I'll give you a good answer based on my experience. I personally worked with DSPs, FPGAs and uCs. DSPs are very good but it is expensive to give a start with DSPs. Most of the time you will have to buy the evaluation module (it is not easy to make one by yourself like for uCs), JTAG emulation module and Code Composer Studio (CCS) software (60-day trial in available). All these are not easy and worth to afford for a single project. This is where most of the people are struck. If there is an investor for you or your university is willing to invest for the project, we can offer you a suitable EVM, JTAG and CCS. But it would be not less than $4000 to start with. DSPs are made to perform advance calculations such as FFTs, DCTs, FIRs, IIRs, etc... DSPs mostly have multiple execution paths (parallel processing) to do several calculations at ones, not just one at a time like in most uCs.On some posts on this forum i got to know that it would be better to get some advice from you regarding fpga, dspic applications since you are knowledgeable on this topic. what i have to do fro a certain project is to grab audio or pressure using 6 piezo electric sensors as either audio pickups or force sensors process them and give an audio output. regarding the processing part the project (when considering the worst case scenario), it would require processing the audio from 6 piezo pickups (fft should be performed on all the inputs ) and then depending on the fourier spectrum a new 2 channel audio output is given. so would a dspic alone be capable of doing this or would an fpga dspic combination be cpable of doing this or would i have to consider DSP processors?
this is why i was asking whther you had fpga or dspic30f4011(this was recommended by a member on this forum for audio applications if you have any better component in your mind please do suggest).
If you take FPGAs, those are devices to do specific tasks. In other words, an FPGA is suitable to do a quick processing of output based on set of inputs. So for fixed algorithms such as FFTs, an FPGA can help. However FPGA is not a cheap and easy to make solution. If you like to work on an Altera FPGA, you can buy an evaluation module, byteblaster unit and use the free IDE to program them.
Finally the uCs (microcontrollers). There are two popular manufacturers, Microchip and Atmel. The main microcontroller product series is called PIC. dsPIC is kind of an advance microcontroller with Digital Signal Processing capabilities. Atmel is well known for stability and reliability. They have two popular product lines called ATTiny and ATMega.
Since you application is more biased on Digital Signal Processing (such as application of FFTs, etc..), the best I can suggest is the dsPIC. I don't think you need a dsPIC, FPGA combination since Audio processing is not heavy. So I think you better start with a dsPIC such as 30F4011 and when you think it needs some more power, then we could introduce a co-processor or FPGA to minimise the load on the main uC. But I'm sure if you write optimal codes, this will not be required.
