www.psychosynth.com

The first one makes the program more visually attractive but also more useful. It now shows the signal going through a connection in real time. This lets us make a clearer idea of what is going on in the sound synthesis. By now, as usally audio signals are much faster than control signals, we show the las 0.05 seconds of the signal for audio, and a second for control signals, but it will be configurable in the future. I am considering that, as audio signals are quite fast, it may be convenient to implement a pitch detector to adjust the temporal frame displayed to a multiply of the wave period, so we’d get a better visualization normally.

This also implies that a generic mechanism to observe the audio signals from the client code have been added, which may permit the creation of beat detectors to control light effects or any other interesting visualization.

The other functionality is less eye-candy but it solves on the most recurrent complaints to the program that I have received from real music makers. When you changed a connection between two objects, if the signnal that was being sent was not stabilized, a small but annoying click would be heard. This is caused because you would move suddenly from an arbitrary value to zero, so we were unconsciously generating a pulse, which we hear.

Most software modular synthesizers do not have to fight this problem, because conexions are not changed on the fly. But in our case one of the main purposes is to build the synthesis patch in live, managing objects and connections in live as a fundamental part of the performance. My solution is to apply a very fast -so its effect is not noticeable, I will create specific evelope objects later- linear envelope on the signal, so we actually enqueue the connection change query and wait for the envelope to finish to really run it. Also now when deleting an object we have to wait for all its envelopes to finish before actually removing it.

This solution is performing quite well, but if someone thinks that he can give a better approach the topic is still open to suggestions and improvements.

I hope to be able to go back to the funny part of the job.

I’ll also comment the development news. As I promised I have already written the system to store configuration settings, but it still lacks a window to edit these on the 3D interface and finish tuning the synthesis systems so we can change parameters such as the sample rate or the buffer size without having to restart the session. Suddenly I am quite busy with things appart from the project, but I will try to find a gap to finish this before the next week starts.

On the other hand, improvements don’t finish there. I have broke the plan a little bit to work into an arduous task that I have been postponing a lot of time: separate the library code and the applications and create classes to ease the creation of Psychosynth apps. Also, now both int library and application code dependencies are not compulsory, so now the are optional. For example, you can compile without liblo support if you don not need network communication or lack jack if you are not an advanced user. It is a boring task but fighting with autotools and refactoring have been worth it.

Now, looking forward to the future, I am starting to consider starting to write LADSPA plugins support and doing some important changes in the core. I will keep you informed.

Actually, the lack of activity in the web does not mean that there were no significant news. Now there are LFO’s and a variety of filters, which allow a richer play with the sound, which can be generated with new wave forms and a ring modulator has been added for both audio and control signals. Also, objects can be deactivated without deleting them just by clicking on their connection. There is a multilevel and hierarchical logging system with multiple output. And the generated sound can be recorded into wav files. I have also made important modifications in the objects system to give it more versatility and will allow us to write a wider variety of synthesis objects. I have also made a quite synthetic diagram of the architecture of the program which can be viewed at the documentation section of the forge. And probably there are more things not included in this list.

On the other hand, during the exams period I have used some isolation time to think out the next things to implement and the pause have also served to refresh ideas and rethink some things. So I will comment briefly what are the next things to do:

As I could not find any library that satisfied my taste, yesterday I started to write a system to store the settings of the program. I will be hierarchical and MVC based and the backend will be interchangeable transparently -by now, I will use XML, but one could use gconf or the Windows registry, for example. I will also complete this work with a system to update automatically the configuration parameter and a class PsychosynthApp to ease the development of Psychosynth based applications.

Later I will put my hands to work into a triccky task: the creation of rythm generator objects and the solution of those small clicks that can be heard when connecting and disconnecting objects. Luckly, I have almost decided how to implement all this.

Once this is done I have many interesting ideas in my mind but I do not know in which order I will do them. For example, to do the mapping between the real synthesis objects and the visual shapes that you see on the screen via a XML based language. I also want to write a reacTIVision based controller for the 3D simulator, so we can have a Reactable without expending lots of money in expensive devices. Other things with the same priority is the possibility to play files -adjusting the tempo and such, which would become Psychosynth in an attractive weapon for DJ’s- and the visualization of the signals that the objects emit.

At last, I want to mention my decision of not developing the documentation during the contest. There are still several months of development before the API can be made stable and I would also like not to make it stable until some other developers take part in the project and make suggestions and feel comfortable with the libraries. As you can see there is a lot of work to do and there is no need to hurry too much with that.

With this serveral synthesizers can connect each other through a server to sync their state. If the network latency allows it -by using UDP we’ll get low latencies usually- we’ll be able to perform collaborative music from different places

At christmas’ end I’ll give you a new vide showing this and other improvements that are to come. ¡Happy holidays!

The network system is going forward too, but I’m having difficulties to find free time to expend on this.

• The object visualization and management code has been rewritten and it has been connected with the existing synthesis code, by adding elements to visualize and manipulate the object parámeters both in an abstract reactable-alike manner and by directly entering the actual values.
• Part of the code in the dynamic patcher have been redesigned and some bugs have been solved in this module.
• The camara control have been improved.
• The visual theme have been improved and there are new texturized models for the objects.
• A new argument parser module have been written, despite not being used yet.

All these changes are already on the SVN repository, where you can follow the development until there is a mature version ready. In the downloads section you can find instructions on how to try it.

Now I’ll spend some days in other things but later we’ll go back with the proyect. The next task is the reimplementation of the network module, wich will substitute the existing prototype by new code based on OSC.

PS: In the video the sound is delayed at some places and there are some lags. This is the sceencasting program’s blame, the synthetizer runs fluently.

Because of this, I won’t update the savannah CVS repository until the competition ends, on April 7th. All the development will be done at the contest’s forge. I have just uploaded to its SVN repository all the old code that you can find at Savannah with some improvements done in October.

I find that this contest is a very motivating experience that will boost the development of the project. Along this week I will upload more information about the development roadmap and the design of the internals of the project.