I'm not sure how relevant this is, but in the context of noise figures for receivers, I have noticed in the past that Agilent (in measurement application notes) that SSB NF is 3dB lower than DSB NF, with the my interpretation as if one only includes SSB NOISE, then take DSB NOISE, then the DSB noise will be 3 dB higher (if symetrical).

In Cadence tools like spectreRF (and Razavi's "RF microelectronics"), they describe from the point of viewpoint of a signal. If a signal exists in both sidebands of an oscillator used in RF downconversion (e.g.  ZIF), then this is considered DSB as the signal exists on both sidebands... so the resultant noise figure is 3 dB lower than the SSB signal case. In the SSB signal case, there is no signal at the image, but the noise does exist, so when down converted it ends up increasing the receiver NF by 3dB.

So there seemed to be a different definition for SSB and DSB NF between Agilent and Cadence tools... so probably best thing to do is find out what ELDO is doing and read you aglient apps notes, and be clear to yourself what each tool is doing and what figure of merit is relevant to you...