Measuring End-to-End Delay Times in Digital Video Systems

A video processing system’s audio-visual delay time depends not only on the MPEG compression and (de)coding equipment, but also on other components in the signal chain, such as noise reducers, multiplexers, demultiplexers, and so on. The modulation method and the transmission channel (such as satellite or cable) do not influence the delay difference between audio and video signal reception.

Delay Time Measurements

The CCIR (Comité Consultatif International des Radiocommunications) recommends that the time difference between the signal’s sound and vision components should not exceed 20 ms if the sound is advanced with respect to the picture, or 40 ms if the sound is delayed with respect to the picture. But in practice, the Digital Video Systems Division at Philips Eindhoven requires that its own equipment has a maximum delay of about 10 ms, which allows for delays that other components in the signal processing chain introduce.

You can measure audio-video delay simply by applying a test signal to the input of the video compression system. Taking a measurement after the video decoder reveals whether the delay between the audio and video components across the system is within specification. When you have characterised the delay of a particular system, you can add a correction factor within the broadcast signal chain. This correction usually entails passing the audio signal through a delay line.

To measure audio-visual delay, apply a test signal to the input of the system under test. A suitable test signal comprises a black video signal with four green video fields (two frames) at 20-ms intervals that repeat every 2 s, together with an audio “beep”. Apply the audio and video components from the decoder’s output to a digital oscilloscope, such as Fluke’s ScopeMeter 123. Configure the scope to trigger on the audio burst, and capture the audio and video signals simultaneously. You can see the audio-video delay as the time difference on the horizontal axis between the rising flanks of the first video frame and the audio burst. Figure 1 shows a typical example, where you can see the four video frames in the upper signal and the audio burst in the lower signal.

Figure 1. The audio-visual delay test signal as it appears on a Fluke ScopeMeter’s screen.

Using a DSO

Using a digital storage scope (DSO) permits more accurate between-cursor measurements and greatly simplifies your documentation tasks. You can download measurements to a PC running dedicated analysis software such as the “FlukeView” package (Figure 2). Set the cursor to the top of the rising flanks of the audio and the video signal, and the software automatically calculates the difference, displaying results in a box alongside cursor values. The example shows the audio-visual delay time to be 10.4 ms, well within the CCIR’s recommendation.

Figure 2. You can make accurate cursor measurements with analysis software such as the FlukeView package that runs on a PC.

A DSO also suits this type of measurement because an analogue scope cannot display low repetition-rate signals clearly. But you need to consider the specification that you require from a DSO for making relatively low bandwidth measurements. For example, typical DSOs simply adapt their sampling rate to the timebase setting, adjusting the number of samples to the number of pixels per division (about 25 samples per division, or 25 samples/20 ms). For audio-visual delay measurements, these low sampling rates result in waveform aliasing. When such aliasing occurs it’s impossible to measure the delay with any accuracy.

You could use a high-bandwidth, long-memory scope, but its expense usually limits such devices to the R&D lab. However, DSOs such as Fluke’s 20-MHz ScopeMeter maintain a constant sampling rate over a wide timebase range. Even at a timebase setting as low as 20 ms, this instrument samples at 5 Msamples/s — some 4000 times higher than a typical DSO. Rapid sampling with peak detection also lets you see the maximum and minimum values for every pixel on screen, displaying an envelope waveform similar to the example shown in Figure 1.

For more information, contact Fluke Europe B.V., P.O. Box 1186, 5602 BD Eindhoven, The Netherlands, +31-40-2-678-213, fax +31-40-2-678-210.