1

Topic: Marking of video of audio of the data

The people! Prompt  where to dig and what to esteem for implementation of the following thing: Is RAW video (YUV) and audio (PCM) flows, it is necessary to add markers of video of frames or audio of units in flows, then flows press close with loss given mp2/mp4/h264/h265/aac/ac3 codecs, are then decoded reversely in RAW the data, and from this data it is necessary to pull out markers. That found that is used for content or comparing identification. And it is necessary for me to pull out the binary data, the small size 128/64-, but without loss of the data. Quality of coding I can influence. Imported distortions - without problems, but the person should have possibility to understand the changed flows. What advise? . The ideas are, but probably the bicycle is invented, and it probably to use

2

Re: Marking of video of audio of the data

Describe the task in other words, and that is not clear if you push them in  then cannot enkodit-dekodit if will not alter certainly and the codec the data which needs be pushed and got a bin should to be lost-free? Then without fec not to manage not clearly where as well as what for you plan to apply it but a format mpegts and everyones rtp protocols are invented for a long time

3

Re: Marking of video of audio of the data

Hello, gwg-605, you wrote: G6> That advise? I would write as subtitles which times-data to a separate file.

4

Re: Marking of video of audio of the data

The ultimate goal is not absolutely clear. But here pair of ideas: 1) watermarking - is more detailed in Google 2) to use user data which is defined in many main standards 3) to store the information at container level

5

Re: Marking of video of audio of the data

R> describe the task in other words, and that is not clear To eat system  video/audio of the data. It is necessary to check in an automatic mode that output data  input (there is passed data/duplikatov,  no data, synchronization is not broken and ) those markers should be allocated in . To transfer them parallely with  it is not meaningful. Markers should be steady against coding-> to decoding and the elementary transforms (resolution, loudness). To begin with planned to use a synthetic material, but further wanted to test on a real content.

6

Re: Marking of video of audio of the data

Hello, DiPaolo, you wrote: DP> the ultimate goal is not absolutely clear. Here aim more in detail: http://rsdn.org/forum/media/6852488.1 the Author: gwg-605 Date: 27.07 23:31 But here pair of ideas: DP> 1) watermarking - is more detailed in Thanks Google! It missed for some reason.

7

Re: Marking of video of audio of the data

Hello, gwg-605, you wrote: R>> describe the task in other words, and that is not clear G6> There is a system  video/audio of the data. It is necessary to check in an automatic mode that output data  input (there is passed data/duplikatov,  no data, synchronization is not broken and ) those markers should be allocated in . To transfer them parallely with  it is not meaningful. Markers should be steady against coding-> to decoding and the elementary transforms (resolution, loudness). To begin with planned to use a synthetic material, but further wanted to test on a real content.   becomes locally once? Or then  the data is transferred where that on a network and further them who that should ? I do not remember in known codecs which you would name that it was possible to push also meta data and mp2 and mp4 it already containers i.e. To add in  a format the labels what that you cannot in any way unless rewrite the code of all codecs and if receiving parties are not under your control, type of any standard decisions they also cannot   betray yours parallely it is variant push together c video-audio the data and  the data in one container of type mp2 mp4 mpegts on study standard decisions of the same type ffmpeg, that bicycle which you can try to invent is already invented  for a long time

8

Re: Marking of video of audio of the data

G6> There is a system  video/audio of the data. It is necessary to check in an automatic mode that output data  input (there is passed data/duplikatov,  no data, synchronization is not broken and ) those markers should be allocated in . To transfer them parallely with  it is not meaningful. Markers should be steady against coding-> to decoding and the elementary transforms (resolution, loudness). To begin with planned to use a synthetic material, but further wanted to test on a real content. Well in the elementary case if only fact is important, there are problems or not it is possible to use any metrics of quality, for example, PSNR. It shows, if frames are confused, passed etc. For more exact analysis it is necessary to approach more inventively in  from your specific conditions and requirements. There are variants in  more low more in detail. Still the important point - if at you at transcoding is change  and-or change of frequency for audio it introduces additional difficulties for the analysis.