older disk-based cmix instruments
superceded by rtgetin, rtaddout and rtbaddout and
Here is the original documentation:
These routines will read or write a block of samples from or to the disk. The arrays ‘input’ and ‘output’ are in floating point form, even if the disk file is in 16-bit integer form. ‘fno’ is the Cmix unit number for the file being accessed, and ‘size’ is the total number of samples in the array. Note that ‘size’ is NOT the number of samples per channel, but rather the total number of samples. The routines act like their sample-by-sample namesakes: blayout() and LAYOUT(), baddout() and ADDOUT(), bwipeout() and WIPEOUT(). They do not return a value, however, but rather the named array, loaded with samples, kicked up by ‘size’ from the last call. As with the other routines, they are positioned by setnote() and cleaned up by endnote. They are considerably faster than these units, however, and should be used when optimizing code, perhaps in conjunction with the block unit generators.