A new formalism to analyze peculiar velocity surveys is presented. Results from these surveys are shown to be dominated by small-scale noise, aliasing, and incomplete cancellations. The formalism allows us to filter out the signal from scales that are not of interest and thus provides us with a clean signal that probe large scales. We use maximum likelihood techniques to analyze the filtered data and compare it to the analysis of the full dataset. The filtered analysis gives a much better parameter estimation than the full analysis, leading us to conclude that, indeed, the large-scale signal is obscured by small-scale noise.