recorded the images on so-called USB keys (which, moreover, they purchased with their own money). This practice admittedly demonstrates the importance placed by officers on the storage of images: such images making it possible, if necessary, to give proof of blameless conduct. However, it is unacceptable in its very principle. Images of persons deprived of liberty cannot be placed in media which are easily copied and accessible to third parties without difficulty. Practices of this kind should be totally prohibited. The manner in which the necessary equipment is installed often shows a lack of discernment87. In many cases, rather than the being directly installed in the cells, cameras are installed in corridors and directed in such a manner as to make the interior of the cells visible. In general, in recently-built police stations, the cells are equipped with toilets, screened-off by a low wall. However, in some cases persons in police custody are shown using the toilets in the images. Conversely, camera angles sometimes leave blind spots, rendering the effectiveness of the apparatus doubtful. Cameras are sometimes focused on places which should be protected from external view: such as medical examination rooms, where present, and rooms provided for consultation with lawyers: doctors and legal advisers should be aware of the need to respect confidentiality of treatment and professional secrecy. In cases where they may feel a certain apprehension, such concerns should be resolved by a manual alarm system, rather than by the permanent presence of the “eye” of camera surveillance. And finally what can be said of the cameras that are installed in the rooms (or rather cubby-holes) reserved for body searches? They should be rapidly removed. The case of rooms for sobering up, in which persons in a public and manifest state of drunkenness are placed, poses a special problem. The Contrôleur général has sometimes hoped that these cameras might make it possible to check on what is happening in these rooms, in view of the health risks affecting these persons. Indeed, provision is made for patrols (in principle every quarter of an hour); but it is difficult to establish the reality thereof, especially in the absence of any document attesting to them. Video surveillance would therefore make it possible to attract immediate attention in case of danger. Conversely, should the collection of images of persons in a (momentary) state of weakness be permitted? The latter, unlike other persons locked up at the police station, are not actually in police custody [garde “à vue”, i.e. literally, held and “watched over”] and are not therefore meant to remain within constant view of those responsible for their custody, without overstepping the nature of the measure applied to them. In view of the difficulty of finding the necessary balance between risk and intrusion into the life of the person concerned, it would be preferable to seek to provide proof that the patrols are conducted, without installing cameras. On this point, the Contrôleur général has changed its position. The report for 2009 had also mentioned the mediocre quality of the images in many of the situations inspected and problems of maintenance of the equipment: in many premises inspected, the equipment is out of order and cannot therefore fulfil its expected role. Finally, problems sometimes also arise with regard to the viewing of images on the monitors. In the first place, the person responsible for viewing them may have many other tasks, thereby lengthening the response time should any problems arise. Moreover, the monitors may be poorly positioned: in one police station inspected, the monitors were placed in such a way that they were behind the officer’s back, in the latter’s usual position, while, on the other hand, they directly faced members of the public, who were thus able to view their content. 2.3.2.2 Sound Sound equipment is sometimes installed in a makeshift manner: in one police station inspected, a radio had been placed in the corridor playing music at maximum volume, in order to avoid the 87 Here we are merely returning to the observations made in the annual report for 2009: op. cit. p.99. 43