The art and science of steganography are dedicated to concealing the presence of a secret message from a third party, such that only the sender and recipient are aware of its content. Various types of media can be used to conceal these communications. When information is hidden in an audio signal, this is referred to as audio steganography. In this paper, two distinct steganography techniques are combined with a multi-level steganography approach: the initial message is embedded in an audio cover at the initial stage, employing, a modified LSB technique, additionally, the second message is embedded in the output from the first level, using a phase coding approach at the second level. A stego audio file is the second level's output containing two audio covers with secret messages. The message is split in multiple ways, with varying proportions between the two levels, in order to investigate how the message's size affects the two procedures used here as well as the levels. The PSNR, MSE, and histogram metrics are used to compare the original and stego audio, in order to assess the effectiveness of the suggested approach. The optimum outcome is achieved when the message is divided in the ratio (1:1). The worst outcome is achieved when the message is divided in the ratio (3:1)