Would you say the same about (for example) a Leica S2 or other 'medium format' digital camera as compared with 24x36mm?
Cheers,
R.
No, I wouldn't.
The lenses for medium format cameras combined with the much larger sensor surface areas present a clear advantage when it comes to recording the analog signals from the sensor.
The ratio of sensor areas for APS-C (DX), 24 x 36 and the S2 sensors are:
1.0 : 2.34 : 3.67
So an APS-C lens would have to deliver almost 4 times as much light if the S2 sensor's performance (full-well capacity, quantum efficiency and inherent read noise at base ISO) was equal to the APS-C sensor cameras'. A lens for an APS-C camera would have to be large, heavy and expensive to deliver that much light when subject perspective is held constant.
By contrast some APS-C lenses of a reasonable size and cost can deliver about twice as much light to the sensor as lenses for 24 X 36 mm cameras. A recent example is the excellent Fuji 14/2.8 XF lens which is approximately equivalent to 20/4 lenses for 24 x 36 mm cameras. Of course a 35/1.4 lens for a 24 x 36 mm camera will provide an advantage compared to a 23/1.4 lens on an APS-C camera.
The lens' T-factors, the absence of AA filters and varying read noise characteristics of different generation cameras make direct comparisons of maximum signal capacity complicated. Shadow noise behaves differently in different camera data stream designs too. But a factor of 3.67 more area combined with longer focal length lenses (when subject perspective is held constant) is a significant advantage. I mention the longer focal lengths because fast, short focal-length lenses are expensive to design and manufacture.
Because the signal from the sensor is an analog DC voltage, its signal-to-noise ratio is the first factor in the data stream that affects image quality. This is why SNR is important.
Many other factors affect how those gallery prins' aesthetics. But this does not diminish the impact of SNR and dynamic range on what happens after the signals leave the sensor sites.