Audio Recording from Terminal Server

This is a small client-server program used to record audio from a windows remote desktop session using the microphone at the remote workstation. Recording takes place entirely at the remote workstation (no audio stream redirection), a file is produced and transferred to the terminal server session. Program should be written in c# (2.0 and above). Client is a dll residing at the terminal server, exposing its methods also via COM (must be instantiated and invoked by Lotus Notes) Server is a windows service installed at the client station, performing the actual recording and file format conversion to GSM codec. **A full working project exists (needs to be extended)** ## Deliverables **We have already created a mock-up project that does the recording part. Both the client, the server and the installer exist. Project is in VS 2005. The mock-up needs to be finalized into a reliable tool and provided with user interface/monitoring.** Client Requirements (Terminal Server Part) A)? DLL format with COM interfaces, instantiatable from external programs via createObject (Lotus notes), already present in mock-up B) Methods invokable from external programs via COM (Lotus Notes), already present in mock-up C) Configuration from file (already present in mock up) D) Logging of activities to file (already present in mock up) E)? Non blocking startrecording method which sends call to server via remoting (already present in mock up) F) Non blocking stoprecording method wich sends call to server via remoting receives the recored file bytes, saves them to a temp file and returns the caller with temp file name? (already present in mock up) G)? Small form (always in foreground on top of all other opened programs) with effective real time display of the recorded audio volume (similar to WIndows audio recorder Fourier transform),? green/red semaphore in case of right/low audio input? with possibility to change the audio device and modify audio recording volume? (non present in mock up). H)? Small form indicating the progress bar in the file transfer status (during stop recording method) With reference to points G and H (the parts not present in mock up) some further explanation is due. The actual limitation of the mock-up is due to the fact that no visual feedback is given to the user about the actual stuff being recorded: if microphone is muted or recording volume is too low, there is no way to know if recording is taking place? correctly. When recording is completed the file gets transferred from the server (user PC) to the client (Terminal server) and we need to provide a feedback about the transfer activitiy with an estimation of the time left. The two forms implementation can be done through a taskbar service which pulls out the foreground forms (eventually the whole client part can be implemented in this way). Since at present ? the call from the external program is startrecording and then the dll gets unloaded, in the new version? it is necessary to eventually move to a service in order to have the program still running after the invocation. It needs always to be kept in mind that we are working in a Terminal Server session (multiple users on same machine) so we need to take care of the session to know which of the client we need to connect to (already done in mock-up) Server Requirements (User's PC Part) A) Windows Service listening through remoting on configurable port (already present in mock up) B) Possibility to establish a user/password? (already present in mock up) C) Configuration from file (already present in mock up) D) Logging of activities to file (already present in mock up) E) Implementation of the remoted recording activitiy? (already present in mock up, needs to be modified/extended) F) At the end of recording, conversion of recorded audio to GSM mono format? (already present in mock up, needs to be modified/extended) G) Transfer of recorded data to client? (already present in mock up, needs to be modified/extended) With reference to points E,F,G the actual recording uses a native win32 wrapper to invoke kernel recording methods. This approach need to be "changed" since we need to provide the client with the actual recording audio level (in real time) via a Fourier transform or similar: this needs to be exposed a a remoting method/event in order to display the level in the client (Client part, point G). This means, eventually, moving away from the Win32 api and use Managed DirectX/DirectSound or other approach. So the specifications summary consist in: - taking the existing working mock up as a sample - write the client / server application adding the real time audio monitor functionality, device chooses, volume manager, file transfer status window. Pleas feel free to ask for clarifications.