a) that sound or television broadcast programmes can include interviews or other interactive situations which involve inserts shot at different physical locations, linked together at a base location;
b) that such inserts can be delayed due to propagation time or due to signal processing in codecs;
c) that Question ITU-R 35/6 – Tolerable round-trip time delay for sound programme and television broadcast programme inserts, has requested submissions on delay models for programme contribution loops and on tolerable limits for delay around such loops,
1 that the signal path model in Annex 1 should be used as a basis for modelling delay in programme contribution loops.
This Annex outlines a model for calculating signal delay in programme contribution loops and outlines system design guidelines and operational practices which can be used to minimize the loop delay, minimize the programme degradation arising from loop delay and echo as well as minimize the disturbance to the programme participants from loop delay and echo.
One half of the insert loop may be modelled as shown in Fig. 1.
Typical delays in the stages of this half-loop are shown in Table 1. Note that more than one row of Table 1 may apply to a given link e.g. calculation of delay for digital video with MPEG 2 encoding transmitted via an asynchronous transfer mode (ATM) carrier needs information from two rows of the Table. These values are examples only. Measured or calculated values for the system under study should be used to determine the actual delay in that system.
Typical delays of loop components using the half-loop model of Fig. 1*
Three general principles should be observed in designing and configuring facilities for programme contribution loops:
3.1 Number of encode/decode stages
The number of encode/decode processes within each half-loop should be minimized. This minimizes the encode/decode delay and has the added benefit of reducing other forms of signal degradation such as group delay and non-linear distortion.
3.2 Differential audio-video delay
The differential sound-to-picture delays should be kept within the limit of 25 ms to –100 ms as specified in Recommendation ITU-R BT.1359 – Relative timing of sound and vision for broadcasting. This may generally be done by judicious selection of audio and video codecs and by compensation with audio delay units if necessary.
3.3 Foldback leakage limit
Audio leakage between half-loops should be kept below the threshold indicated in Fig. 2. The 15 dB upper limit on each half-loop ensures that the loop has at least a substantial gain margin for stability and that the frequency response deviations from comb filtering are limited to less than 1.5 dB. The slope of the curve is derived from experimentally determined audibility thresholds. The –50 dB lower limit is set as the S/N for speech rarely exceeds 50 dB even under studio conditions. It may be necessary in some circumstances to reduce the leakage further than 50 dB, especially for long delays. The leakage limit should be applied to the worst-case sum of both the electric and the acoustic leakage.
No picture transmitted along the return half-loop should show a live monitor fed by the forward half-loop. If an on-screen monitor is required, either use a blue screen and key the picture in downstream OR (preferably) use a separate camera to feed the return half loop, with no monitor in shot.