There's a Key Release Noise aspect to all the pianos in Pianoteq (in "Action") but why is there no key press noise? I am trying to get the perfect "peaceful piano" feel to my song (if you don't understand, listen closely to the "Peaceful Piano" Spotify playlist (the one created by Spotify) and you'll realize that they have a lot of different key press noises). I feel like this is something that should be implemented.
There's a Key Release Noise aspect to all the pianos in Pianoteq (in "Action") but why is there no key press noise? I am trying to get the perfect "peaceful piano" feel to my song (if you don't understand, listen closely to the "Peaceful Piano" Spotify playlist (the one created by Spotify) and you'll realize that they have a lot of different key press noises). I feel like this is something that should be implemented.
I think a YouTube link to a specific track or two would be easier for users here to help you.
I don't know if this is what you mean but a lot of trendy piano recordings have lots of pedal noise, mechanical key noise, fingernails occasionally tapping on the keys, very close microphones.
I like the idea.
There are some reasons for not having it as a prominent feature I guess.
When playing as per a real piano in a room:
- Users' dpaino key strikes already make enough noise in their environment.
When used in recordings:
- Studios take pains generally to remove anything like 'noise' unless genre is OK with it.. and even most pop wants everything super clean (subsets wanting crushed/glitched noisy instruments generally just use samples/loops and prob not a selling point to have key hit noise tbh), and it's possibly mostly cinematic music which pushes piano noise.. beyond that, it may get down to individual artists who use their piano in a way where they highlight the live quality of a real piano sound. I guess fractionally, it's a small listener quantity.
Anyway, just a few reasons why it's probably not yet been on the radar, off the top of my head, not sure how likely it might be to come in future - BUT I do really like the idea and would say it would interest me!
(I can imagine it being handled somewhat like key release noise - but not sure if there are impediments to it.)
There's a Key Release Noise aspect to all the pianos in Pianoteq (in "Action") but why is there no key press noise? I am trying to get the perfect "peaceful piano" feel to my song (if you don't understand, listen closely to the "Peaceful Piano" Spotify playlist (the one created by Spotify) and you'll realize that they have a lot of different key press noises). I feel like this is something that should be implemented.
Interesting topic. However, I may be wrong, but it seems to me that Pianoteq has no way to know the parameters that drive the noise of the keys as there is no sensor to capture this shock and final velocity of the key doesn't tell you how the key was stroke. So, noise of key press would have to be artificial and seems not to be compatible as such with the modelling approach where all calculations are made with real input parameters and piano characteristics. Moreover in a live situation, that wouldn't work well as it would interfere with the real noise produced by the fingers.
For recordings, in absence of a sensor to capture the info , a possible workaround would be to use good microphones to record the ambient noise , capturing the noise made by key presses, record the piano track without outputting the sound to external speakers and to mix the 2 recordings in a DAW. worth trying to see it that works.
Alex_Fullam wrote:There's a Key Release Noise aspect to all the pianos in Pianoteq (in "Action") but why is there no key press noise? I am trying to get the perfect "peaceful piano" feel to my song (if you don't understand, listen closely to the "Peaceful Piano" Spotify playlist (the one created by Spotify) and you'll realize that they have a lot of different key press noises). I feel like this is something that should be implemented.
Interesting topic. However, I may be wrong, but it seems to me that Pianoteq has no way to know the parameters that drive the noise of the keys as there is no sensor to capture this shock and final velocity of the key doesn't tell you how the key was stroke. So, noise of key press would have to be artificial and seems not to be compatible as such with the modelling approach where all calculations are made with real input parameters and piano characteristics. Moreover in a live situation, that wouldn't work well as it would interfere with the real noise produced by the fingers.
For recordings, in absence of a sensor to capture the info , a possible workaround would be to use good microphones to record the ambient noise , capturing the noise made by key presses, record the piano track without outputting the sound to external speakers and to mix the 2 recordings in a DAW. worth trying to see it that works.
Actually, for some electric pianos (including mine), the velocity curves section that measures it for you and creates the curves automatically does include the feature of how fast the key was struck. I'm don't remember what it's called but I know for sure that it can sense how far down the key is. So, for some people it could be a feature, but for others it could not.
joannchr wrote:Alex_Fullam wrote:There's a Key Release Noise aspect to all the pianos in Pianoteq (in "Action") but why is there no key press noise? I am trying to get the perfect "peaceful piano" feel to my song (if you don't understand, listen closely to the "Peaceful Piano" Spotify playlist (the one created by Spotify) and you'll realize that they have a lot of different key press noises). I feel like this is something that should be implemented.
Interesting topic. However, I may be wrong, but it seems to me that Pianoteq has no way to know the parameters that drive the noise of the keys as there is no sensor to capture this shock and final velocity of the key doesn't tell you how the key was stroke. So, noise of key press would have to be artificial and seems not to be compatible as such with the modelling approach where all calculations are made with real input parameters and piano characteristics. Moreover in a live situation, that wouldn't work well as it would interfere with the real noise produced by the fingers.
For recordings, in absence of a sensor to capture the info , a possible workaround would be to use good microphones to record the ambient noise , capturing the noise made by key presses, record the piano track without outputting the sound to external speakers and to mix the 2 recordings in a DAW. worth trying to see it that works.Actually, for some electric pianos (including mine), the velocity curves section that measures it for you and creates the curves automatically does include the feature of how fast the key was struck. I'm don't remember what it's called but I know for sure that it can sense how far down the key is. So, for some people it could be a feature, but for others it could not.
Physically speaking there is a delta of time between the sound produced by the finger impact and the moment the hammer hits the sound . A VST has no way to predict when you are going to hit the key , so matter how you want to solve the problem , the result will be hasardeous , secondly in order to measure initial speed of the impact , you would need to have 2 sensors ( minimum to calculate speed) located at the very top of the key just for that purpose . Sensors situated in the lower part of the key dip don’t provide this info , as speed can very a lot during the key descent . You can touch the key gently but inject a lot of force to accelerate the key during its descent . So it seems very difficult to achieve in a synchronous and exact way , and I guess it is not worth it as the true effect exists naturally .
An excellent observations joannchr!
I like the idea of dpianos including more sensors for measuring more accurately things like first touch and time before hammers are struck. Not aware of any dpianos with that onboard but it seems a good kind of 'ghost in the machine' thing to solve with that.. could result in future excellent products with better handling for skilled pianists for sure.
For now, without an initial movement sensor, I'm just imagining it possible with a small latency drag of a few ms, to insert some appropriate pre-strike noise.
An excellent observations joannchr!
I like the idea of dpianos including more sensors for measuring more accurately things like first touch and time before hammers are struck. Not aware of any dpianos with that onboard but it seems a good kind of 'ghost in the machine' thing to solve with that.. could result in future excellent products with better handling for skilled pianists for sure.
For now, without an initial movement sensor, I'm just imagining it possible with a small latency drag of a few ms, to insert some appropriate pre-strike noise.
Regarding your last suggestion, of course if it works that could be an acceptable solution , but unfortunately I’m not even sure it’s possible as you can only trigger the calculation when the first midi message is passed to pianoteq at which point it’s already a race against time to be able to render the sound with the minimum possible latency, so to get a realistic effect you would need to differ the production of the piano sound so that the interval of time between key press and sound of note remains realistic but then you end up with a latency issue inherent with this approach. So only possible approach would be to have 2 sensors at the top of the key bed plus a specific way to transport this message via midi CC , however there is no midi CC message , part of the protocol designed for that , so …. Kind of a simple problem with a very complex solution ….
Yes, seems in hindsight quite an overlooked aspect of dpiano + MIDI.
During early MIDI days, no dpiano action was enough like real pianos and first draft MIDI did well for available hardware/tools, at least for contemporary music production. Understandable really.. but now and near future, perhaps key press noise is something which products like Pianoteq now make desirable (seemingly so close to real piano playing now etc.).
In more recent years, with much more piano-like actions and more abilities baked into MIDI.. it does seem like a rather unobserved moment in a piano key strike, to be missing.. the moment the finger initiates pressure downward.. the speed and also pre-strike 'acceleration' or 'relaxation' (different kinds of staccato for example, one punching through, the other using kinetic drive to action the key but with some immediate release rather than driving through with the same initiating force).
Thinking of note-OFF (really nice to have working well), the idea of 'note ON' really should have been (for piano-like use cases) a sensor beginning before hammer hits.. the idea of note-on as it is should be something more like 'hammer strike'.
But - IDK for sure if there are not great possible ways to work with all within the latest MIDI implementations. Stopped looking into it a while ago. More happy than ever with my setup, but it would be amazing to find out if someone really is, or has solved this somehow.
Re latency - absolutely, many like latency as low as possible. My ideal with latency and Pianoteq, is to experiment with the buffer size in options.. depending on full loop latency, the latency numbers seen in Pianoteq may feel more piano-like if they not so close to zero. On my current system with so-so full latency loop, I like about 5 to 10ms.. sweet spot around 8 or 10 really. My last machine though had a faster full loop latency and I could enjoy 10 to 15ms with Pianoteq.
When I set to lowest latency for Pianoteq, my dpiano feels too much like a normal synthesizer controller. I love a little old Casio for example, which was my chief controller for some years in the past (not used for piano music) - as it (and other tools like it were too) was practically instant from key press to 'note ON' (suitable for MIDI timing focussed work like playing beats, not so much for 'pianistic feel').. but I do believe if some kind of true 'note on' (at first sign of downward pressure) even if it added some 4ms latency, could still be in most users' acceptable range of latency.. without more knowledge of what's to come within the later MIDI standards, it still seems reasonable even with perhaps a 2ms shorter noise maybe.
Fun to think about - and probably it would be something I should check out more info about. Thanks for the enjoyable idea crunching j - Cheers!