RTCRtpStreamStats.qpSum - Web APIs 编辑
The qpSum
property of the RTCRtpStreamStats
dictionary is a value generated by adding the Quantization Parameter (QP) values for every frame sent or received to date on the video track corresponding to this RTCRtpStreamStats
object. In general, the higher this number is, the more heavily compressed the video data is.
Syntax
var qpSum = RTCRtpStreamStats.qpSum;
Value
An unsigned 64-bit integer value which indicates the sum of the quantization parameter (QP) value for every frame sent or received so far on the track described by the RTCRtpStreamStats
object. Since the value of QP is typically larger to indicate higher compression factors, the larger this sum is, the more heavily compressed the stream generally has been.
Note: This value is only available for video media.
Usage notes
Quantization is the process of applying lossy compression to a range of values, resulting in a single quantum value. This value takes the place of the range of values, thereby reducing the number of different values that appear in the overall data set, making the data more compressible. The quantization process and the amount of compression can be controlled using one or more parameters.
It's important to keep in mind that the value of QP can change periodically—even every frame—so it's difficult to know for certain how substantial the compression is. The best you can do is make an estimate. You can, for example, use the value of RTCReceivedRtpStreamStats.framesDecoded
if receiving the media or RTCSentRtpStreamStats.framesEncoded
if sending it to get the number of frames handled so far, and compute an average from there. See Calculating average quantization below for a function that does this.
Also, the exact meaning of the QP value depends on the codec being used. For example, for the VP8 codec, the QP value can be anywhere from 1 to 127 and is found in the frame header element "y_ac_qi"
, whose value is defined in RFC 6386: 19.2. H.264 uses a QP which ranges from 0 to 51; in this case, it's an index used to derive a scaling matrix used during the quantization process. Additionally, QP is not likely to be the only parameter the codec uses to adjust the compression. See the individual codec specifications for details.
Example
Calculating average quantization
The calculateAverageQP()
function shown below computes the average QP for the given RTCRtpStreamStats
object, returning 0 if the object doesn't describe an RTP stream.
function calculateAverageQP(stats) {
let frameCount = 0;
switch(stats.type) {
case "inbound-rtp":
case "remote-inbound-rtp":
frameCount = stats.framesDecoded;
break;
case "outbound-rtp":
case "remote-outbound-rtp":
frameCount = stats.framesEncoded;
break;
default:
return 0;
}
return status.qpSum / frameCount;
}
Specifications
Specification | Status | Comment |
---|---|---|
Identifiers for WebRTC's Statistics API The definition of 'RTCInboundRtpStreamStats: qpSum' in that specification. | Candidate Recommendation | Initial definition. |
Identifiers for WebRTC's Statistics API The definition of 'RTCOutboundRtpStreamStats: qpSum' in that specification. | Candidate Recommendation | Initial definition. |
Browser compatibility
BCD tables only load in the browser
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论