VSR1 burst hardware injections
        
        
          
            by
          Marie Anne Bizouard
          
          —
        
        
        last modified
        
        2007-08-14 13:39
        
        
        
        
       
    
First checks of the burst hardware injections using EGC pipeline
            EGC is tuned for [150 Hz, 2 kHz] runs on the 2nd, 3rd, 4th and 5th (loud injections) periods of burst hardware injections
Sine Gaussian with frequency 1304, 914, 850, 554, 393, 235, 153, 105, 70 (Q=9) are injected
Nominal SNR: 7.5 and 15 (75 and 150 for loud injections)
During HI 2, there are still some high frequency SG injected (3068 and 2000 Hz).
As expected, SG with frequency out of the EGC frequency band are not well reconstructed or just missed. That concerns high frequency HI 2 SG and 70 and 105 Hz SG injections for all the other injections.
Concerning the SNR recovery, for standard HI injections the spread of the ratio reconstructed SNR / injected SNR is quite large (XX +/- .25) where the average XX seems to be around .9
For the loud HI injections, the average depending on the SG frequency goes from .8 up to .9 with a rather small deviation 0.02.
We think that this can be explained by some difference in the sensitivity curve used to calibrate the HI and the one at the instant of the injections.
 
        
    
    
                    
                      Sine Gaussian with frequency 1304, 914, 850, 554, 393, 235, 153, 105, 70 (Q=9) are injected
Nominal SNR: 7.5 and 15 (75 and 150 for loud injections)
During HI 2, there are still some high frequency SG injected (3068 and 2000 Hz).
As expected, SG with frequency out of the EGC frequency band are not well reconstructed or just missed. That concerns high frequency HI 2 SG and 70 and 105 Hz SG injections for all the other injections.
Concerning the SNR recovery, for standard HI injections the spread of the ratio reconstructed SNR / injected SNR is quite large (XX +/- .25) where the average XX seems to be around .9
For the loud HI injections, the average depending on the SG frequency goes from .8 up to .9 with a rather small deviation 0.02.
We think that this can be explained by some difference in the sensitivity curve used to calibrate the HI and the one at the instant of the injections.
|  |  |  | 
 Forgot your username or password?
           Forgot your username or password?
        