The future of sampling...is here!

Cheez

Moderator
First there were hardware samplers. Realistic instruments could finally be captured. But then limitation was memory.

Then we have software samplers - broke the memory limitation problem by using PC. But even then, size of samples are limited by the amount of RAM that's able to be loaded. So samples are limited in their articulations.

Then came the first sample streaming technology by Gigastudio. Samples are streamed directly off hard-drives. That solves the RAM problem as only the first part of samples need to be loaded into RAM while the rest of the sample is streamed in real time. This lets one instrument go up into the GB size instead of just MB and yet allow for low latency. Sample size increases dramatically with multiple articulations. Others like Kontakt and various VI soon follows.

But samples are still pretty dead in itself. It's just a recording of the instrument played. Expressiveness is great, but it it's hard to get the different tonal colours of the instruments that varies during live playing. This is especially true for solo instruments which is pretty transparent. Others tried to add true legato sampling (actual recording of legato intervals between 2 notes played - without using portamento) - like VSL, and East West. They did pretty well, but doesn't address change in tone when playing one note (which most would just apply filters to emulate, or using crossfading of samples - which is again limited in realism).

Then come modeling + sampling hybrid technology. Basic sampling is there, but modeling is added to give greater expressiveness LIVE instead of sequenced. Garritan and Giorgio Tommasini came up with their first - Stradivari Violin (and later Gofriller Cello). Wallander follows with their trumpet (and now woodwinds) - simply amazing sample library!

AND NOW....

Based on the same modeling and sampling hybrid, others are taking it on and pushing up realism to the next level. Sample Modeling presents: THE TRUMPET (by none other than Giorgio Tommasini who is the first who came up with the sampling modeling technology) and Peter Siedlaczek (grandfather of sample libraries - for those who recall his famous 5 discs Advanced Orchestra Library - that still sits in my hard-drive).

Sample Modelled Instruments - The Trumpet

You need to listen to the demos to be convinced! I'm a trumpet player - and I never dreamed the day will come sample can achieve this kind of realism. When they do that for the saxophone, then it's totally game over for real players. Composers no longer need to hire real players for recording - they will be out of job.
 
Last edited:
had a listen to it...wow freekin scary i can't tell the diff.....maybe just maybe only if one were Miles Davis then maybe can make out the diffs.

well we are approaching a reality where VR is everything.........and its simultaneiusly exciting and horrifying!!!!

one day can make music with yr mind!!!
 
cool really cool. espically the 'ends' of every note, tons of expression in them. wonder if they are gonna control it via velocity levels.
 
Zenguan, by "velocity levels", I gather you meant after-touch. By far, after-touch is not the easiest to control. There will be key-switches and mod wheel controls - both which are much easier to control. The question is whether they will program it for the breath controller.
 
Thanks all

Thanks Cheez for your kind presentation.

A seamless continuous crossfade between the dynamics is possible due to the Harmonic Alignment technique that I developed, and used for the Stradivari and the Gofriller. Thus, no velocity layers, but CC11 allows to pass from ppp to fff.

Please feel free to ask questions. I'm very happy to be a new member of this nice community. :)

Best,

Giorgio
 
I think the best way to verify the identity is to ask him questions! :D

This is indeed a good idea :)

That's me anyway. I found this nice presentation of the Trumpet over the Internet, and I felt it was right to partecipate.

Please consider the time lag in case I don't reply promptly.

Best,

Giorgio
 
That's great!

So, Giorgio, why did you do the Trumpet instead of other instruments?
How long does it take for you to come up with this breakthrough technology in sampling realism?
Do you mind to also share with us what sparked you off in pushing realism to this stage?
 
Last edited:
after Dr Giorgio Tommasini answers pf's question, we need to warn all the softies in this thread that "Elvis has entered the building"... hehehehe
 
stories

pf,

We chose the trumpet because we thought it was a real challenge. The development of the adaptive model gave us many insights into the natural behaviour of the instrument itself, and on the interaction with the player.

Not a physical modelled, but a physically oriented VI, indeed.

I've been performing research in this field since almost ten years. The thrust obviously was my dissatisfaction with traditional sample libraries.

As a Cardiologist I was interested in applications of computers in medicine since 1974. Imagine that my PDP 11-23 had a 32 KB RAM, and read floppy disks whose capacity was 384 KB. :)

Now it goes faster. ;)

Best,

Giorgio
 
Hang on Doctor...what has your development of samples got to do with your interest in the applications of computers in medicine?

I can understand that music is your hobby. I also can understand that probably you weren't too happy with the transitioning of the different dynamics in traditional sample libraries, so you come up with harmonic alignment technology. I can perfectly understand your interest in computers as well.

Your field is actually in cardiology. How do you use your professional area of expertise to help your research in sampling technology?

By the way, I was told by a band conductor that the most difficult instruments to play are the French Horn and the Oboe. Have you studied the differences in the musicians' interactions with other instruments (such as these) to finally decide on the trumpet?
 
Last edited:
pf,

I spent some time on modeling, several years ago, compartmental models to study enzyme kinetics, to measure the size of an infarct. Modeling on electromaps (surface recording of the ecg). I developed a rotational technique applied to coronary angiography, and a software for automatic quantification of coronary stenoses. Image processing by AI. The step towards sample processing was not an unexpected one, you see.

Oboe and French Horn are being analysed. Of course what will come next will remain a small secret. Perhaps unexpected. :rolleyes:

Giorgio
 
That is awesome work on integration and application of your knowledge, doctor! No wonder you can write 150 research papers.

Have you learnt to play the trumpet in the process of your study? :)

What are other difficulties that you faced in developing the application of the Trumpet (in obstructing the sample to be as realistic as can be) besides the velocity response issue of midi keyboards? And how do you mitigate them?
 
Last edited:
wow... this guy's been impersonating the doctor for 3½ decades! heheh just kidding.

Dr giorgio, I have 0 knowledge and barely any experience about midi (i'm a recording sorta guy), but my deepest respects to your commitment and contribution towards today's audio technology.

pf : remember to get a copy and mail it to the doc for an autograph haha

cheers
 
Back
Top