Look at the schematic and service manual closely. The 100mv is being measured across two sides (the PNP side and NPN side). Each side has 2 transistors in parallel. So in effect 50mv per side. And the way it measures the voltage is that each .5 ohm emitter resistor is in series with a 22 ohm resistor. If the service manual, instead of asking for measurement from pin 7 to pin 19 told you to measure from pin 10 to 13, or pin 8 to pin 13 this would measue the voltage directly across the Emitter resistor and I could understand a lot lower reading. The Spec 4 actually has the reading taken directly across the emitter resistor pairs in series, not through 22 ohm resistors. (and has you verify that the other pairs are OK, in that this is a similar 4 output trANSISTOR per channel (configured as a PNP and NPN Dalington).
As a cross check on the SX-1250, take voltage measurements between pin 8 to 10, 10 to 11 then pin 16 to 17, and 16 to 18. All of these measurements should read at least 0.5v foward bias on these transistors identifying them as conducting. If they are not reading at least 0.5v then the bias is too low.
But great to see other feedback!!!