Yes - with an audio output transformer secondary, it's OK to 'ground one end' of the winding and put a cro probe to the other end to take voltage measurements.
Remembering that a CRO probe is 'internally grounded', one can simply put the CRO probe across the audio transformer winding without having explicitly grounding one end.
The process of grounding one end of the audio transformer winding makes it 'unbalanced' or 'single ended' and thus suitable to measure with a typical (non-differential) CRO probe.
Note that debalancing a transfomer winding and taking a voltage measurement will return a +6dB level shift over the 'notional' signal on each phase. ie. V single ended = V+diff - (V-diff) = 2x Vdiff
Of course, one can use the X and Y channels of a dual input CRO, set one channel to 'invert' and use the CRO's dual mode input of type (x+y) - this will function as a true differential probe with no grounding of the winding at all.
Also note that although making sensible 'phase-to-ground' measurements in a differential audio output is not guarranteed to show a pair of 'same-amplitude signals of opposite phase', it is often the case anyway - it depends on the circuit.
For example, a solid state balanced output's '+' and '-' connection signal will typically be symmetrical about ground, so it's fine to take the expedient single ended readings of each 'phase-to-ground' style.
This stuff is the kind of thing you can verify, for example, when 'lining up' or calibrating, an audio interface (dB) with a CRO (voltage).
ie. you want to measure voltage peak-to-peak on the CRO (or AC true rms voltmeter), and convert back and forth between RMS, dBu (600 ohm load) and so on.