feat: add Sample::bits_per_sample()

[roderickvd]

We already had Sample::sample_size() which returns the size in bytes of a format. This does not always correspond with the number of bits. For example, I24 is stored in 4 bytes but with 24 bits storing audio, not 32. This PR introduces a new function Sample::bits_per_sample() to do that.

For F32 or F64 we could argue what it should return. This method returns 32/64 bits of bit depth (storage) instead of 24/58 bits of precision (effective resolution).

[Copilot]

Pull Request Overview

This PR adds a new method bits_per_sample() to the SampleFormat enum to distinguish between storage size and actual bit depth. The key motivation is that some formats like I24 use 4 bytes for storage but only 24 bits for audio data.

• Adds SampleFormat::bits_per_sample() method that returns the actual audio bit depth
• Refactors sample_size() method to use explicit type-specific mem::size_of calls instead of grouped patterns
• Updates documentation to clarify the distinction between quantized sampling depth and storage formats

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/samples_formats.rs	Adds new bits_per_sample() method and refactors sample_size() for clarity
CHANGELOG.md	Documents the new Sample::bits_per_sample method

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}