Messing with GPO.....

[sec2]

... this time GPo standing not for "Personal Orchestra" but "Pocket Orchestra".

Garritan Pocket Orchestra is a set of Garritan samples, and relevant sfz programs, which comes with any version of Cakewalk Dimension. I am not sure, but I believe the samples are the same as GPO, only a smaller set as they do not cover the full chromatical range.
Nice thing is that they come in .wav format and I could not resist the temptation of importing them into my soft sampler. Just for fun, but also a little bit to see how they would behave in a different environment with more options to control them and make an impact on them. So I pulled the samples of the string sections and imported them.

Said samples include long bows, short bows, pizzicato and tremolo sets.

To cut a long story short, at the end of the process I found myself with 6 patches. 4 patches are the ones relevant to the available sets of articulations. The remaining 2 were a mere product of experimentation.
The "programming" (well... yes... I am laughing at myself right now ) is common to all patches.
Basically, I adopted a common AHDSR envelope, which is though fully adjustable (in real time via mappable CC) and added a little bit of behaviour.
Having just one velocity available, I had to apply a little variation in tone, and did it in proper GP(ersonal)O style but, instead of velocity/cc1, most is done here via velocity only, which controls volume, a lowpass filter frequency, the intensity of the AHDSR envelope and the attack speed.
To move things around a bit, notes' velocity varies randomly within a -10/+10 range through a random generator.
Sections have been panned in, roughly, orchestral positions, but I also added a random generator on the panning, which "pans the panning" -15/+15, thus trying to simulate the physical "lenght" of the section.
Adjustable glide modulator (portamento) was also defined, but I have not really been able to test it properly. Legato behaviour is in also, and an adjustable vibrato via a LFO applied on the pitch of the note, controllable via CC1. As the LFO is adjustable in real time in intensity (cc1) and frequency via mappable cc, it is also an acceptable simulator of trills.

The really nice things, though, happened with the two extra patches I mentioned above.
Basically, these are compound patches built layering the short and long bows. Violins have two sample sets of short bows, remaining instruments just one. (well.. so it seems, as naming and consistency of the sets is a bit chaotic. This is the way I handled them, anyway).
In the case of violas, cellos and basses, this compound "sus+short" patch triggers the layered long and short bows samples, but in case of fast repeated notes triggers alternatively the layered sample and a modified version of the long bow sample (different starting point). In case of the violins, there are two separated sus/short patches, each created by layering the long samples with each of the two short sets of samples. They get triggered alternatively, round robin style, plus a third set of samples (again, long bows with different attack) gets triggered in fast repetitions alternatively with the other two.

This might sound a bit "messy"... told you this was just experimentation. The point is... these actually sound interesting.
I would like to share here a couple of tests I've done with these patches applied on standard midi files, mainly for curiosity and to know your opinion (I reckon there are pretty qualitative "ears" on this forum) but I do not know whether I would infringe any copyright law regarding the samples sets if I publish them. So, in case this might be interesting to anyone... and this is LEGAL, I can just forward them.

One last thing: this doesn' t absolutely pretend to be taken seriously. It was and is just a "game".

Bye to everyone.

Fabry

[Peter Jeffrey Gale]

Hi Fabry

It's not legal to share what seems to be a wrapper product that includes the samples, I think, but I'm sure you're free to share wav/mp3 files of any productions you've made using you legitimate copy of a Garritan product, however you've produced it.

If that's the case, just host the the wav/mp3 somewhere on the web and post the link here.

I'd be interested enough to give them a listen.

Peter

[sec2]

Hallo Peter, thanks for the reply.

Yep, sharing the mp3's is exactly what I had in mind.
And I believe, too, that you are right about the copyright issues, so I' m giving the links to the files. Goes without saying that I' ll remove everything if something is not correct.

So, there are 4 of them. All make use of the said compound patches. These are not perfect renderings, as they were made just for testing purposes. Plus, they absolutely have no extra programming: just notes velocities (obviously...) are defined, no extra controller, no volume envelopes, nothing. These are just the plain raw patches at work, exception made for convolution reverb which has been added on the master track but could as well be programmed into the patches themselves.

Die Zauberflote, Ouverture. This is a full orchestral rendering, strings are though overmixed to make them more noticeable. The file got cut during the transfer, don't know why. But, seen the purpose of the file, it should not be much of a problem:

http://www.4shared.com/audio/zOOoBEQ...Ouverture.html

Beethoven' s 5th, 1st movement. Same as above, strings are overmixed. Winds are from GPO4, here:

http://www.4shared.com/audio/LeGnJrQq/Beeth5.html

Die zauberflote again. This time the isolated strings tracks of the first Aria from the Queen of the Night. Don't recall the actual name of the aria, I just call it "the first aria..." :

http://www.4shared.com/audio/MT9cxLN...te_Queen1.html

Eine Kleine Nachtmusik:

http://www.4shared.com/audio/Cu0ZD2dF/Eine_kleine.html

Well, that's all, hope you enjoy.

Bye,

Fabry

[Peter Jeffrey Gale]

Hi Fabry

I can see now that you've taken quite a large and useful step towards understanding samples better and make how to make better use of them, or at least the GPO ones.

You're demos are quite pleasant enough to listen to, though I do understand that that is not the primary point of your exercise.

I think I'd be interested now to hear at least one of the pieces with no IR on as a comparison, and I'd be interested to know what conclusions you have drawn for yourself and any lessons you feel you have learned from the exercise?

[fabiolcati]

Fine Beethoven direction.
I guess there's some low frequency floating here or there?

[sec2]

Thanks for the replies, guys....

I bear no credit for how pleasant to hear those files were. Let's thank the guys who made them (unfortunately they lay somewhere on the web, unknown ), and of course... Mozart & Beethoven.

Peter,

http://www.4shared.com/audio/D-ms-eBh/Queen1_NoIR.html

http://www.4shared.com/audio/qdBcmi3...eine_NoIR.html

These are the "First aria of the Queen...." and Eine Kleine Nachtmusik rendered without convolution reverb. I am glad you asked for this, as it shows how powerful a full AHDSR envelope is. Just switching off reverb on the original files made them sound terrible, as fast bpm coupled with short notes and therefore cut tails is way too ugly. Just reshaping the envelope using longer release values put things in place again. Now the patches are really "nude".

Fabio,

what you point out might be due to the adjustable vibrato, the "random
panner" which is actually supposed to behave this way but might need a small adjusting, or a difference in volume between the alternating samples, which I have to check.
Three days ago this patches did not exist even in the back of my mind! Some fine tuning is definitely due....

My opinions? I would not like to express one. By hearing the files anyone can make up his.
But one thing must be said: making your own patches is definitely funny. Nicest patch I did is a fretless bass, always from the Dimension collection of samples. This one has several layers which play alternatively and has different articulations that play at different velocities. In one single key switch it plays long, vibrato, thumb, slap, slide, harmonics, fx plus all of the above with various noises added, for fast repetitions. I did a rendering of Birdland with it and it really shines in there.

[Peter Jeffrey Gale]

Thanks for doing that Fabry.

Needless to say it wasn't very pleasant listening to these new versions! LOL

I did a similar exercise myself not so long ago using the raw samples and off the shelf midi files, and I drew the same conclusion as you just have ... not just that applying (convolving) a well chosen IR sample (or AHDSR envelope as you call it) is a powerful way to way to improve the sound of the GPO samples, but that it is the first and most important step that should be taken when using these sounds, even before worrying too much about midi editing and other effects (panning, delay etc), which are all of secondary importance.

Peter