With most DSLRs, metering is designed around electronic lenses that stay wide open for viewing/focusing, then stop down only at exposure. The camera knows the lens’s maximum aperture and selected aperture, so it can calculate exposure correctly.

With a non-electronic manual lens, the camera usually does not know:

• the lens’s maximum aperture
• the aperture you’ve set
• whether the lens is already stopped down

So the camera may meter under assumptions that are no longer true. You’re effectively doing stop-down metering, while the camera’s normal metering system expects open-aperture behavior.

That’s why Av often works better: the camera simply measures the light actually coming through the stopped-down lens and chooses a shutter speed to match. In M, the meter/preview behavior can be less reliable because the camera has no aperture data and may not simulate exposure the same way.

Also, stopping down a manual lens darkens the view/live view, which can make composition and preview harder.

Practical approach:

• use Av mode for easiest metering
• set aperture on the lens, let the camera choose shutter speed
• use exposure compensation or AE lock as needed
• in M, meter manually and check histogram/playback rather than trusting preview alone

So the missing electronic lens data—not just the light level—is the main reason metering is less accurate or less predictable.