Been in Scotland for near 2 years already, working at RealtimeWorlds, more specifically in a game called APB.
Times does fly when we’re having fun!
Meanwhile, I did work a bit in czPlayer, but nothing really new. Mostly bugfixes or small features for a client.
So despite not having any updates here in the blog, the following things happened to czPlayer
- Some games using czPlayer were release
- A client added iPhone support. Was as simple as creating a single class with about 160 lines of code.
- Implemented volume ramping, which helps a great deal in removing clicks, especially for chiptunes
- Added some other small features, again with minor effort. 
- Fixed some bugs, especially related to Symbian.
Also, apart from czPlayer, I managed to squeeze in some other small freelance work.
The last few months I’ve been thinking where to go next with czPlayer, and I’ve decided to go with a new design.
Personally, I think the current design is really good performance-wise, minimizing copying memory around unnecessary or other unnecessary overhead.
For example, internally a Mixer object contains MixerInputGroup objects, and each MixerInputGroup contains individual MixerInput objects. Each of those classes have a SetVolume method, which allows me to have several levels of volume (aka as Global volume and channel volume) which get multiplied to calculate a final volume to apply in a single mixing pass, avoiding the need to mix at the MixerInputGroup level, and then again at the Mixer level.
Although the current design has proven successful, it does have severe limitations in flexibility, has for example not being able to apply an hypothetically effect to a single sound or sound group. It’s only possible to apply an effect to the final mix.
So I decided to go with a new design which should be more flexible, but I’ll still use the old codebase as reference, especially because of the small fixes/features I’ve implemented meanwhile.
Some of the new features I’m planning:
- DSP plugins. Will probably provide just 1 or two as an example.
- Configurable DSP network. This will make it possible to create a custom DSP network and play sounds in specific point in the network.
- Streaming support. Only Ogg Vorgis planned so far, but should pave the way for other formats
- Maybe virtual voices (ala FMOD style), if I get the time.
Again, those features shouldn’t add significant overhead until used, of course.
If the user sticks with the default settings, performance should hopefully be comparable to the current version.
I’ll be dropping support for old Symbian editions (< 3rd edition), which should make some things easier.
I wish I could drop Palm OS, which is a real pain to develop for, but a particular client still requires it.