In the context of distributed joint source–channel coding, we conceive reduced-complexity turbo trellis coded modulation (TTCM)-aided syndrome-based block decoding for estimating the cross-over probability $p_{e}$ of the binary symmetric channel, which models the correlation between a pair of sources. Our joint decoder achieves an accurate correlation estimation for varying correlation coefficients at 3 dB lower SNR, than conventional TTCM decoder, despite its considerable complexity reduction.