Changes

Page history
Update home authored by Wei Changfeng's avatar Wei Changfeng
Hide whitespace changes
Inline Side-by-side
home.md
View page @ 58a5374d
...@@ -290,12 +290,12 @@ Fig.4 & Fig.5 ...@@ -290,12 +290,12 @@ Fig.4 & Fig.5
> dec=-1.01 > dec=-1.01
Their posterior look very similar. Their posterior look very similar.
### (iv)windowing signal helps... ### (iv)windowing signal...
Windowing may so
## (14)VItamin: bilby BBH waveform test ## (14)VItamin: bilby BBH waveform test
After remaking Hunter's data, I'm going to see how well VItamin estimates bilby waveform. After remaking Hunter's data, I'm going to see how well VItamin estimates bilby waveform.
Fig.6 bilby signal Fig.6 bilby signal
![signal](uploads/2e2d1c8e16d9d6228d3cb08d7eaf977e/signal.png) ![signal](uploads/2e2d1c8e16d9d6228d3cb08d7eaf977e/signal.png)
Fig.7 bilby whitened signal Fig.7 bilby whitened signal
...@@ -311,8 +311,15 @@ Fig.9 bilby signal's posterior ...@@ -311,8 +311,15 @@ Fig.9 bilby signal's posterior
> theta_jn=2.62 > theta_jn=2.62
> ra=3.14 > ra=3.14
> dec=0 > dec=0
As we can see, mass_1, t_zero, dec are not estimated very well, whose truths are on the edge. As we can see, mass_1, t_zero, dec are not estimated very well, whose truths are on the edge.
(15)VItamin: Korean BBH waveform test ## (15)VItamin: Korean BBH waveform test
For Korean waveform, signal is generated with points less than 256. We need to pad it to 256 as VItamin requires. I pad it on the back because start time is already set when generating signal.
Next is to find t_zero's value.
In my understanding of t zero:
![微信截图_20201203011518](uploads/9eb61191e74b0e31c8298f7577418589/微信截图_20201203011518.png)
t_zero = t_peak - ref_geocent_time.(Is that right?)
Then we get t_zero's truth.
Fig.10 Korean signal Fig.10 Korean signal
![000](uploads/44643fae14b16ea57b031dc8364cda8e/000.png) ![000](uploads/44643fae14b16ea57b031dc8364cda8e/000.png)
Fig.11 Korean whitened signal Fig.11 Korean whitened signal
...@@ -320,7 +327,7 @@ Fig.11 Korean whitened signal ...@@ -320,7 +327,7 @@ Fig.11 Korean whitened signal
Fig.12 Korean whitened signal with Gaussian white noise Fig.12 Korean whitened signal with Gaussian white noise
![003](uploads/fdadcff96b9e14f23a6b57b88e680a45/003.png) ![003](uploads/fdadcff96b9e14f23a6b57b88e680a45/003.png)
Fig.13 Korean signal's posterior Fig.13 Korean signal's posterior
![vitamin_corner_timeseries-0](uploads/c24ff1d516e0fa8f87c93ab70835b23a/vitamin_corner_timeseries-0.png) ![vitamin_corner_timeseries-0](uploads/c24ff1d516e0fa8f87c93ab70835b23a/vitamin_corner_timeseries-0.png)
> mass_1=57.5 > mass_1=57.5
> mass_2=57.5 > mass_2=57.5
> luminosity_distance=2236 > luminosity_distance=2236
...@@ -328,7 +335,10 @@ Fig.13 Korean signal's posterior ...@@ -328,7 +335,10 @@ Fig.13 Korean signal's posterior
> theta_jn=6.08 > theta_jn=6.08
> ra=5.61 > ra=5.61
> dec=-0.07 > dec=-0.07
(16)Comparison between Korean waveform
I generate 4 samples for this test. Plots above are of the first one. As we can see, theta_jn and ra are totally wrong. Other parameters are good. Furthermore, theta_jn and ra also get a bad estimation in other 3 samples.
## (16)Comparison between Bilby waveform and Korean waveform
# 2.Data # 2.Data
## (1)Korean merger wavform ## (1)Korean merger wavform
... ...
......