Text this: WaViT-CDC: Wavelet Vision Transformer With Central Difference Convolutions for Spatial-Frequency Deepfake Detection