According to the HD standards, signals with image frequencies between 23.976 and 60 fps may be used. HD cameras that capture higher speeds therefore use a buffer, which is then playing out at a standard image frequency.

Variable speeds (ramps) are realized using similar techniques. A camera will deliver a continuous signal at e.g. 50 fps, while the camera head is running a ramp from 20 to 48 fps. As at the beginning of the ramp, more than twice as many frames are required to fill the 50 fps signal as are actually captured by the camera head, the 20 fps are simply duplicated, as required. To get rid of duplicate frames, only the exposed frames marked with a so-called ‘flag’. Some recording systems have a buffer of their own and discard duplicate frames directly during recording. Other systems skip duplicates during playback, or leave handling the duplicates for the following steps in postproduction.

In the meantime, some cameras also work with data based outputs. When a recorder is connected via a data based interface (also called IT-interface), the images are simply stored at whatever frequency the camera can deliver them. The limiting factor is the maximum data rate, which can be achieved by the interface or inside the camera. Some manufacturers use a trick to inflate their camera’s image frequency, as explained under Windowing. The RED ONE, for instance, can output the full sensor aperture up to 30 fps. By using a smaller aperture, the data volume per frame is reduced, which in turn, allows the RED to double or triple the 30 fps.

The D-21 also offers data output, but through a video-interface instead of an IT-interface. For transmission, the camera has to pack the date into a signal resembling a video stream, which then needs to be unpacked on the recording side. Due to the underlying HD standards, the frame rate cannot be chosen freely as with pure data interfaces. The D-21 thus only offers a hybrid video/data transfer solution.