To treat data from temperature modulated differential scanning calorimetry (TMDSC) in terms of complex heat capacity, one should know heat transfer and apparatus influences on experimental results. Concise theoretical model for power compensated DSC is given, which includes description of most important heat transfer paths and apparatus effects. Estimations for real experimental conditions are made. An algorithm to determine parameters for calibration to exclude heat transfer and apparatus effects is proposed. This allows to increase upper frequency limit and accuracy of TMDSC measurements. Based on theoretical results the key points of the correction and some experimental evidence are discussed.