Encoding and decoding to loudspeakers in Ambisonics introduces interchannel crosstalk to channel-based material. This study investigates the perceptual impact of this phenomenon for different signals, ambisonic orders, decoders, and listening positions with respect to timbral and spatial quality aspects. Open-access recordings made with different microphone arrays, as well as noise signals are encoded in Ambisonics with different orders. Then they are decoded to a loudspeaker setup matching the channel-based layout using either the AllRAD approach or a sampling decoder. In a MUSHRA-like listening experiment for center and off-center listening positions, perceptual differences to a direct loudspeaker playback of the material are investigated.