SVSNet+: Enhancing Speaker Voice Similarity Assessment Models with Representations from Speech Foundation Models