PPG Phonem (Advanced Vocal Synthesizer) official released

[c_voltage]

I also think about grab this, very like different speech generators, but only thing what about re-registration availability ? I see there write "Online Activation", but it's support license deactivation ? I mean, in case of computer upgrade for example, would not want to have any problems with it. I just haven't dealt with the Wolfgang instruments yet.

[ErikH]

Thanks for Phonemklatura, Simon!
Phonem is eccentric as it5 is powerful and I can use every example I can get. The factory soundbanks are very good, and Phonemklatura is on par with it. Yet, knowing your personal style better, your creations are easier for me to analyse and learn from.

This said, I have a question: I am on a Windows 7 machine, and I notice that in my PhonemSounds folder (the folder in which the Resources folder also resides) five Temporary Host Banks reside (0 to 4) , and all five of them are filled with exactly the same .utd and .wtz files as the Phonemklatura folder (wich resides in that same PhonemSounds folder).
Being temporary folders, shouldn't at least four of these and perhaps even all five not be emptied when a Phonem /Cubase session is ended?

My pet peeve is still the fact that Init New does not open an empty session.
Apart from this, I also hope that in a future version a Preferences setting can be organised that would allow the user to specify a default base/folder/bank of choice which Phonem will use. I mean: if it is a factory setting that nothing can be saved into the factory bases, then why not change to, say, the User Base automatically when a user wants to do something? Including of course the Wave Base. This would be much more user friendly than that pop-up window telling the user to choose the User Base. This may not be possible - I am not a programmer - but in my opinion it would be way more simple if Phonem would save temporal work into a general User folder/bank from which it then could be copied into a final bank.

[Sampleconstruct]

Thanks for Phonemklatura, Simon!
Phonem is eccentric as it5 is powerful and I can use every example I can get. The factory soundbanks are very good, and Phonemklatura is on par with it. Yet, knowing your personal style better, your creations are easier for me to analyse and learn from.

This said, I have a question: I am on a Windows 7 machine, and I notice that in my PhonemSounds folder (the folder in which the Resources folder also resides) five Temporary Host Banks reside (0 to 4) , and all five of them are filled with exactly the same .utd and .wtz files as the Phonemklatura folder (wich resides in that same PhonemSounds folder).
Being temporary folders, shouldn't at least four of these and perhaps even all five not be emptied when a Phonem /Cubase session is ended?

My pet peeve is still the fact that Init New does not open an empty session.
Apart from this, I also hope that in a future version a Preferences setting can be organised that would allow the user to specify a default base/folder/bank of choice which Phonem will use. I mean: if it is a factory setting that nothing can be saved into the factory bases, then why not change to, say, the User Base automatically when a user wants to do something? Including of course the Wave Base. This would be much more user friendly than that pop-up window telling the user to choose the User Base. This may not be possible - I am not a programmer - but in my opinion it would be way more simple if Phonem would save temporal work into a general User folder/bank from which it then could be copied into a final bank.

Thank's Erik, I guess those temporary host folders which get created when you open the FX template prject from Phonemklatura are related to Cubase and the amount of MB it can actually store with a project, because those folders don't get created when I use Logic with Phonem, I don't know the limit exactly, it must be different from DAW to DAW.

[c_voltage]

Ok another question
What about automatization in Wolfgang synths ? For example - this is very well implemented in Phonem ? i.e. quite deep ? For example there is i see some partials system like additive in Phonem, so this is all automatable ? Or no.

[ErikH]

... and this is quite frustrating: I created a text ( one two three four can I have a little more)and tweaked it so that it was very understandable and decided to save it. I was working where I wanted to, but Save As did not work as the bank was full. Huh? Then, on the Brows tab it appeared I was not working where I believed I was. So I had to create a new bank and saved in there. All good end good?
No! What was saved was the original setting before the about three hours of tweaking.

The synth itself is extremely flexible but this saving system is about the worst I ever experienced. This is simply counter-productive, and I only have one life. I won't write how I really feel right now as I respect the friendliness of this meeting place, but I guess I'll better now take my bike and go for a few hours' ride to calm down. It feels like having the car of your dreams but no gear box in it. And then it appears it has already gotten better...

[Sampleconstruct]

But you did watch my tutorial and the necessary steps it takes before starting to tweak things...just saying...but I know how you feel, the first 10 days with Phonem were extremely frustrating for me as well.

EDIT: But the utterance you edited should be in your user library anyway, no? Did you not save anything during those hours?

[ErikH]

Thanks for your quick reply.
Innfact, I came back to specify things:

The project that was saved includes my matrix- LFO- and Vibrato settings. But neither the phonemes I added - many are better when doubled, like for example one is better as w/ah/n/n/ - nor the lengths of the individual phonemes I set.
Same unfortunately for the Cubase project.

I added phonems by copying and pasting, just like the help and you describe. And it was after I added them that I tired to save.

The frustrating thing is that, yes, Phonem is *the* missing link but what is saved is not on the synth level but only on specific parts of it.I must try again untill I find it, but I am not a very patient person and I cannot spend hours and hours trying to follow the rituals of a genial mind. If I want to make coffee, I take my Bialetti, pour in water and coffe and heat it untill finished. I do not start by running up and down the stairs five times whilst reciting the alphabet backwards.I mean: I am pretty shure Wolfgang must have his reasons for this, and the moment I grasp these, the problem might be solved. So I take full responsability and blame it on my lack of experience.
But it is frustrating indeed.

[Sampleconstruct]

The general thing to keep in mind: resources in Phonem and other PPG synths can be regarded like samples in a sampler - so when you save a “patch“, the samples are not saved with that patch, only the shell is saved. Now in Phonem we have 2 different sorts of “samples“ -> utterances and wavetables, so actually 3 things have the be saved individually. Saving a preset only saves the shell, but not the assets.


For that reason all the blues light show up when you have changed an utterance and/or wavetable, I suggest to enable auto-save in your DAW when working with Phonem and to frequently hit save, but the first step should always be to select an active Bank on the left side of the browser and to save a resource into your user library once created. I haven't lost a thing in months now and don't think about it anymore.

[Sampleconstruct]

After testing the latest Beta from Wolfgang, I'm happy to say that he drastically reduced the CPU-usage, like really drastically. E.g. in the old version with one of the pads from Phonemklatura I could play 5-6 voices at a sample buffer of 128 samples on my 12-core Mac Pro until it started to crackle, not it is like 16+ voices until the CPU gets fried.

Also the synced LFOs finally sync how they should, the last improvement wasn't really quite there yet, but now the synced LFOs play absolutely tight (the low notes still lag a bit, but that can't be changed)!

And finally Wolfgang also removed the nagging warning message “please select user directory“ when trying to save utterances/assets into one of the factory folders, because now the user directory gets automatically selected. That should make some people here a little happier

Wolfgang, you rock!

[ErikH]

Hey, that is good news. I must confess I worry a little that Wolfgang would assume I don't like Phonem, but really, if I didn't I would not waste any time on it. Thing is I've never been beaten by a file system/save before.

Your post on samplers confirms my assumptions. I guess that for someone with late-seventies and more recent hardware synths the concept of banks must be completely natural, but I never had one of these. I literally jumped from a Roland System 100 to Absynth.

This said: I decided to have a good night's sleep over it, and it all begins to dawn:

- At the heart of Phonem there is a filter that can be driven by three sources: a sound generator that mimicks the organics that make the human voice, wavetables (.wtz) and time-corrected samples (.tcs) . The latter two can be created in WaveGenerator and Wavemapper but need first to be translated.

- The sound generator can be tweaked with an Excitation Editor that has three controller options: Loudness, Tilt and steepnes
- The filter has twelve formant resonators that can be individually tweaked.
- Phonemes in Phonem are specific settings of the resonator filter and the sound generator that have been factory saved.
- A string or chain of phonemes and intervals like, for example breathing space, is called an 'utterance'. Each utterance in Phonem is saved in its own file in a specific subfolder in the Resources folder with the .utd extention. A wavetable is stored in a subfolder named 'wr' inside said folder with the .wtz extension.
- to be able to save utterances of our own, we have to create our own folder inside the Resources folder.
- Banks are sets of max. 128 units of utterances. Banks have the extension .fbx .
- When an active soundbank (Browser page) is exported, it gets the extension .phbnk .
- When importing a .phbnk file (like Phonemklarura) it creates a .phbnk file, a .fbx file and its folder inside Phonem(PhonemSounds in Windows) and this .fbx links to a user-defined folder('Simon') in which the utterances and wavetables are found.

When I add here your likeness with samples in a sampler, it as good as completely falls together.
So, if I may, the final things are:

- Does the Base show resources?
- Are Programs another word for utterances?

Sorry, but my brain needs a good understanding of the skeleton before I can start adding muscles and tissues...

And thanks once agin.
I look forward to that update.

[Sampleconstruct]

No, programs only contain the data for envelopes, LFOs, modulation assignments, FX setup, utterances are assets (like samples) and are stored as .utd-files in -> resources -> user/whatever name.

Select a base - e.g. Simon - and have a look at the resources contained in that Base using the little arrow.

[ErikH]

OK.
Now I know enough.

I could play with Phonem and make it generate the most fabulous and impossible sounds, but now I now how I have to save them and, what's more important to me: why it has to be done like that.

Thanks!

[Sampleconstruct]

Version 1.02 has been released, thank's for all the improvements, Wolfgang!

[pdxindy]

Version 1.02 has been released, thank's for all the improvements, Wolfgang!

What sort of improvements?

[Sampleconstruct]

Version 1.02 has been released, thank's for all the improvements, Wolfgang!

What sort of improvements?

Mentioned a few posts back, massive CPU reduction, LFO sync-final fix.