Large space launchers emit jet noise that is harmful to their payloads, and are typically mitigated by use of large quantities of water injected into their exhausts, typically ~6 times the exhaust mass flow rate. In the present work, this approach is comprehensively examined with the aim of optimizing the water injection strategy, by the use of (a) scale models of the launch-pad at different size levels and exhaust temperatures including actual rocket exhaust. Three regimes during the vehicle lift-off are distinct in their noise spectral identities: (i) the exhaust jets are fully inside the deflector ducts; (ii) the jets partially impinge on the launch pedestal; (iii) the jets behave as in a free field. Shock noise predominates at later times during lift-off. The surrounding sound field indicates peak acoustic radiation away from the vehicle along the jet deflectors, shielded by the launch-pad structures. Typically, a staged water injection is evaluated for the launch scenario: (i) on the deflecting jets beneath the launch pedestal and from the shielding surfaces; (ii) above the pedestal. The hot and cold tests deliver similar results for the case without water injection, indicating that the temperature does not influence the aerodynamic sound generation in such high supersonic jets (M > 3). With water injection, hot jets were relatively quite distinct, particularly on the effect of injection pressure and flow rate. Hot jets registered higher sound level reduction due to possible influence of temperature on the water jet breakup and evaporation in the jet. Around 6-8 dB reduction is achieved with the 1:12.8 scale model involving actual rocket exhausts with a water-toexhaust ratio of ~4. Copyright © (2011) by the International Institute of Acoustics & Vibration.