The Audine camera was photometrically calibrated by observing numerous stars in various regions of the sky, referred to as Selected Areas. The magnitude of the selected stars in the RGBI system is readily available from star catalogues. By observing these stars with a telescope fitted with the Audine camera and the appropriate RGBI filters, it is possible to produce the complete spectral characteristics of the camera in a standard photometric reference system. It is important to note that the calibration coefficients described below are only completely valid for the specific observation site used, the elevation of the objects observed relative to the local horizon and to some degree the atmospheric conditions at the time of observation. There is also the potential for significant differences in spectral sensitivity between different KAF-0400 CCD arrays. The coefficients provided should therefore be taken as an approximate indication of the relative performance across the spectrum only.

The Schott filters used for producing the RGBI pass bands were as follows:

1 mm BG12 + 2 mm GG385 + 1 mm BG18
2 mm GG495 + 1 mm BG18 + 1 mm WG305
2 mm OG570 + 2 mm KG3
3 mm RG9 + 1 mm WG305
The filters used, mounted in an AUDE group filter holder.

The observation site used was in the Toulouse suburb of Ramonville Saint-Agne (code UAI 959). The stars were observed at an average elevation of 50 degrees relative to the horizon.

The telescope used was a flat field camera of 190mm, F/D of 4 and a central obstruction of 0.3. The resulting images were analysed using the QMiPS32 software. The Audine camera used was a prototype whose characteristics are very similar to the production model.

The Audine camera mounted on the "flat-field camera". The colour filters are placed between the camera and the telescope in the AUDE filter holder.

The following table gives the signal strength observed as a function of the filter used, for a G0-spectral-type star (B-G = 0.58, G-R = 0.52, G-I = 0.93) :

Relative Signal Strength
No filter
This table shows, for example, that to obtain the same signal strength with a B filter as with a KG3 filter, an image collection cycle of 3.25/0.25 = 12.5 times as long is required, for a G0-spectrum star.

The data below define the equations linking the system of magnitudes described above for the equpment, and the magnitudes given in the star catalogues. The instrument magnitudes are based on an image collection cycle of 60s.

b, g, r, i, M are the magnitudes recorded by the instrument
R, G, B, I are the catalogued magnitudes

If we therefore observe a star with a colour index G-R of 0.4 using an R filter, over a collection cycle of 120 seconds, this star would give a signal of 500 counts (this signal is measured by summing the photometry values of the pixels associated with the star, taking care to eliminate the part of the signal emanating from the deep sky). So how is the magnitude of the star calculated?

In 60s, the signal would be 500/2=250 counts. We can now transform this signal into the instrument's magnitude value using Pogson's formula which defines the system of magnitudes:

m = -2.5 log (signal count)

or :

m = -2.5 log (250) = -5.99

We are looking for a magnitude R, and we have m = r = -5.99

From the third equation above, we can write:

R = r + 21.616 + 0.043 (G-R)

or :

R = -5.99 + 21.616 + 0.043 . 0.4 = 15.64

The magnitude R sought is therefore 15.64.

The same star observed in the absence of any filter would give an instrument magnitude as follows for a 60s collection cycle:

M(no filter) = R - 23.197 - 0.033 (G-R) = 15.64 - 23.197 - 0.033 . 0.4 = -7.57

or in terms of signal counts over a 60s collection cycle:

Signal counts = 2.512-M =  2.5127.57 = 1067

The ratio of signal strength without filtering to strength with the R filter is therefore 1067/250=4.27, i.e. the value indicated in the table of relative signal strengths above.

The new Kodak KAF-0401E CCD array shows different characteristics to those above due to a general improvement in the CCD's sensitivity, particularly noticeable in the blue region of the spectrum. This CCD array is compatible with the Audine camera, at the cost of certain minor modifications.

A photometric calibration was performed with an Audine camera equipped with a KAF-0401E, and the results for the R, G, B, I bands for a viewframe of SA101 are now presented:

b, g, r, i, M are the instrument magnitudes
B, G, R, I are the catalogued magitudes

The relative sensitivity for the KAF-0401E between the B, G, R and I bands is as follows (again for a star of spectral type G0) :
Relative signal strength
Therefore, for a colour exposure, if it is necessary to have an image collection cycle of 15 minutes for a good signal in the red part of the spectrum, it is necessary to have a collection cycle of 15/0.4=37.5 minutes in the blue, to ensure a good colour balance (with a KAF-0400, it would take 58 minutes).

The following table shows the increase in sensitivity of a KAF-0401E over a KAF-0400 for a star of spectral type G0:

Ratio of sensitivity
On average, therefore, the KAF-0401E can be considered as 1.6 times more sensitive than the KAF-0400.