摘要： Measuring the temporal shape of ultrashort laser pulses is an important standard task in ultrafast optics. A multitude of methods has been devised for that purpose [1]. However, in many cases a specific, iterative algorithm has to be applied to retrieve the pulse shape from the measurement. In this work we present a common pulse retrieval algorithm (COPRA) that can be universally applied to such pulse measurement methods and compares favorably in terms of speed and accuracy to existing approaches [2]. For example, COPRA can be applied to the well-established frequency-resolved optical gating (FROG) [3], as well as, more recent techniques, such as dispersion scan (d-scan) [4] or time-domain ptychography (TDP) [5] (see Fig. 1). Generally, it is designed for a broad class of measurements where sets of pulse spectra after a nonlinear process such as second-harmonic generation (SHG) or third-harmonic generation (THG) are measured. Specifically, the measurement process has to be tunable by some parameter δ that forms the second measurement dimension. We call this class of measurements parametrized nonlinear process spectra (PNPS). COPRA now finds the pulse spectrum ?E(ω) that matches best to a PNPS measurement T meas(ω, δ ) in the least-squares sense. It does so in two stages: first the local iteration subsequently optimizes ?E(ω) with respect to single spectra, then the global iteration optimizes ?E(ω) with respect to the entire measurement. COPRA usually finds a least-squares solution to the pulse retrieval problem within a few hundreds iterations, i.e., within few tens of seconds on a standard workstation for usual measurement dimensions. This compares very favorably to existing approaches. Additionally, the retrieved pulses are more robust in the presence of Gaussian measurement noise than those retrieved by generalized projections or ptychography. Many existing pulse measurement schemes can directly benefit from COPRA for more robust and faster retrieval. Also, the framework of PNPS measurements may allow to gain more insight on fundamental questions of pulse measurement, will make measurement schemes more comparable and may lead to new and improved methods.