ROG Ally X speaker audio analysis:

The upper mids and treble >1 kHz have less controlled directivity than the Steam Deck, with the treble becoming uneven with even slight angular changes, so you need to point it in a specific direction to get stereo imaging to sound ideal, which fortunately happens to mean pointing a portable device at your face for the Ally X. The treble is also more ragged (https://files.catbox.moe/3d0tka.jpg) than the Deck's in general, but those dips and peaks are narrow and mostly get smoothed away by human hearing perception, potentially negatively affecting vocal timbre in some edge cases.

Otherwise, the entire range >1 kHz is surprisingly good, falling within the speaker gold standard of [+/-]3 dB in the ideal case even above 10 kHz. With the treble slope being flatter compared to the slight upwards rise of the Steam Deck and lack of peak in the upper treble >10 kHz to smear transients, I can understand why people would call the Ally X's sound superior to the Deck's. Having measured it with Dolby Atmos on, I can also understand why people call Dolby awful.

A tuning problem with the drivers would be a small +3 dB peak centered at around 2 kHz. It contributes to the Ally X's slightly harsh tonality, and while the peak adds the perception of extra clarity by pushing some sounds into the foreground, it does so by pushing others into the background. Most headphone/IEM manufacturers would kill to have such small tonal deviations, though.

The remaining flaws would be the result of inherent physical limitations. There's the usual half-wave resonance caused by the Ally X as a whole vibrating, which is centered at 640 Hz in this case, as would be predicted for a device about 10.5 inches wide. It has a significantly detrimental effect on the sound here, as it leads to a steep drop at 500 Hz that effectively cuts the lower midrange in half and creates a hole in the mids at 800 Hz. There's also the typical lack of bass/lower midrange extension caused by the speaker's small size. Ignoring the aforementioned resonance, the speaker's midrange slopes downwards at 1000 Hz, starts dropping significantly at 600 Hz, and falls off a cliff starting at 350 Hz, well above the bass range.