Abstract
This paper discusses the application of voice recognition for Egyptian Radio and Television Union's cloud. It is used as secure access to the cloud by authorized group (AuthGs). The voice of each member from AuthGs is watermarked using singular value decomposition and then encrypted by Chaotic map. The results are transmitted through channel under several conditions and received at the receiver side, and then the recognition rate is calculated for various extraction of watermarking after decryption method. It is done to find out the suitable technique for AuthGs to access the cloud and insure the security and privacy of the cloud. Many tests are performed to compare between the voice before and after process to grantee the high robustness of the signal from illegal eavesdrops or any abuse behaviour. The feature extraction is performed using artificial neural network to store it in a database to compare with.