And here is a preview of another new grand piano sample called the EMPEROR. This clip is made with 24 bit version audio with 10x10x4 layers (10 pedal up and 10 pedal down), separate release layer, pedal activation noise, body resonance and room IR.
may I permit to offer a critic concerning pedal action noise?
I always teach my piano students to use the pedal very gently and to avoid all possible disturbing noises coming from its mechanism. Expecially in slow movements like Beethoven\'s demo one, the foot should move very carefully and very slowly, thus there should almost be no noise at all. The only noise you could have is that coming from dampers leaving strings in the case you completely release pedal, but if it is too evident it\'s a sign of a piano in not so good conditions. Anyway you shouldn\'t hear so clearly the mechanic noise!
Piano pedals use is an art in itself, you can produce infinite colour variations through a fine use and combination of the two! It\'s a pity actual technology doesn\'t concentrate yet on a more sophisticated pedal simulation!
So true! The pedal noise is often a very disturbing factor.
In case of sampling a piano the magic word is \"control\".
The pedal noise in this demo is triggered by the activation of the sustain pedal. Going from 1 to 127 (or vice versa) triggers the corresponding sample at a certain velocity. In the demo the pedal noise sample is triggered at a fixed velocity. In the ideal situation you would want to have the speed of depressing and releasing the sustain pedal to determine the velocity value of the pedal noise samples. And this exists already: The better Yamaha Clavinova pianos series have this feature implemented. The pedal movement is measured and transmitted as continuous midi data values.
Most old-fashion midi keyboards have a simple pedal connected that acts as a switch with two positions (ON and OFF, value 127 or 1). So, using a Yamaha keyboard would be more realistic since the sound that comes from stepping on the sustain pedal with full force will result in the pedal noise as heard in my demo and gently pushing the pedal a little would not result in any audible sound. Just a distant but realistic whoosh.
When you step on the pedal with full force the sustain (and the symphatic resonance of notes sounding) will drastically change. When you hardly touch the pedal there will be far less resonance and less sustain. This is another issue that can be solved; when you use two samples (sustain pedal up and sustain pedal down samples) for the sustain pedal dimension you have the ON-OFF scenario, which is an ugly immitation compared to the real piano behaviour. When you use convolution to emulate the resonance and sustain you get full control over the resonance dimension. When you play a chord and push the sustain pedal afterwards you WILL hear resonance build up!).
Another issue is repedalling. When you re-pedal a note (first play the note with pedal down, keep the key depressed and release the sustain pedal, then push the pedal again) you want to hear the resonance behaving exactly as in real life. When you use ON-OFF samples only there will not be any effect at all!
Since midi controller 64 data is sensitive to all values from 1 to 127 you can easily create your custom sustain pedal controller by routing continuous data to controller 64 using a midi processor in your sequencer.
yes, the only actual way to emulate the full behaviour of piano pedal would be the use of a resonance model by convolution. Piano resonance should be sampled at various pedal depressing depth (including no pedal which should be useful for the undampered piano high notes) and applied by convolution at the corresponding pedal down value (I would not call it velocity).
A separate velocity sensor for the pedal should trigger the pedal action noise which is an independent factor. That gives us 2 values for the pedal controller: depth and velocity. Does the Yamaha Clavinova read them both?
Will GS3 be able to apply convolution in \"real(!)\" real time? Then we could have some very enjoyable time in the future!
Also, I do not clearly understand from your post if your piano is already responsive to this kind of pedal model!
Convolution is realtime in GST3. When I played the new GST3 during the NAMM show, Jim van Buskirk did a short demo of the new Gigapiano that uses body resonance from Gigapulse. There was no delay in the GigaPulse sound.
Sampling sustain pedal down notes can still have good results but their use is limited to specific -rare- situations.
I already have all three of your new piano libraries, and then you\'re coming up with two more that interest me a lot, both the old lady and the emperor.... Makes me wonder just how many Post Pianos will end up in my system....
By the way, have you received the private message that I sent you a while ago?
[ QUOTE ]
I thought some pretty cool things were afoot when the PMI Iowa was released.
I like this piano alot. Now when you gonna stop teasing and release it for us to buy? I\'ve been itching for a new piano library [img]/ubbthreads/images/graemlins/grin.gif[/img]
[/ QUOTE ]
Thanks for the compliments, really appreciated.
Stop teasing? No way. These two pianos will be available in the next moth.
We are currently making a stable version (with all small details fixed) and will probably release this stuff in GST3 first, since the high resolution quality and the full range of features (pedal action, re-pedalling, body resonance, room resonance and smart midi functions) are all included in GST3 and GigaPulse.
So to answer your question (and the one others have) the release will be when GST3 is released.