Filter design ============= This document explains guidelines that should be observed (or ignored with good reason) when writing filters for libavfilter. In this document, the word “frame” indicates either a video frame or a group of audio samples, as stored in an AVFrame structure. Format negotiation ================== The query_formats method should set, for each input and each output links, the list of supported formats. For video links, that means pixel format. For audio links, that means channel layout, sample format (the sample packing is implied by the sample format) and sample rate. The lists are not just lists, they are references to shared objects. When the negotiation mechanism computes the intersection of the formats supported at each end of a link, all references to both lists are replaced with a reference to the intersection. And when a single format is eventually chosen for a link amongst the remaining list, again, all references to the list are updated. That means that if a filter requires that its input and output have the same format amongst a supported list, all it has to do is use a reference to the same list of formats. query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism to try again later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions ========================================== Principle --------- Audio and video data are voluminous; the frame and frame reference mechanism is intended to avoid, as much as possible, expensive copies of that data while still allowing the filters to produce correct results. The data is stored in buffers represented by AVFrame structures. Several references can point to the same frame buffer; the buffer is automatically deallocated once all corresponding references have been destroyed. The characteristics of the data (resolution, sample rate, etc.) are stored in the reference; different references for the same buffer can show different characteristics. In particular, a video reference can point to only a part of a video buffer. A reference is usually obtained as input to the filter_frame method or requested using the ff_get_video_buffer or ff_get_audio_buffer functions. A new reference on an existing buffer can be created with av_frame_ref(). A reference is destroyed using the av_frame_free() function. Reference ownership ------------------- At any time, a reference “belongs” to a particular piece of code, usually a filter. With a few caveats that will be explained below, only that piece of code is allowed to access it. It is also responsible for destroying it, although this is sometimes done automatically (see the section on link reference fields). Here are the (fairly obvious) rules for reference ownership: * A reference received by the filter_frame method belongs to the corresponding filter. * A reference passed to ff_filter_frame is given away and must no longer be used. * A reference created with av_frame_ref() belongs to the code that created it. * A reference obtained with ff_get_video_buffer or ff_get_audio_buffer belongs to the code that requested it. * A reference given as return value by the get_video_buffer or get_audio_buffer method is given away and must no longer be used. Link reference fields --------------------- The AVFilterLink structure has a few AVFrame fields. partial_buf is used by libavfilter internally and must not be accessed by filters. fifo contains frames queued in the filter's input. They belong to the framework until they are taken by the filter. Reference permissions --------------------- Since the same frame data can be shared by several frames, modifying may have unintended consequences. A frame is considered writable if only one reference to it exists. The code owning that reference it then allowed to modify the data. A filter can check if a frame is writable by using the av_frame_is_writable() function. A filter can ensure that a frame is writable at some point of the code by using the ff_inlink_make_frame_writable() function. It will duplicate the frame if needed. A filter can ensure that the frame passed to the filter_frame() callback is writable by setting the needs_writable flag on the corresponding input pad. It does not apply to the activate() callback. Frame scheduling ================ The purpose of these rules is to ensure that frames flow in the filter graph without getting stuck and accumulating somewhere. Simple filters that output one frame for each input frame should not have to worry about it. There are two design for filters: one using the filter_frame() and request_frame() callbacks and the other using the activate() callback. The design using filter_frame() and request_frame() is legacy, but it is suitable for filters that have a single input and process one frame at a time. New filters with several inputs, that treat several frames at a time or that require a special treatment at EOF should probably use the design using activate(). activate -------- This method is called when something must be done in a filter; the definition of that "something" depends on the semantic of the filter. The callback must examine the status of the filter's links and proceed accordingly. The status of output links is stored in the frame_wanted_out, status_in and status_out fields and tested by the ff_outlink_frame_wanted() function. If this function returns true, then the processing requires a frame on this link and the filter is expected to make efforts in that direction. The status of input links is stored by the status_in, fifo and status_out fields; they must not be accessed directly. The fifo field contains the frames that are queued in the input for processing by the filter. The status_in and status_out fields contains the queued status (EOF or error) of the link; status_in is a status change that must be taken into account after all frames in fifo have been processed; status_out is the status that have been taken into account, it is final when it is not 0. The typical task of an activate callback is to first check the backward status of output links, and if relevant forward it to the corresponding input. Then, if relevant, for each input link: test the availability of frames in fifo and process them; if no frame is available, test and acknowledge a change of status using ff_inlink_acknowledge_status(); and forward the result (frame or status change) to the corresponding input. If nothing is possible, test the status of outputs and forward it to the corresponding input(s). If still not possible, return FFERROR_NOT_READY. If the filters stores internally one or a few frame for some input, it can consider them to be part of the FIFO and delay acknowledging a status change accordingly. Example code: ret = ff_outlink_get_status(outlink); if (ret) { ff_inlink_set_status(inlink, ret); return 0; } if (priv->next_frame) { /* use it */ return 0; } ret = ff_inlink_consume_frame(inlink, &frame); if (ret < 0) return ret; if (ret) { /* use it */ return 0; } ret = ff_inlink_acknowledge_status(inlink, &status, &pts); if (ret) { /* flush */ ff_outlink_set_status(outlink, status, pts); return 0; } if (ff_outlink_frame_wanted(outlink)) { ff_inlink_request_frame(inlink); return 0; } return FFERROR_NOT_READY; The exact code depends on how similar the /* use it */ blocks are and how related they are to the /* flush */ block, and needs to apply these operations to the correct inlink or outlink if there are several. Macros are available to factor that when no extra processing is needed: FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink); FF_FILTER_FORWARD_STATUS_ALL(outlink, filter); FF_FILTER_FORWARD_STATUS(inlink, outlink); FF_FILTER_FORWARD_STATUS_ALL(inlink, filter); FF_FILTER_FORWARD_WANTED(outlink, inlink); filter_frame ------------ For filters that do not use the activate() callback, this method is called when a frame is pushed to the filter's input. It can be called at any time except in a reentrant way. If the input frame is enough to produce output, then the filter should push the output frames on the output link immediately. As an exception to the previous rule, if the input frame is enough to produce several output frames, then the filter needs output only at least one per link. The additional frames can be left buffered in the filter; these buffered frames must be flushed immediately if a new input produces new output. (Example: frame rate-doubling filter: filter_frame must (1) flush the second copy of the previous frame, if it is still there, (2) push the first copy of the incoming frame, (3) keep the second copy for later.) If the input frame is not enough to produce output, the filter must not call request_frame to get more. It must just process the frame or queue it. The task of requesting more frames is left to the filter's request_frame method or the application. If a filter has several inputs, the filter must be ready for frames arriving randomly on any input. Therefore, any filter with several inputs will most likely require some kind of queuing mechanism. It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced. request_frame ------------- For filters that do not use the activate() callback, this method is called when a frame is wanted on an output. For a source, it should directly call filter_frame on the corresponding output. For a filter, if there are queued frames already ready, one of these frames should be pushed. If not, the filter should request a frame on one of its inputs, repeatedly until at least one frame has been pushed. Return values: if request_frame could produce a frame, or at least make progress towards producing a frame, it should return 0; if it could not for temporary reasons, it should return AVERROR(EAGAIN); if it could not because there are no more frames, it should return AVERROR_EOF. The typical implementation of request_frame for a filter with several inputs will look like that: if (frames_queued) { push_one_frame(); return 0; } input = input_where_a_frame_is_most_needed(); ret = ff_request_frame(input); if (ret == AVERROR_EOF) { process_eof_on_input(); } else if (ret < 0) { return ret; } return 0; Note that, except for filters that can have queued frames and sources, request_frame does not push frames: it requests them to its input, and as a reaction, the filter_frame method possibly will be called and do the work.