readme.txt

各种视频压缩格式的网络传输的模拟信道分析

/*
 ==========================================================================================
 ITU-T Telecommunications Standardization Sector      Document:   VCEG-N80
 Study Group 16 Question 6                            Filename:   README.TXT
 Video Coding Experts Group (VCEG)                    Generated:  13 September, 2001
 ----------------------------------------
 Fourteenth Meeting: Santa Barbara, CA, USA, 21-24 September, 2001


 Intention: 
 ~~~~~~~~~~~~
 Simple offline software simulator for RTP/IP over 3GPP/3GPP2 bearers


 Source:
 ~~~~~~~
 Thomas Stockhammer, Guenther Liebl                  Tel:   [phone]
 Institute for Communications Engineering            Fax:   [phone]
 Munich University of Technology                     Email: {stockhammer,liebl}@ei.tum.de
 80290 Munich, Germany
 ==========================================================================================
*/


1. Content

This archive includes the following files, which are part of the offline software simulator
distribution:

- README (this file): contains all information necessary to compile and run the simulation tool

- Makefile: contains all the rules to compile the tool on a Linux workstation

- global.h: header file containing global definitions common to all source modules

- mobile_ip.cpp: primary source module

- loss_generator.h: header file containg the type and class definitions for the
                             3G loss generator module

- loss_generator.cpp: source module containing the explicit loss generator routines

- packet_data_agent.h: header file containing the type and class definitions for the
                                PDCP/PPP layer functionality

- packet_data_agent.cpp: source module containing the explicit PDCP/PPP layer routines

- radio_link_agent.h: header file containing the type and class definitions for the 
                               RLC/RLP layer functionality

- radio_link_agent.cpp: source module containing the explicit RLC/RLP layer routines

- settings_template.txt: sample parameter file containing the correct syntax to supply the simulation
                     tool with all the required parameter settings

- 18681.3, 18681.4, wcdma_64kb_3kph_5e-04.bin, wcdma_64kb_50kph_2e-04.bin, 
  wcdma_128kb_3kph_5e-04.bin, wcdma_128kb_50kph_2e-04.bin: error pattern files


2. Compilation: 

UNIX:
After unpacking the archive, just type 

>make

at the command prompt, and the tool will compile itself
according to the given Makefile. On other Unix-architectures than Linux machines, you probably
have to change some settings regarding the compiler in the Makefile.
Successful compilation will result in an executable called <mobile_ip>.


WINDOWS:
After unpacking the archive, create a new workspace in your favorite C++ developer plattform and add
all the source and header files.
Use the built-in routines of the developer environment to produce an executable called



3. Running the tool:

3.1 The parameter file

Before you can actually start the simulation tool, you have to create a suitable parameter file 
containing the desired settings for a specific simulation run. The easiest way to do this
is to just copy the given sample file <settings_template.txt> to your own parameter file (which we
will simply call <parameter-file> in the following) and edit it afterwards. In the following,
you will find a list of all possible settings together with the required type of the input values:


ENTRY								       INPUT VALUE TYPE			CLASS		   
-------------------------------------------------------------------------------------------------------------------------------------
file for writing simulation log report					STRING				mandatory
file for writing final statistics					STRING				mandatory	

file containing the bit error pattern					STRING				mandatory
start position in the pattern (in bytes)				POSITIVE INTEGER		mandatory

file containing the input RTP stream					STRING				mandatory
file containing the output RTP stream					STRING				mandatory
3GPP/3GPP2 system type (UMTS or CDMA-2000)				STRING				mandatory
compressed RTP/UDP/IP header size (in bytes)				POSITIVE INTEGER		mandatory
packet agent header size (in bytes)					POSITIVE INTEGER		mandatory

nominal radio bearer bitrate (in kbit/s)				POSITIVE INTEGER		mandatory
frame size (in bits)							POSITIVE INTEGER		mandatory
radio link agent header size (in bytes)					POSITIVE INTEGER		mandatory
acknowledged mode (y/n)							y (for yes) or n (for no)	mandatory
fully-persistent ARQ (y/n)						y (for yes) or n (for no)	only needed, if above is y
maximum number of retransmissions per frame in acknowledged mode	POSITIVE INTEGER		only needed, if above is n


It should be mentioned that all of the above STRING values are limited to a total of 99 characters!

The actual entries depend on the desired simulation scenario. Suitable combinations can be either found in the document 
VCEG-M77 or have to be decided by the user.



3.2 Running the simulation

After having created a valid <parameter-file>, a single simulation is started by simply entering

> mobile_ip <parameter-file>

where the last expression has to be replaced by the actual name of your personal parameter file!

The tool will then start to read the RTP source stream from the specified input file,
perform a system level simulation according to the requirements in document VCEG-M77 and the actual
parameter settings, and write the resulting RTP stream to the specified output file.



3.3 Simulation results

The results of the simulation run can be found in three different files, whose location and names are
all set in the <parameter-file>:
- The file for the output RTP stream, which conforms to the I/O-format specified in document VCEG-M77.

- The simulation log file, which contains a trace of all major events that have occurred during
  the run. This file is mainly for the purpose of verification of parameter settings and debugging
  in case of strange simulation results.

- The statistics file, which contains all relevant statistics about packet counters, error rates, loss
  rates, bitrates, ...
  In the following, we will give a brief definition of what the different elements mean and how they
  are computed:


STATISTIC							DEFINTION
----------------------------------------------------------------------------
LOSS-GENERATOR: Statistics:
 
total number of bits transmitted:				total number of bits that would have been sent in a real system

number of erroneous bits:					total number of erroneous bits among them (according to the specified bit error pattern)

bit error rate:							resulting bit error rate at radio frame level
 
total number of frames transmitted:				total number of radio link frames that would have been sent in a real system

number of erroneous frames:					total number of corrupt ones among them (i.e. that contain at least one erroneous bit)

frame error rate:						resulting frame error rate at radio link level

----------------------------------------------------------------------------
PACKET-DATA-AGENT: Statistics:
 
total number of RTP packets transmitted:			total number of RTP packets that would have been sent in a real system (equals
                                                                the number of RTP packets in the input stream)

number of lost RTP packets:					total number of dropped RTP packets (i.e. that contain at least one corrupt segment)

packet loss rate:						resulting packet loss rate at RTP level
 
effective bitrate (including RTP header):			effective bitrate at the RTP receiver (including the RTP header overhead), i.e.

						                    sum ( all bits in the correctly received RTP packets, including RTP header ) 
                                                                  ---------------------------------------------------------------------------------
												simulation-time

effective net bitrate (excluding RTP header):			effective bitrate at the RTP receiver (excluding the RTP header overhead), i.e.

						                  (# all bits in the correctly received RTP packets, excluding RTP header )
                                                                  ---------------------------------------------------------------------------------
												simulation-time

							        -> can be used to determine an appropriate payload bitrate of the video application

----------------------------------------------------------------------------
RADIO-LINK-AGENT: Statistics:
 
total number of radio link frames transmitted:			total number of radio link frames that would have been snet in a real system

number of retransmissions:					total number of retransmitted frames among them in case of acknowledged mode operation

number of dummy frames:						total number of dummy frames among them; dummy frames are generated each time the
								source stream bitrate falls below the effective radio bearer bitrate. This corresponds to
                                                                a lack of application data at the transmitter in a real system, where the synchronization
                                                                of the radio bearer has to be kept up by simply sending idle frames. 
                                                                A similar procedure is performed whenever a radio frame can only be partly filled with
                                                                application data (so-called stuffing), but this type of frames is not considered
                                                                in the dummy statistic since it partly contains application data. However, the log file
								contains an entry each time one of these two stuffing procedures has occurred.
								NOTE: This feature is mainly aimed at providing the user with an indication of
                                                                      how good the generated RTP source stream matches the available radio bearer bitrate.
                                                                      If the payload bitrate is chosen correctly, in practice dummy frames should not
								      appear at all!
 
total bitrate (including radio link header):			total bitrate (including radio link overhead) at radio link level, which should
                                                                equal the nominal radio bearer bitrate setting in the <parameter-file>, and is computed as
								
								# transmitted bits at radio link level
								--------------------------------------
									simulation-time

total net bitrate (excluding radio link header):		total bitrate (excluding radio link overhead) at radio link level, which is computed as

								sum ( all payload bits in the transmitted radio link frames )
								-------------------------------------------------------------
										simulation-time
										
								-> represents the maximum possible payload bitrate in the error-free case at radio
                                                                   link level, and should be used in unacknowledged mode to determine a suitable
                                                                   RTP-level framing and application bitrate


effective net bitrate (including only correct data bits):	effective bitrate at radio link level, which is computed as

								sum ( all payload bits in the correclty received radio link frames )
								--------------------------------------------------------------------
										simulation-time

								-> represents the maximum possible payload bitrate in the erroneous case at radio
                                                                   link level (possible retransmissions included in acknowledged mode), and should
								   be used in acknowledged mode to determine a suitable RTP-level frameing and 
								   application bitrate

----------------------------------------------------------------------------

total simulation time:						total simulation time in ms, which is always a multiple of the transmission time
								interval (time between two successive equally sized frames) with respect
								to the selected radio bearer
----------------------------------------------------------------------------            


NOTE: All of the above bitrates are usually in units of kbit/s, where the conversion factor between kbit and bit is assumed to be 1000!


4. Questions, comment, suggestions, bug reports

... should be adressed directly to the authors of this software tool as mentioned in the header of this file.