keywords speaker recognition; speaker identification; emotions; stress conditions; data augmentation; synthetic stress