new sounds. What we can see, they are working with old (for a decade) recorded samples, only remixing, blending and reprocessing them. (Many DPs have a long way to go until they arrive at throwing out the loops they are castrated to and have all the recorded velocity layers.) Binaural recording would mean completely new sampling. The market of SW Pianos is saturated. But perhaps just for that reason, one or the other could decide to aspire for of an outstanding status by daring such a complete new approach.

My view of the silents: the horrible investition can halfway be justified if only at least the sound module is fully upgradable/exchangable. This is the single most amortising component of the instrument. Otherwise a minimum life cycle of say 25 Years isn't realistic.

I'm pretty sure Mike Martin said at some point that the Privia x50 models detect the headphone and alter the sound in some way to allegedly make it more headphone friendly. I'm not sure exactly what sort of change. My impression is that I hear too much piano noise through the headphones, e.g. the sound of the strings releasing when you press the pedal is very pronounced.

It's possible, but these binaural samples are specifically aimed at acoustic piano silent systems. The Silent SH series uses new binaural samples recorded with microphones in a dummy head. Hence the samples are optimized for headphones, not playing through speakers. The justification for binaural sampling in the Silent SH series is that these are acoustic pianos and therefore the sampled sound is expected to be used primarily for playing with headphones, otherwise the user would be playing the actual acoustic piano. The claim (or inference) in the video that the same samples will sound just as good played back through an audio system with speakers is contrary to the history of binaural recording. Binaural recording is not at all optimum for listening through speakers in a room.

It seems like an opportunity for digital piano manufacturers to include an additional binaural sample set for headphone use, but manufacturers have tried to keep sample memory requirements to a minimum for years, else we would have seen larger sample sets with more sampling levels already. So it will be interesting to see if Yamaha (or other manufacturers) decide to add binaural samples to new designs and market it as an additional feature.

As long as the module is MIDI, never mind how crappy the sounds are already or will be in a few weeks/months/years.

But your view is completely out of the question for another reason : the most important part of the piano playing has nothing to do with the way it sounds, but with the way it plays under the fingers. No DP can compete at this time, not even come close, with what piano learning implies in order to get in shapely contact in minutes with ANY piano you might encounter in your musician's life. This is where the silent piano is not an expansive investment, not only do you learn to play your piano well, but you learn to play any piano, whereas on a DP training you only get to be good on your DP keyboard.

Well, I take Your point, while I don't think, that the weakest part of quality DPs their action would be.

You can use a silent as the most expensive MIDI controller - but the most realistic at he same time, if the electronics behind the keybed is on that level, as in an AG, which I would assume. (In the schema I could see only 2 sensors, but this arrangement would allow detailled enough event capturing when the single hammer sensor a true velocity or stress sensor is, measuring impact energy, e.g. ).

However, in theory, there should be no advantage over an AG regarding MIDI controller functionality only.

Interestingly, all of a sudden I begin to feel a GAS (but sadly out of reach) :-(. Would be very exciting to compare electronic and acoustic response directly with the very same keyboard!

And optimised binaural sound having at first in a silent is a logical development, for the typical use of the digital is over headhones, as Macy pointed out correctly.

Hi Temperament and all,

I happen to be studying some material about binaural recording/rendering and think that your comment is absolutely spot-on. I have been wondering myself why there seems to be so little interest in binaural rendering in the DP market, considering the fact that most DP users spend a major fraction of their time playing on headphones.

As the posters correctly stated, DP makers probably see little practical benefit from including binaural samples nowadays (with the welcome exception of Yamaha on the newer silent pianos, where they are bound to see a benefit).

All things considered, in my opinion it would be worth investigating a DSP-based conversion from stereo to binaural for a better rendition over headphones. This would save the cost of doubling sample size and still provide a better experience when playing silent.

My suspicion is that the limited quality of available binaural converter plugins (see bs2b) is more due to the simplistic implementation than to a conceptual problem. Full HRTF-based algorithms should fare much better than simple crossfeed schemes.

Ciao,
Paolo

A binaural recording/playback system would be best for a headphone application. Yes, you would need a different sample set for speaker playback, but the Yamaha seems to have that as well. That module attached to Yamaha's new silent pianos looks very cool, I wonder if they may ever sell something like that separately.

Thx, Paolo, could You please some more specific about this, (at the time being I have simply no time to investigate this, as a have liked it). Especially, what do you mean with "limited quality" (like "gimmick" by Macy)?

Are these implementations in your opinion better than nothing (as I think), useless, unusable (because of say deteriorating sound quality to such extent). I have experience with VLC, Pianoteq, Cockos ReaSurround only so far. Thx.

Binaural is a very funny thing. I think it works partly through (human) memory apart from purely physical or other psycho acoustic means.

Over 30 years ago I did some experiments in binaural recording using two mikes strapped to my head. One of the sounds was the gentle tap-tapping of my door with the draft running under it. When I replayed the tape I would have the absolute sensation of the door banging 'over there' outside my head, a few feet away. It was uncanny - even after the door had been repaired and no longer made that noise, the recording would give me the exact experience all over again, just as it had been.

But when I asked a friend to listen, he did not have this experience anything like so vividly as I did.

There are two explanations for this:

1. It really depends acutely on the exact shape of your head to get the right results.

2. It involves precise memories of sounds and their location.

...or 3, I'm bonkers.

I could make a very elementary observation by chance in a big shopping mall for some years. There was a very young baby, I can't know exactly of which age, but perhaps between 3-10 Month, it was kept on the arms and shoulder of his dad. And suddenly there came a shop personal announcement from the speakers built in the above ceiling. And in the very moment the sound started the baby turned its face straight upwards the built in speaker was direct above its head.

It was unbelievable, how resolute its reflex-like reaction was.
And in this very early age. (We don't turn our heads upwards, because we do know, where it is coming from, therefore we are not even aware of this amazing capabality of us.) There is no other explanation for this, than that our brain can analyse much more complex and subtle patterns in sound image, than the simple phase shift due to the different distance from a source and the two ears opening. The reflections from the ear produce probably also such phase-shifts (at some magnitudes weaker amplitudes, and even this very subtle anomalies in the mix bear obviously vital information for our brains.)

Toddy, Your observations are very interesting, they can be the result of our information processing capabilites. We have to learn the physical characteristics of our own phenotype (ear lobes, skulp as the resonator) and analyse the small anomalies in sound image against this mental model.

If this is the case, my original objectives (to pay more attention to this aspect of DP sound, listened to both through boxes and through headphones) seems more than justified.

I don't have know enough time for the moment, but is definitively an interesting theme for further study.

If anyone hasn't heard this binaural recording yet, take a listen to this "virtual barber shop" with headphones: http://www.youtube.com/watch?v=IUDTlvagjJA

It's quite a realistic experience.

Hi Temperament,



I have only tried b2sb so far and do agree that it is better than nothing, even if I didn't test it on many stereo recordings. What I mean with 'limited quality' of this kind of plugin is the rather moderate improvement of the 'sound-in-the-head' sensation that sometimes arises when listening to stereo material over headphones. The 'virtual barber shop' demo recommended by Allan is a striking example of what can be done with a true binaural recording in terms of sound source localization.

As you correctly hypothesized based on your and Toddy's observations, according to psychoacoustics studies, the brain does perform a kind of 'inverse filtering' on ear signals to disambiguate among multiple sound source locations that would have identical delay/attenuations. This is based on a learning of the frequency and/or phase response of the body (especially head, ears and torso) to wavefronts arriving from different directions.

In case you are interested, you may look up the keyword HRTF (head-related transfer functions) to get some introductory material on this subject.

Let me add that, after listening with some attention to my DP's sound to headphones, I didn't find it annoying at all and, in fact, quite pleasant. My guess is that the choice of mic positions for stereo recording of the piano samples was made to ensure a resonable rendition over both loudspeakers and headphones.

However pleasant, the sound I hear through headphones is still unreal, i.e. it doesn't sound like a piano in a real room. This is where DSP post-processing could help, by reproducing an approximation of the 3-D reverberations of the piano sound in a variety of virtual rooms, and possibly of the HRTF-based transfer functions from the virtual room to the player's ears.

Ciao,
Paolo

Hi,

For those interested, I just came across this paper from Yamaha R&D describing a scheme where simple binaural cues are synchronized with player head movements by means of a head tracking device. This supposedly enhances the illusion of sitting at a real piano when playing through headphones:

http://www.acoustics.asn.au/conference_proceedings/ICA2010/cdrom-ICA2010/papers/p762.pdf

Ciao,
Paolo