In this work we consider the problem of quantifying the detection latency of weak signals- A signal whose energy is much lower than the noise floor. We define detection latency as the total number of samples needed to correctly detect the weak signal. We characterize the latency by considering a single-input, single-output (SISO) wireless communications scenario where the receiver is tasked with detecting the transition from a channel probing waveform to a 500 kHz signaling tone. We consider constant modulus linear frequency modulated (LFM) chirp and the GPS L2C (M) pseudorandom noise waveforms for channel probing. We also introduce the notion of hypothesis tracking built on the sequential detection framework to reduce detection latency at the receiver. We compare the performance between two hypothesis tracking methods.