Whilst writing the constraints section for the project proposal a number of considerations have been bought to attention.
Frequency range of modelling accuracy:
Time constraints mean that not all frequencies of a guitar will be recorded, analysed and modelled so how much will be? The first 12 frets of a guitar contain 31 unique frequencies (which some cross over beyond the 12th fret).
Volume, Gain and Tone controls:
The amplifier has three settings controlling the pre amplifier level, the power amplifier level and the tone stack of the amplifier. Due to non linearity of analogue circuitry difference between 2 - 4 and 4 - 6 in terms of gain will not be the same however not all 1 - 10 values can be calculated. Tone stack should prove easier to model as it acts as a single band EQ.
A key aspect of guitar amplifier sounds are due to the speaker and speaker configuration of an amplifier. This project however does not include speaker emulation as one of its main aims and instead focuses on the pre and power amplifier valve sections. There are two key issues with this, first of all if the output of the amplifier is to be recorded then using a microphone to measure the speaker will colour the output with the frequency response of both the speaker and microphone. There is however a speaker output on the amplifier for connection to external cabinets with an 8/16 ohm resistance output. If this could be used as a line in then an uncoloured signal could be used to truly represent the output of the amplifier sections. The second issue is how to test the final model as a model of the amplifier, a similar speaker emulation (using convolution plug-ins) will need to be used to test the effectiveness of the model as a guitar amplifier.