diff options
Diffstat (limited to 'pjmedia/build/Jbtest.dat')
| -rw-r--r-- | pjmedia/build/Jbtest.dat | 39 |
1 files changed, 28 insertions, 11 deletions
diff --git a/pjmedia/build/Jbtest.dat b/pjmedia/build/Jbtest.dat index b2f4999d..4f5a67fd 100644 --- a/pjmedia/build/Jbtest.dat +++ b/pjmedia/build/Jbtest.dat @@ -19,11 +19,11 @@ # # 3. Success conditions, started with '!', followed by condition name # and its maximum tolerable value, in frames unit. Recognized condition -# names are: burst, discard, lost, empty, delay. These conditions will -# be verified with jitter buffer statistics after all session test data -# are executed. +# names are: burst, discard, lost, empty, delay, delay_min. These +# conditions will be verified with jitter buffer statistics after all +# session test data are executed. # Example: -# !delay 10 <- maximum average delay of jbuf is 10 frames +# !delay 10 <- average delay of jbuf is 10 frames # # 4. Session test data, containing sequence of jitter buffer events, # an event is represented by a character as follow: @@ -295,7 +295,7 @@ PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG !discard 50 <- frames discarded for delay adaptation !lost 50 <- ticket #1188, normal frame after discarded frame is flagged 'lost' to align signal !empty 0 -!delay 25 <- average delay, JB is able to adapt the delay +!delay_min 2 <- minimum delay, JB is able to adapt the delay PPPPPPPPPPPPPPPPPPPP PPPPPPPPPPPPPPPPPPPP PPPPPPPPPP PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG @@ -311,13 +311,30 @@ PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG +PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG . = Fixed mode prefetch 5, with two empty events %fixed 5 !burst 1 -!discard 0 -!lost 0 +!discard 4 <- the burst level is about 1, but prefetching will cause delay by 5 frames prefetching, delay adjustment may take place later on +!lost 4 <- progressive discard drops frames as if they were lost !empty 10 !delay 5 G @@ -329,8 +346,8 @@ PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG = Fixed mode prefetch 5, with random burst %fixed 5 !burst 3 -!discard 0 -!lost 0 +!discard 4 <- (see above test scenario) +!lost 4 <- (see above test scenario) !empty 5 !delay 5 PGPGPPGGPPPPGGPGGGPG PGGGGPPPGPPGPPPGGPGG PGPGPPGGPPPPGGPGGGPG @@ -349,7 +366,7 @@ PGGGGPPPGPPGPPPGGPGG PGPGPPGGPPGGPPPGGGPG PGGGGPPPGPPGPPPGGPGG !discard 50 <- frames discarded for delay adaptation !lost 50 <- ticket #1188, normal frame after discarded frame is flagged 'lost' to align signal !empty 0 -!delay 20 <- average delay, twice of minimal prefetch +!delay_min 20 <- minimum delay, twice of minimal prefetch PPPPPPPPPPPPPPPPPPPP PPPPPPPPPPPPPPPPPPPP PPPPPPPPPP PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG @@ -369,7 +386,7 @@ PGPGPGPGPGPGPGPGPGPG PGPGPGPGPGPGPGPGPGPG PGPGPGPGPG = Large PUT burst at beginning, then normal with burst level 10 and periodic burst spikes %adaptive 0 0 40 -!burst 10 +!burst 12 !discard 300 <- not so relevant for long period session with many delay adjustments needed (i.e: for first burst and periodic spikes) !lost 300 <- ticket #1188, normal frame after discarded frame is flagged 'lost' to align signal !empty 60 <- delay adjustment effect, as there is actually no drift |