You probably mean these pdf and Word-documents:
http://www.subsim.com/radioroom/down...o=file&id=3402 [EDIT: Sailor Steve assigned a different download page to it. Corrected it. ]
This application is I think actually more usefull. Because it is a pain to select and make a list of all the units that fit to the frequencies that you find, by hand!.
http://www.subsim.com/radioroom/showthread.php?t=82476
(the link is dead so I uploaded it from my hardrive:
http://www.gamefront.com/files/20897758)
I uploaded it just now to Gamefront, and asked Sailor Steve to add it to the Subsim.com download section. [EDIT: now at Subsim:
http://www.subsim.com/radioroom/down...=file&id=3403]
Basically the idea is to measure with the (mouse) cursor in the waterfall window (depends on sub which display it is, upper middle or bottom) which frequencies you find in the direction of the sensor/contact. There could be multiple contacts overlapping eachother and so you could find more frequencies that are near to one another but are from distinct units. (like 320 hertz next to 340 hz, as there are no units that emit these 2 frequencies at the same time. Each unit emits 1 specific frequency in 5 bands). You could start with the lowest frequencies but then you'll get a large list of possible units classes. Instead I would start with the highest visible frequency that the sensor detects. (You may need to increase your brightness/contrast/gamma settings of your videocard/monitor) That will narrow it down tremendously. Then wiggle the sensor direction a bit and see which lower freqencies fade out in sync with the higher frequency. These likely belong together. (Unless the different units are really close together bearingwise.) Noting which frequencies fade together limits the amount of possibilities even more. Then if you still have a list, instead of a definate identification of class, then I would compare the frequencies received of the other sonar sensors. Bow sonar doesn't detect frequencies the towed array is capable of detecting. Assuming you are close enough for them to be detectable.
And if it is still not definate you should start to combine it with ping frequencies of active sonar intercept, and/or ESM receptions.