A mobile network employing a so-called cloudbased architecture requires an additional fronthaul network to forward digitized samples. With increasingly dense deployments of small cells, the availability of high quality fiber fronthaul cannot be expected at all sites. Longer fronthaul distances or the employment of wireless technologies renders fronthaul links imperfect in the sense that they introduce bit-errors into the digital samples. In this work, we build upon so-called hardware noise resilient decoders to investigate decoders for imperfect fronthaul that can achieve very good performance even under high bit-error rates on the fronthaul.