An image-taking apparatus is disclosed, which has an effective arrangement of sound input or output portion. The image-taking apparatus comprises a main body, a lens barrel provided on a front face side of the main body, an illumination unit which illuminates an object and is provided on the top face of the main body. The front face of the illumination unit is arranged at a position closer to the object than the front face of the main body. The apparatus further comprises a sound input/output portion. In a direction orthogonal to the bottom face of the main body, at least a part of the sound input/output portion is provided in an area between the periphery of the lens barrel and the illumination unit on the front face of the main body.