Reduce sync demodulation FFT work

[Josalo-III]

Checklist

• I have searched the open pull requests to confirm this change has not already been submitted.
• My branch is up to date with the target branch.
• I have tested my changes and all existing tests pass.
• Documentation is unchanged because this does not alter user-facing behavior or configuration.
• My code follows the project's coding standards (see CONTRIBUTING.md).

Description

Use the existing phase-adjusted real-FFT filter for sync-only demodulation.

Motivation

The sync-only path still performed a full FFT, full inverse FFT, and time-domain roll for a real-valued result. The main demodulation path already precomputes the equivalent positive-frequency filter with its sample offset applied in the frequency domain.

Related Issues

N/A

Changes Made

• Replace the sync path's full transform and inverse transform with rfft and irfft.
• Reuse the existing phase-adjusted 0.5 MHz half-spectrum.
• Extend the deterministic NTSC and PAL regression comparisons to cover sync-only demodulation.

Testing

• Full suite: nix develop --command python -m pytest -q — 84 passed.
• Focused tests: nix develop --command python -m pytest -q tests/test_demod_fft.py — 2 passed.
• Synthetic benchmark against the former implementation:
  • NTSC: 1.934 → 1.639 ms/block (15.3% lower)
  • PAL: 1.965 → 1.701 ms/block (13.4% lower)

Screenshots (if applicable)

N/A

Additional Notes

No dependency or output-format changes.