Audio and video testing focus

Voice call test point:

(1) echo

Echo refers to the voice that a user hears during a voice call (with a certain delay ).

In traditional PCs, users usually have headsets, and there is no echo problem.

When talking on a mobile phone, the echo problem becomes more prominent when the voice (receiver or speaker) is usually put out.

Therefore, the speech engine performs echo cancellation.

 

(2) Noise

The voice sender may experience noise in the voice due to background noise or mobile phone devices.

The Speech Engine implements noise reduction, minimizing noise and keeping the sound clear.

 

(3) latency

A certain delay occurs during network transmission or during jitter buffering.

A good engine should reduce latency as much as possible.

 

(4) Sound Quality

For example, whether the speaker's original tone is maintained, and whether the voice is clear and round.

 

(5) network bandwidth usage

The bandwidth used by voice transmission should not be too large. Otherwise, problems may occur in areas with poor mobile network conditions.

 

(6) CPU usage

CPU usage should be as small as possible to adapt to low-end mobile phones with poor hardware performance.

 

(7) Installation Package Size

Minimize the installation package size. Such as cropping unnecessary functions.

 

Video Call test point:

(1) Definition

The appearance of mosaic should be avoided or reduced as much as possible, and the Blur degree should be reduced in the area of image change.

 

(2) Smoothness

The actual frame rate cannot be too low. Otherwise, the frame rate may be choppy.

 

(3) latency

Large latency may occur when the video stream is too large or the network bandwidth is poor.

The engine can reduce the Latency by modifying parameters.

 

(4) shadow

When the image changes dramatically, there may be obvious shadow.

A good video engine should minimize the occurrence of shadow.

 

(5) Audio and Video Synchronization

The voice and the speaker's lip movements should be synchronized as much as possible to reduce the video lag time.

 

(6) network bandwidth usage

The bandwidth occupied by the video is significantly higher than that occupied by the voice. Therefore, this part should be optimized as much as possible. Otherwise, the video effect will be greatly affected.

 

(7) CPU usage

Similarly, the CPU usage of a video is significantly higher than that of a speech, especially for Codec and image processing.

This optimization is equally important.

 

(8) Installation Package Size

In the same voice part, the smaller the installation package, the better. Crop the package if necessary.

 

Because audio and video testing is more subjective, we recommend that you:

(1) Blind test

Under the same conditions, the evaluators can experience multiple products of the same type (including Coco ),

But don't tell them the product name and then let them compare different products.

 

(2) more testers make an appointment

Different people have different feelings due to their subjectivity.

Therefore, try to increase the number of testers so that they can score different test points and then take the average value.

 

(3) various environments

Perform tests in a variety of typical Call environments (including quiet and noise environments ).