The bypass ratio of newer turbofan engines has been increasing steadily and has reached $10; even higher bypass ratios are being considered for improving aircraft fuel efficiency. An accurate method for the prediction of jet noise over a wide range of bypass ratio, from $5 to ultra-high bypass ratios ($20), is required to address current and future needs. The main objective of the current study is the development of a procedure for real-world application that would permit the accurate prediction of jet noise, which in turn would enable the quantification of the non-jet noise component. A new empirical method for prediction of noise from realistic dual-stream jets is developed and validated against an extensive database acquired at model scale; the range of validity covers a velocity ratio (V s /V p) ! $0.6, as this is the range of interest in real turbofan engines. Accurate absolute spectral predictions at all angles and all cycle conditions are first demonstrated at model scale. The same method is then extended to full-scale predictions, both from static engine tests and airplane flyover tests. The range of Strouhal number of interest in full-scale tests spans $0.1 to $100. Nearfield effects on the jet noise from dual-stream jets have been quantified and a simple and improved procedure for incorporating spectral effects has been developed. Accurate spectral predictions are obtained for noise from static engine tests and flyover tests, over a wide range of engine operating conditions and bypass ratio. Therefore, the good quality and accuracy of the new prediction method have been confirmed at both model scale and full scale, with and without forward flight. An absolute prediction takes $1 s on a workstation, as only a single scaling equation is utilized. The application of the new prediction method in gaining better quantification and understanding of the non-jet noise components, and their reduction, is described.