... output4.1
The input pupil is the plane where the electric field coming from the object enters the interferometer. The output pupil is the plane where the electric field exiting from each optical train before being combined on the detector. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... theorem)6.1
Note that the Fourier transform of a time series is not defined. However, in this context, we only work with finite sections of a time series. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... filters)6.2
Because $ \tau$ is restricted to a maximum time lag, this instrumental gain factor does not describe long-term variations. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... correlator6.3
For the sake of completeness, it should be mentioned that this is a special case of the so-called Hilbert transform, which property is to change signal phases by $ \pi/2$, but to leave amplitudes unchanged. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... band6.4
Note that at Bure the total IF bandwidth available is limited by the receivers to the range $ [100,650]$ MHz 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... distribution17.1
In the following, we will assume an uniform noise rms, i.e. we do not take into account variation of the noise introduced by the imaging process (see Chapter 15 by S. Guilloteau). 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... mapped17.2
We have considered observations in different directions, performed with the same $ uv$-coverage. The analysis presented here shows that such an experiment is somehow equivalent to a single observation of the whole source, but with a denser $ uv$-coverage. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... mosaic17.3
More precisely, this file contains the non normalized mosaic $ \Sigma B_i^t \times F_i$. The proper normalization (see Eq. 17.12) is further done by the deconvolution procedures. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.