gstreamer/docs/design/part-negotiation.txt
Stefan Kost 2e8d4a491d docs/design/: Fix two typos.
Original commit message from CVS:
* docs/design/draft-framestep.txt:
* docs/design/part-negotiation.txt:
Fix two typos.
2008-06-29 15:00:26 +00:00

268 lines
11 KiB
Text

Negotiation
-----------
Capabilities negotiation is the process of deciding on an adequate
format for dataflow within a GStreamer pipeline. Ideally, negotiation
(also known as "capsnego") transfers information from those parts of the
pipeline that have information to those parts of the pipeline that are
flexible, constrained by those parts of the pipeline that are not
flexible.
GStreamer's two scheduling modes, push mode and pull mode, lend
themselves to different mechanisms to achieve this goal. As it is more
common we describe push mode negotiation first.
Push-mode negotiation
---------------------
Push-mode negotiation happens when elements want to push buffers and
need to decide on the format. This is called downstream negotiation
because the upstream element decides the format for the downstream
element. This is the most common case.
Negotiation can also happen when a downstream element wants to receive
another data format from an upstream element. This is called upstream
negotiation.
The basics of negotiation are as follows:
- GstCaps (see part-caps.txt) are refcounted before they
are attached to a buffer to describe the contents of the buffer.
It is possible to add a NULL caps to a buffer, this means that the
buffer type did not change relative to the previous buffer. If no
previous buffer was received by a downstream element, it is free to
discard the buffer.
- Before receiving a buffer, an element must check if the datatype of
the buffer has changed. The element should reconfigure itself to the
new format before processing the buffer data. If the data type on
the buffer is not acceptable, the element should refuse the buffer by
returning an appropriate return value from the chain function.
The core will automatically call the set_caps function for this purpose
when it is installed on the sink or source pad.
- When requesting a buffer from a bufferpool, the prefered type should
be passed to the buffer allocation function. After receiving a buffer
from a bufferpool, the datatype should be checked again.
- A bufferpool allocation function should try to allocate a buffer of the
prefered type. If there is a good reason to choose another type, the
alloc function should see if that other type is accepted by the other
element, then allocate a buffer of that type and attach the type to the
buffer before returning it.
The general flow for a source pad starting the negotiation.
src sink
| |
| accepts? |
type A |---------------->|
| yes |
|<----------------|
| |
get buffer | alloc_buf |
from pool |---------------->|
with type A | | Create buffer of type A.
| |
check type |<----------------|
and use A | |
| push |
push buffer |---------------->| Receive type A, reconfigure to
with new type| | process type A.
| |
One possible implementation in pseudo code:
[element wants to create a buffer]
if not format
# see what the peer can do
peercaps = gst_pad_peer_get_caps (srcpad)
# see what we can do
ourcaps = gst_pad_get_caps (srcpad)
# get common formats
candidates = gst_caps_intersect (peercaps, ourcaps)
foreach candidate in candidates
# make sure the caps is fixed
fixedcaps = gst_pad_fixate_caps (srcpad, candidate)
# see if the peer accepts it
if gst_pad_peer_accept_caps (srcpad, fixedcaps)
# store the caps as the negotiated caps, this will
# call the setcaps function on the pad
gst_pad_set_caps (srcpad, fixedcaps)
break
endif
done
endif
# if the type is different, the buffer will have different caps from
# the src pad -- setcaps will get called on the pad_push
buffer = gst_pad_alloc_buffer (srcpad, 0, size, GST_PAD_CAPS (fixedcaps));
if buffer
[fill buffer and push]
elseif
[no buffer, either no peer or no acceptable format found]
endif
The general flow for a sink pad starting a renegotiation.
src sink
| |
| accepts? |
|<----------------| type B
| yes |
|---------------->|
| |
get buffer | alloc_buf |
from pool |---------------->|
with type A | | Create buffer of new type B.
| |
check type |<----------------|
and | |
reconfigure | |
| push |
push buffer |---------------->| Receive type B, reconfigure to
with new type| | process type B.
| |
Use case:
videotestsrc ! xvimagesink
1) Who decides what format to use?
- src pad always decides, by convention. sinkpad can suggest a format
by putting it high in the getcaps function GstCaps.
- since the src decides, it can always choose something that it can do,
so this step can only fail if the sinkpad stated it could accept
something while later on it couldn't.
2) When does negotiation happen?
- before srcpad does a push, it figures out a type as stated in 1), then
it calls the pad alloc function with the type. The sinkpad has to
create a buffer of that type, src fills the buffer and sends it to sink.
- since the sink stated in 1) it could accept the type, it will be able to
create a buffer of the type and handle it.
- sink checks media type of buffer and configures itself for this type.
3) How can sink request another format?
- sink asks if new format is possible for the source.
- sink returns buffer with new type in allocfunction.
- src receives buffer with new type, reconfigures and pushes.
- sink can always select something it can create and handle since it takes
the initiative. src should be able to handle the new type since it said
it could accept it.
videotestsrc ! queue ! xvimagesink
- queue implements an allocfunction, proxying all calls to its srcpad peer.
- queue proxies all accept and getcaps to the other peer pad.
- queue contains buffers with different types.
Pull-mode negotiation
---------------------
Rationale
.........
A pipeline in pull mode has different negotiation needs than one
activated in push mode. Push mode is optimized for two use cases:
* Playback of media files, in which the demuxers and the decoders are
the points from which format information should disseminate to the
rest of the pipeline; and
* Recording from live sources, in which users are accustomed to putting
a capsfilter directly after the source element; thus the caps
information flow proceeds from the user, through the potential caps
of the source, to the sinks of the pipeline.
In contrast, pull mode has other typical use cases:
* Playback from a lossy source, such as RTP, in which more knowledge
about the latency of the pipeline can increase quality; or
* Audio synthesis, in which audio APIs are tuned to producing only the
necessary number of samples, typically driven by a hardware interrupt
to fill a DMA buffer or a Jack[0] port buffer.
* Low-latency effects processing, whereby filters should be applied as
data is transferred from a ring buffer to a sink instead of
beforehand. For example, instead of using the internal alsasink
ringbuffer thread in push-mode wavsrc ! volume ! alsasink, placing
the volume inside the sound card writer thread via wavsrc !
audioringbuffer ! volume ! alsasink.
[0] http://jackit.sf.net
The problem with pull mode is that the sink has to know the format in
order to know how many bytes to pull via gst_pad_pull_range(). This
means that before pulling, the sink must initiate negotation to decide
on a format.
Recalling the principles of capsnego, whereby information must flow from
those that have it to those that do not, we see that the two named use
cases have different negotiation requirements:
* RTP and low-latency playback are both like the normal playback case,
in which information flows downstream.
* In audio synthesis, the part of the pipeline that has the most
information is the sink, constrained by the capabilities of the graph
that feeds it. However the caps are not completely specified; at some
point the user has to intervene to choose the sample rate, at least.
This can be done externally to gstreamer, as in the jack elements, or
internally via a capsfilter, as is customary with live sources.
Given that sinks potentially need the input of sources, as in the RTP
case and at least as a filter in the synthesis case, there must be a
negotiation phase before the pull thread is activated. Also, given the
low latency offered by pull mode, we want to avoid capsnego from within
the pulling thread, in case it causes us to miss our scheduling
deadlines.
The time to do capsnego, then, is after activate_pull() has succeeded,
but before the sink has spawned the pulling thread. Because of the
latency concerns just mentioned, capsnego does not occur in the pulling
thread.
Mechanism
.........
The sink initiates the negotiation process by intersecting the results
of gst_pad_get_caps() on its sink pad and its peer src pad. This is the
operation performed by gst_pad_get_allowed_caps(). In the simple
passthrough case, the peer pad's getcaps() function should return the
intersection of calling get_allowed_caps() on all of its sink pads. In
this way the sink element knows the capabilities of the entire pipeline.
The sink element then fixates the resulting caps, if necessary,
resulting in the flow caps. It notifies the pipeline of the caps by
calling gst_pad_set_caps() on its sink pad. Sink pads should proxy the
setcaps() to their peer src pads. In the simple passthrough case, src
pads' setcaps() functions proxy the setcaps() to all of their sink pads,
which then set_caps() on their peers, and so the entire pipeline becomes
configured before dataflow has started. All pads have fixed caps.
If the sink element could not set caps on its sink pad, it should post
an error message on the bus indicating that negotiation was not
possible.
In this way all pads are negotiated before data starts flowing, so all
getrange() requests have a defined meaning for the number of bytes being
pulled.
During dataflow, gst_pad_pull_range() checks the caps on the pulled
buffer. If they are different from the sink pad's caps, it will return
GST_FLOW_NOT_NEGOTIATED. Because of the low-latency requirements,
changing caps in an activate pull-mode pipeline is not supported, as it
might require e.g. the sound card to reconfigure its hardware buffers,
and start capsnego again.