Bedload transport assessment is important for geomorphological, engineering, and ecological studies of gravel-bed rivers. Bedload can be monitored at experimental stations that require expensive maintenance or by using portable traps, which allows measuring instantaneous transport rates but at a single point and at high costs and operational risks. The need for continuously measuring bedload intensity and dynamics has therefore increased the use and enhancement of surrogate methods. This paper reports on a set of flume experiments in which a Japanese acoustic pipe and an impact plate have been tested using four well-sorted and three poorly sorted sediment mixtures. Additional data were collected in a glacierized high-gradient Andean stream (Estero Morales) using a portable Bunte-type bedload sampler. Results show that the data provided by the acoustic pipe (which is amplified on 6 channels having different gains) can be calibrated for the grain size and for the intensity of transported sediments coarser than 9 mm (R-2 = 0.93 and 0.88, respectively). Even if the flume-based calibration is very robust, upscaling the calibration to field applications is more challenging, and the bedload intensity could be predicted better than the grain size of transported sediments (R-2 = 0.61 and 0.43, respectively). The inexpensive impact plate equipped with accelerometer could be calibrated for bedload intensity quite well in the flume but only poorly in the field (R-2 = 0.16) and could not provide information on the size of transported sediments. (C) 2015 Elsevier B.V. All rights reserved.