[home][back to index]

How to setup the VST engine

First make sure that a working audio output device has been selected by visiting VST|Audio Out Settings. Then you will have to make a choice between the two possible rendering modes:

The Immediate (Direct Sound) method is called 'immediate' since it will calculate the corresponding audio data immediately after MIDI data has been received. This mode requires a fast computer (a 2 GHz CPU is a must) as well as Direct X installed (will always be the case on XP, but might not be on older systems or Windows emulators). If immediate processing should fail for some reason, there is still the Delayed (MME) method which uses the slow, but stable Windows MME audio driver. This method will most likely run on even the most miserable machine, but it has the downside of a very large latency, i.e. a gap between the reception of MIDI data and corresponding audio output, which means that real time recording becomes virtually impossible (every note you play needs nearly a second to become audible).

Another important issue to consider when choosing the mode is the CPU load. Audio data processing is a calculation-intensive task and can drain CPU power quickly. First sign of an overloaded CPU is the emergence of sudden pops and clicks that obviously do not come from the audio data produced by the plugins. Regarding these the Delayed Mode is more stable than the other one and has better chances to perform accurately, even when CPU load suddenly increases for a moment, as e.g. when other programs start or stop. The Immediate mode is quite vulnerable in this respect and produces clicks and pops quite easily when CPU conditions become averse. So it is a good idea to switch to Immediate Mode when composing or recording from a keyboard and to revert to Delayed Mode when it comes to capturing the audio output.

You can keep yourself informed about the current CPU load by consulting the CPU monitor, which is located on the far right of the main frame window. The value shown here corresponds to the percentage of CPU time used up by the plug-ins.

If the display approaches 100 % there are definitely too many plugins playing too many notes and it's time to stop adding more.

In order to safeguard against a complete CPU overload (which would mean a near standstill of every other activity inside the machine, a really unpleasant thing!) there are two internal "guards" installed, that shut down audio production automatically when trouble is approaching.

The CPU Load Tolerance Guard lets you specify the maximum CPU time a single audio block may consume, If this setting is at 100 % audio blocks may not consume more CPU time than it takes to play them back. If two consecutive blocks take more than that there will be general shutdown and a "Time Out" warning will be issued.

The Tick Delay Tolerance Guard measures the gap between two consecutive sequencer ticks and blocks the engine if there are two consecutive gaps larger than 10 msec (a normal tick occurs every 5th millisecond). This will take place e.g. when too large bulks of MIDI data are send to the MIDI outputs, thereby endangering audio processing as well as sequencer timing.

Moreover you may change Buffer Size and Bit Depth. If the whole audio sounds sounds as if someone had chopped it into tiny pieces the buffer(s) may be too small. Otherwise keep this value as low as possible. In case your soundcard does support it, you may furthermore increase the bit resolution of the audio output (not possible with "MME" method, as MME doesn't support 32-bit).

Having gone all through this, now you might start to load VST plugins into the Space Toad, as it is described in the topic "Loading and Unloading VST plugins".

[home][back to index]

Privacy Policy