How to calculate a unified continuous zoom range using only physical cameras (like the default camera app)?

[stopkite]

Background

The default camera app provides a seamless continuous zoom experience from ultra-wide → wide → max zoom. The starting zoom ratio varies by device — some start at 0.5x, others at 0.6x, etc.
Notably, even on devices that don't support a logical multi-camera, the default camera app still provides this smooth transition between physical cameras.
I want to replicate this exact behavior in my app using the Camera2 API.

Goal

Combine the zoom ranges of multiple physical cameras into a single continuous zoom range — starting from the correct device-specific starting ratio (0.5x, 0.6x, etc.) up to the maximum zoom — just like the default camera app does.

Example

Physical Camera 0 (Ultra-wide): focal length 2.0mm, zoom range 1.0x–8.0x
Physical Camera 1 (Wide):       focal length 5.0mm, zoom range 1.0x–10.0x

Default camera app behavior:
0.5x (or 0.6x, etc.) ────── 1.0x ────── max
[Ultra-wide camera zone]      [Wide camera zone]

→ Starting ratio varies by device
→ Smooth transition to wide camera around 1.0x

Question

1. How is the starting ratio (0.5x, 0.6x, etc.) calculated?

• Is it the focal length ratio of ultra-wide / wide? (e.g., 2.0mm / 5.0mm = 0.4x?)
• Is the reason it differs per device simply due to different focal lengths?

2. What is the formula to calculate the switch point (crossover) between physical cameras?
3. How should the remaining zoom range of the wide camera be mapped into the unified range after the switch point?
4. Since the default camera app handles this even without a logical multi-camera, is there a recommended approach at the Camera2 API level, or is manual calculation the only

Environment

• Camera2 API directly (not CameraX)
• Targeting devices without logical multi-camera support

[MatiwosKebede]

The key point is that the zoom ratio shown by the default camera app is not the native zoom range of each physical camera. It is a virtual zoom space created by combining focal lengths and sensor crop factors.

For each physical camera, first calculate its equivalent 35mm focal length:

equivalentFocalLength =
focalLength * (36mm / sensorWidth)

Then choose the reference camera (usually the main wide camera). The zoom ratio can be represented as:

zoomRatio =
equivalentFocalLength / wideCameraEquivalentFocalLength

Example:

Ultra-wide:
2mm focal length → around 0.5x compared to 1x wide

Wide:
5mm focal length → 1x reference

Telephoto:
10mm focal length → 2x reference

So the unified range can be created by mapping each physical camera into the same zoom coordinate system:

Ultra-wide:
0.5x → 1.0x

Wide:
1.0x → 3.0x

Telephoto:
3.0x → maximum

For smooth transitions:

1. Detect all physical cameras:

• LENS_INFO_AVAILABLE_FOCAL_LENGTHS
• SENSOR_INFO_PHYSICAL_SIZE
• SENSOR_INFO_ACTIVE_ARRAY_SIZE

2. Calculate the minimum zoom ratio:
   minZoom =
   shortest focal length / reference wide focal length

This gives device-specific starting values like 0.5x or 0.6x.

3. Create zoom segments:
   Each physical camera owns a range in the unified zoom space.

Example:

[
{
cameraId: ultraWide,
start: 0.5,
end: 1.0
},
{
cameraId: wide,
start: 1.0,
end: 3.0
}
]

4. During zoom:

• Use cropRegion inside one camera.
• When reaching the transition point, switch physical camera.
• Apply overlap blending or gradual exposure/focus adjustment to hide the switch.

One important detail: the default camera app usually does not simply switch lenses. It also compensates:

• exposure
• white balance
• focus distance
• stabilization
• image processing differences

So matching the exact stock camera behavior requires reproducing both the zoom mapping and the camera switching pipeline.

For a computer vision app, I would keep the physical camera fixed when possible because switching sensors changes image characteristics and can affect detection accuracy.