I am planning to build a sound I/O system for a PC, and I am interested in a review of the design. If you are interested, could you please look this over? Thank you.
---------------------------------------------------- This is for a system that uses a PC, custom software (written by me), a hardware interface (to be built by a specialist, not me), and one or two hydrophones (off-the-shelf). In the one-hydrophone option, the hydrophone is to be used for input and output. In the dual-hydrophone option, one h-phone is for input and the other is for output. Somewhat a priori numbers: Usable frequency range (underwater acoustics): 1 to 100 kHz For Input: Voltage sampling = 200 kS/s (S/s = Samples per second) Voltage array size = 1024 samples Voltage sample size = 16 bits For output, the application sends a voltage sample to the hardware at the voltage rate, 200 kS/s, and the hardware converts this voltage number into the analog equivalent. If a single hydrophone is used, then the input process may be disabled during the output. The connection standard used in and out of the PC = USB version 2 (max device speed = 1500 kBytes/s) The PC application needs to get a voltage array at each 100th of a second, i.e., the frequency sampling rate is 0.01 second. Frequency Sampling Rate = 100 F/s (F/s = Frequency samples per second). A frequency sample is a single number which is the dominant frequency resulting from an FFT calculation on a voltage array. The frequency sample is done by the software of the application (it is already written). The two input frequency sampling options are these: 1) Continuous feeding of voltage numbers to PC memory: The A/D hardware sends to the PC voltage samples continuously at the rate of 200 kS/s. The voltage samples are stored in a PC memory buffer which is read by the application once every 100th of a second. The data traffic between the PC and the A/D = 400 kB/s = 200 * 2 = voltage sampling * voltage sample size 2) Selective voltage numbers to PC: The A/D hardware sends to the PC a voltage array at each 100th of a second (frequency sampling rate = 0.01 sec), either automatically or when requested by the PC (at each 100th of a second). The data traffic between the PC and the A/D = 200 kB/s = 1024 * 2 * 100 = voltage array size * voltage sample size * frequency sampling The voltage samples are stored in a PC memory buffer which is read by the application once every 100th of a second. I would be grateful for any comments. For example, which option is the most appropriate or effective, or which option is the least expensive (if both options are equivalent in effectiveness). I would also welcome suggestions on who could manufacture the A/D hardware or who already sells such hardware, off-the-shelf, and compliant with these requirements, with a preference for Option 2) Selective voltage numbers to PC, and a single hydrophone instead of 2. thanks serge masse [EMAIL PROTECTED] http://leafy.dev.java.net/
