Cost of flight at various speeds is a crucial determinant of flight behaviour in birds. Aerodynamic models, predicting that mechanical power (Pmech) varies with flight speed in a U-shaped manner, have been used together with an energy conversion factor (efficiency) to estimate metabolic power (Pmet). Despite few empirical studies, efficiency has been assumed constant across flight speeds at 23%. Ideally, efficiency should be estimated from measurements of both Pmech and Pmet in un-instrumented flight. Until recently, progress has been hampered by methodological constraints. The main aim of this study was to evaluate recently developed techniques and estimate flight efficiency across flight speeds. We used the 13C-labelled sodium bicarbonate method (NaBi) and particle image velocimetry (PIV) to measure Pmet and Pmech in blackcaps flying in a wind tunnel. We also cross-validated measurements made by NaBi with quantitative magnetic resonance (QMR) body composition analysis in yellow-rumped warblers. We found that Pmet estimated by NaBi was ∼12% lower than corresponding values estimated by QMR. Pmet varied in a U-shaped manner across flight speeds in blackcaps, but the pattern was not statistically significant. Pmech could only be reliably measured for two intermediate speeds and estimated efficiency ranged between 14% and 22% (combining the two speeds for raw and weight/lift-specific power, with and without correction for the ∼12% difference between NaBi and QMR), which were close to the currently used default value. We conclude that NaBi and PIV are viable techniques, allowing researchers to address some of the outstanding questions regarding bird flight energetics.