jack_iodelay will create one input and one output port, and then measures the latency (signal delay) between them. For this to work, the output port must be connected to its input port. The measurement is accurate to a resolution of greater than 1 sample.
The expected use is to connect jack_iodelay's output port to a hardware playback port, then use a physical loopback cable from the corresponding hardware output connector to an input connector, and to connect that corresponding hardware capture port to jack_iodelay's input port. This creates a roundtrip that goes through any analog-to-digital and digital-to-analog converters that are present in the audio hardware.
Although the hardware loopback latency is the expected use, it is also possible to use jack_iodelay to measure the latency along any fully connected signal path, such as those involving other JACK clients.
Once jack_iodelay completes its measurement it will print the total latency it has detected. This will include the JACK buffer length in addition to any other latency in the signal path. It will continue to print the value every 0.5 seconds so that if you wish you can vary aspects of the signal path to see their effect on the measured latency.
If no incoming signal is detected from the input port, jack_iodelay will print
Signal below threshold... .
every second until this changes (e.g. until you establish the correct connections).
To use the value measured by jack_iodelay with the -I and -O arguments of a JACK backend (also called Input Latency and Output Latency in the setup dialog of qjackctl), you must subtract the JACK buffer size from the result. The buffer size is determined by multiplying the number of frames per period (given to the jackd backend by the -p or --period option) by the number of periods per buffer (given to the jackd backend by the -n or --nperiods option). Note that JACK2 will add an implicit additional period when using the default asynchronous mode, so for JACK1 or JACK2 in synchronous mode, the buffer size is n*p, but for JACK2 in asynchronous mode the buffer size is (n+1)*p. Once the JACK buffer size is subtracted from the measured latency, the result is the "extra" latency due to the interface hardware. Then, if you believe that the latency is equally distributed between the input and output parts of your audio hardware (extremely likely), divide the result by two and use that for input and output latency values. Doing this measurement will enable JACK clients that use the JACK latency API to accurately position/delay audio to keep signals synchronized even when there are inherent delays in the end-to-end signal pathways.
Originally written in C++ by Fons Adriaensen, ported to C by Torben Hohn.