This paper introduces a universal distortion function called UNIWARD (UNiversal WAvelet Relative Distortion) for steganography in arbitrary domains. The function is designed to embed messages in a way that minimizes detectability by focusing on areas of the cover image that are difficult to model, such as textures and noisy regions, while avoiding smooth regions and clean edges. The distortion is computed as a sum of relative changes of coefficients in a directional filter bank decomposition of the cover image. UNIWARD is applicable for embedding in the spatial domain, JPEG domain, and side-informed JPEG domain. The function is simple and suitable for embedding in any domain, as it uses wavelet domain representations of the image. The paper demonstrates that steganographic methods using UNIWARD match or outperform the current state of the art in the spatial domain, JPEG domain, and side-informed JPEG domain. The authors also address technical issues related to zero embedding costs and propose a solution to mitigate the effects of 1/2-coefficients in the DCT domain. The paper concludes that UNIWARD provides a robust and effective method for steganography in arbitrary domains.This paper introduces a universal distortion function called UNIWARD (UNiversal WAvelet Relative Distortion) for steganography in arbitrary domains. The function is designed to embed messages in a way that minimizes detectability by focusing on areas of the cover image that are difficult to model, such as textures and noisy regions, while avoiding smooth regions and clean edges. The distortion is computed as a sum of relative changes of coefficients in a directional filter bank decomposition of the cover image. UNIWARD is applicable for embedding in the spatial domain, JPEG domain, and side-informed JPEG domain. The function is simple and suitable for embedding in any domain, as it uses wavelet domain representations of the image. The paper demonstrates that steganographic methods using UNIWARD match or outperform the current state of the art in the spatial domain, JPEG domain, and side-informed JPEG domain. The authors also address technical issues related to zero embedding costs and propose a solution to mitigate the effects of 1/2-coefficients in the DCT domain. The paper concludes that UNIWARD provides a robust and effective method for steganography in arbitrary domains.