[transmageddon] Further work on multistream handling
- From: Christian Fredrik Kalager Schaller <uraeus src gnome org>
- To: commits-list gnome org
- Cc:
- Subject: [transmageddon] Further work on multistream handling
- Date: Fri, 4 Jan 2013 14:10:47 +0000 (UTC)
commit d05e9d20a3695f1113abaa829667ea644aae40c1
Author: Christian Fredrik Kalager Schaller <uraeus linuxrisin org>
Date: Wed Dec 5 14:56:12 2012 +0100
Further work on multistream handling
src/transcoder_engine.py | 36 +++++++++---
src/transmageddon.py | 149 +++++++++++++++++++++++++---------------------
2 files changed, 108 insertions(+), 77 deletions(-)
---
diff --git a/src/transcoder_engine.py b/src/transcoder_engine.py
index aad5dcb..ee6ee13 100644
--- a/src/transcoder_engine.py
+++ b/src/transcoder_engine.py
@@ -36,14 +36,17 @@ class Transcoder(GObject.GObject):
'got-error' : (GObject.SignalFlags.RUN_LAST, None, (GObject.TYPE_PYOBJECT,))
}
- def __init__(self, FILECHOSEN, FILENAME, DESTDIR, CONTAINERCHOICE, AUDIOCODECVALUE, VIDEOCODECVALUE, PRESET,
- OHEIGHT, OWIDTH, FRATENUM, FRATEDEN, ACHANNELS, MULTIPASS, PASSCOUNTER, OUTPUTNAME,
+ def __init__(self, FILECHOSEN, FILENAME, DESTDIR, CONTAINERCHOICE, AUDIODATA, VIDEOCODECVALUE, PRESET,
+ OHEIGHT, OWIDTH, FRATENUM, FRATEDEN, MULTIPASS, PASSCOUNTER, OUTPUTNAME,
TIMESTAMP, ROTATIONVALUE, AUDIOPASSTOGGLE, VIDEOPASSTOGGLE, INTERLACED, INPUTVIDEOCAPS):
GObject.GObject.__init__(self)
# Choose plugin based on Container name
+ self.audiodata = AUDIODATA
+ print(self.audiodata)
self.container = CONTAINERCHOICE
- self.audiocaps = AUDIOCODECVALUE
+ self.audiocaps = self.audiodata[0]['audiotype']
+ print("audiocaps " +str(self.audiocaps))
if self.container != False:
self.containercaps = Gst.caps_from_string(codecfinder.containermap[CONTAINERCHOICE])
# special case mp3 which is a no-container format with a container (id3mux)
@@ -69,15 +72,18 @@ class Transcoder(GObject.GObject):
self.owidth = OWIDTH
self.fratenum = FRATENUM
self.frateden = FRATEDEN
- self.achannels = ACHANNELS
+ self.achannels = self.audiodata[0]['audiochannels']
+ self.astreamid = self.audiodata[0]['streamid']
self.blackborderflag = False
self.multipass = MULTIPASS
self.passcounter = PASSCOUNTER
self.outputfilename = OUTPUTNAME
self.timestamp = TIMESTAMP
self.rotationvalue = int(ROTATIONVALUE)
- self.vbox = {}
self.missingplugin= False
+ self.probestreamid = False
+ self.sinkpad = False
+
# switching width and height around for rotationchoices where it makes sense
@@ -327,6 +333,16 @@ class Transcoder(GObject.GObject):
bus.add_signal_watch()
bus.connect('message', self.on_message)
+ def padprobe(self, pad, probeinfo, userdata):
+ event = probeinfo.get_event()
+ eventtype=event.type
+ if eventtype==Gst.EventType.STREAM_START:
+ self.probestreamid = event.parse_stream_start()
+ if self.probestreamid==self.astreamid:
+ print("stream id matches")
+ pad.link(self.sinkpad)
+ return Gst.PadProbeReturn.OK
+
def on_message(self, bus, message):
mtype = message.type
# print(mtype)
@@ -363,7 +379,6 @@ class Transcoder(GObject.GObject):
return True
def OnDynamicPad(self, uridecodebin, src_pad):
- # print "src_pad is" +str(src_pad)
origin = src_pad.query_caps(None)
if (self.container==False):
a = origin.to_string()
@@ -390,14 +405,19 @@ class Transcoder(GObject.GObject):
sinkpad = self.encodebin.emit("request-pad", origin)
if c.startswith("audio/"):
if self.passcounter == int(0):
- src_pad.link(sinkpad)
+ self.sinkpad = self.encodebin.emit("request-pad", origin)
+ src_pad.add_probe(Gst.PadProbeType.EVENT_DOWNSTREAM, self.padprobe, None)
+ #if self.probestreamid==self.astreamid:
+ # print("got streamid from probe")
+ # src_pad.link(sinkpad)
elif ((c.startswith("video/") or c.startswith("image/")) and (self.videocaps != False)):
if self.videopasstoggle==False:
if (self.multipass != 0) and (self.passcounter != int(0)):
+
videoencoderpad = self.videoencoder.get_static_pad("sink")
src_pad.link(videoencoderpad)
else:
- # print "self.colorspaceconverter before use " + str(self.colorspaceconverter)
+ src_pad.add_probe(Gst.PadProbeType.EVENT_DOWNSTREAM, self.padprobe, None)
deinterlacerpad = self.deinterlacer.get_static_pad("sink")
src_pad.link(deinterlacerpad)
self.videoflipper.get_static_pad("src").link(sinkpad)
diff --git a/src/transmageddon.py b/src/transmageddon.py
index 4d88223..4738dbe 100644
--- a/src/transmageddon.py
+++ b/src/transmageddon.py
@@ -224,7 +224,7 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.videorows=[]
self.audiocodecs=[]
self.videocodecs=[]
- self.flag=False
+
# set flag so we remove bogus value from menu only once
self.bogus=0
@@ -275,6 +275,7 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.containerchoice.connect("changed", self.on_containerchoice_changed)
self.presetchoice.connect("changed", self.on_presetchoice_changed)
self.audiorows[0].connect("changed", self.on_audiocodec_changed)
+ self.audiorows[0].set_name("audiorow0")
self.videorows[0].connect("changed", self.on_videocodec_changed)
self.rotationchoice.connect("changed", self.on_rotationchoice_changed)
@@ -363,7 +364,7 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.videopasstoggle=False # toggle for passthrough mode chosen
self.audiopasstoggle=False
self.videopass=False # toggle for enabling adding of video passthrough on menu
- self.audiopass=False
+ self.audiopass=[False]
self.containertoggle=False # used to not check for encoders with pbutils
self.discover_done=False # lets us know that discover is finished
self.missingtoggle=False
@@ -386,7 +387,8 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.nocontainernumber = int(13) # this needs to be set to the number of the no container option in the menu (from 0)
self.p_duration = Gst.CLOCK_TIME_NONE
self.p_time = Gst.Format.TIME
- self.streamids=[] # (list of stream ids)
+ self.audiostreamids=[] # (list of stream ids)
+ self.videostreamids=[]
self.language=[]
self.audiostreamcounter=int(-1)
self.videostreamcounter=int(-1)
@@ -399,6 +401,7 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.markupaudiocodec=[]
self.markupvideoinfo=[]
self.markupvideocodec=[]
+ self.audiodata =[] # create list object to hold the rows
# Populate the Container format combobox
# print("do we try to populate container choice")
@@ -443,6 +446,15 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.waiting_for_signal="False"
+
+ # define the audiodata structure here as the canonical location. This structure should include everything
+ # needed for the audio pipelines. Use strings to describe the type of data to be stored.
+ # The two caps values should be GStreamer caps objects, not caps in string format.
+ # When a row is added, any data you are missing at that point should be replaced by 'False' values
+ def add_audiodata_row(self, audiochannels, samplerate, inputaudiocaps, outputaudiocaps, streamid, passthrough, language):
+ audiodata = {'audiochannels' : audiochannels, 'samplerate' : samplerate, 'inputaudiocaps' : inputaudiocaps, 'outputaudiocaps' : outputaudiocaps , 'streamid' : streamid, 'passthrough' : passthrough, 'language' : language }
+ return audiodata
+
# Get all preset values
def reverse_lookup(self,v):
for k in codecfinder.codecmap:
@@ -614,25 +626,24 @@ class TransmageddonUI(Gtk.ApplicationWindow):
for i in info.get_stream_list():
if isinstance(i, GstPbutils.DiscovererAudioInfo):
streamid=i.get_stream_id()
- if streamid not in self.streamids:
+ if streamid not in self.audiostreamids:
self.audiostreamcounter=self.audiostreamcounter+1
- print(self.audiostreamcounter)
- self.inputaudiocaps.append(i.get_caps())
- self.streamids.append(streamid)
- self.language.append(GstTag.tag_get_language_name(i.get_language()))
- self.audiostreams.append( \
- GstPbutils.pb_utils_get_codec_description(self.inputaudiocaps[self.audiostreamcounter]))
- self.audiochannels.append(i.get_channels())
- self.samplerate.append(i.get_sample_rate())
+ # self.inputaudiocaps.append(i.get_caps())
+ self.audiostreamids.append(streamid)
+ # self.language.append(GstTag.tag_get_language_name(i.get_language()))
+ # self.audiostreams.append( \
+ # GstPbutils.pb_utils_get_codec_description(self.inputaudiocaps[self.audiostreamcounter]))
+ # self.audiochannels.append(i.get_channels())
+ # self.samplerate.append(i.get_sample_rate())
self.haveaudio=True
- self.audiodata = { 'audiochannels' : self.audiochannels[self.audiostreamcounter], \
- 'samplerate' : self.samplerate[self.audiostreamcounter], 'audiotype' : self.inputaudiocaps[self.audiostreamcounter], \
- 'clipduration' : clipduration }
-
+ self.audiodata.append(self.add_audiodata_row(i.get_channels(), i.get_sample_rate(), i.get_caps(), False, streamid, False, i.get_language()))
+ #print(self.audiodata)
if self.audiostreamcounter > 0:
combo = Gtk.ComboBoxText.new()
self.audiorows.append(combo)
self.audiobox.add(self.audiorows[self.audiostreamcounter])
+ self.audiorows[self.audiostreamcounter].connect("changed", self.on_audiocodec_changed)
+ self.audiorows[self.audiostreamcounter].set_name("audiorow"+str(self.audiostreamcounter))
self.audiorows[self.audiostreamcounter].show()
self.containerchoice.set_active(-1) # set this here to ensure it happens even with quick audio-only
@@ -640,14 +651,14 @@ class TransmageddonUI(Gtk.ApplicationWindow):
if isinstance(i, GstPbutils.DiscovererVideoInfo):
streamid=i.get_stream_id()
- if streamid not in self.streamids:
+ if streamid not in self.videostreamids:
self.inputvideocaps=i.get_caps()
videotags=i.get_tags()
interlacedbool = i.is_interlaced()
if interlacedbool is True:
self.interlaced=True
self.havevideo=True
- self.streamids.append(streamid)
+ self.videostreamids.append(streamid)
self.populate_menu_choices() # run this to ensure video menu gets filled
videoheight=i.get_height()
videowidth=i.get_width()
@@ -670,28 +681,26 @@ class TransmageddonUI(Gtk.ApplicationWindow):
if self.waiting_for_signal == True:
if self.containertoggle == True:
if self.codeccontainer != False:
- # self.check_for_passthrough(self.codeccontainer)
- print("snoke")
+ self.check_for_passthrough(self.codeccontainer)
else:
# self.check_for_elements()
if self.missingtoggle==False:
self._start_transcoding()
if self.codeccontainer != False:
- #self.check_for_passthrough(self.codeccontainer)
- print("hei")
+ self.check_for_passthrough(self.codeccontainer)
# set UI markup, will wary in size depending on number of streams
if self.haveaudio:
self.markupaudioinfo=[]
if self.audiostreamcounter==0:
- self.markupaudioinfo.append(''.join(('<small>','Audio channels: ', str(self.audiochannels[self.audiostreamcounter]), '</small>',"\n", '<small>','Audio codec: ',str(GstPbutils.pb_utils_get_codec_description(self.inputaudiocaps[self.audiostreamcounter])), "\n", 'Language: ', self.language[self.audiostreamcounter],'</small>')))
+ self.markupaudioinfo.append(''.join(('<small>','Audio channels: ', str(self.audiodata[self.audiostreamcounter]['audiochannels']), '</small>',"\n", '<small>','Audio codec: ',str(GstPbutils.pb_utils_get_codec_description(self.audiodata[self.audiostreamcounter]['inputaudiocaps'])), "\n", 'Language: ', self.audiodata[self.audiostreamcounter]['language'],'</small>')))
self.audioinformation.set_markup("".join(self.markupaudioinfo))
else:
if self.audiostreamcounter==1:
x=0
while x <= self.audiostreamcounter:
- self.markupaudioinfo.append(''.join(('<small>','<b>','Audiostream no: ',str(x+1),'</b>','</small>',"\n",'<small>','Audio channels: ', str(self.audiochannels[x]), '</small>',"\n", '<small>','Audio codec: ',str(GstPbutils.pb_utils_get_codec_description(self.inputaudiocaps[x])), "\n", 'Language: ', self.language[x],'</small>',"\n")))
+ self.markupaudioinfo.append(''.join(('<small>','<b>','Audiostream no: ',str(x+1),'</b>','</small>',"\n",'<small>','Audio channels: ', str(self.audiodata[x]['audiochannels']), '</small>',"\n", '<small>','Audio codec: ',str(GstPbutils.pb_utils_get_codec_description(self.audiodata[x]['inputaudiocaps'])), "\n", 'Language: ', self.audiodata[x]['language'],'</small>',"\n")))
self.audioinformation.set_markup("".join(self.markupaudioinfo))
x=x+1
@@ -718,9 +727,9 @@ class TransmageddonUI(Gtk.ApplicationWindow):
def check_for_passthrough(self, containerchoice):
videointersect = Gst.Caps.new_empty()
audiointersect = []
- for x in self.streamids:
+ for x in self.audiostreamids:
audiointersect.append(Gst.Caps.new_empty())
- if containerchoice != False: # or self.usingpreset==False): <- Need to figure out what this was about
+ if containerchoice != False:
container = codecfinder.containermap[containerchoice]
containerelement = codecfinder.get_muxer_element(container)
if containerelement == False:
@@ -736,26 +745,37 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.vsourcecaps = videointersect
if self.haveaudio == True:
x=0
- while x < self.audiostreamcounter:
- x=x+1
+ count=len(self.audiostreamids)
+ while x < count:
if audiointersect[x].is_empty():
- audiointersect[x] = sourcecaps.intersect(self.audiodata['audiotype'])
- # print(audiointersect[x])
+ audiointersect[x] = sourcecaps.intersect(self.audiodata[x]['inputaudiocaps'])
self.asourcecaps = audiointersect[x]
+
+ if audiointersect[x].is_empty():
+ if len(self.audiopass)>=x+1:
+ self.audiopass[x]=False
+ else:
+ self.audiopass.append(False)
+
+ else:
+ if len(self.audiopass)>=x+1:
+ self.audiopass[x]=True
+ else:
+ self.audiopass.append(True)
+ x=x+1
+
+
+
if videointersect.is_empty():
self.videopass=False
else:
self.videopass=True
- if audiointersect[x].is_empty():
- self.audiopass=False
- else:
- self.audiopass=True
-
+
# define the behaviour of the other buttons
def on_FileChooser_file_set(self, widget):
self.filename = self.builder.get_object ("FileChooser").get_filename()
- self.audiodata = {}
+ self.audiodata = []
if self.filename is not None:
self.haveaudio=False #make sure to reset these for each file
self.havevideo=False #
@@ -795,20 +815,11 @@ class TransmageddonUI(Gtk.ApplicationWindow):
vwidth=False
ratenum=False
ratednom=False
- if self.haveaudio:
- achannels = self.audiodata['self.audiochannels']
- if self.audiopasstoggle == False:
- audiocodec = self.AudioCodec
- else:
- audiocodec = self.asourcecaps
- else:
- audiocodec=False
- achannels=False
self._transcoder = transcoder_engine.Transcoder(filechoice, self.filename,
- self.outputdirectory, self.codeccontainer, audiocodec,
+ self.outputdirectory, self.codeccontainer, self.audiodata,
videocodec, self.devicename,
- vheight, vwidth, ratenum, ratednom, achannels,
+ vheight, vwidth, ratenum, ratednom,
self.multipass, self.passcounter, self.outputfilename,
self.timestamp, self.rotationvalue, self.audiopasstoggle,
self.videopasstoggle, self.interlaced, self.inputvideocaps)
@@ -954,7 +965,7 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.passcounter=int(1)
self.StatusBar.push(context_id, (_("Pass %(count)d Progress") % {'count': self.passcounter}))
if self.haveaudio:
- if "samplerate" in self.audiodata:
+ if "samplerate" in self.audiodata[0]:
# self.check_for_elements()
if self.missingtoggle==False:
self._start_transcoding()
@@ -993,16 +1004,11 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.houseclean=True # set this to avoid triggering events when cleaning out menus
x=0
while x <= self.audiostreamcounter:
- #print("audiostream counter is " +str(self.audiostreamcounter))
- #print("x er " +str(x))
- #print("self.audiocodecs is " + str(self.audiocodecs))
if self.audiocodecs:
for c in self.audiocodecs[x]: #
self.audiorows[x].remove(0)
if x==self.audiostreamcounter:
self.audiocodecs=[]
- #print("is it clean " + str(self.audiocodecs))
- # print("is the menu clean " + str(self.audiorows))
if self.havevideo==True:
if self.codeccontainer != False:
for c in self.videocodecs:
@@ -1018,7 +1024,6 @@ class TransmageddonUI(Gtk.ApplicationWindow):
x=0
while x <= self.audiostreamcounter:
self.audiocodecs.append([])
- # print("self audiorows adding list " +str(self.audiocodecs))
if self.usingpreset==True: # First fill menu based on presetvalue
testforempty = self.presetaudiocodec.to_string()
if testforempty != "EMPTY":
@@ -1037,13 +1042,16 @@ class TransmageddonUI(Gtk.ApplicationWindow):
else:
audiolist = []
audio_codecs = supported_audio_container_map[self.codeccontainer]
- # print("vi skal fylle meny")
for c in audio_codecs:
- #print("x is " +str(x))
- #print("c is " +str(c))
self.audiocodecs[x].append(Gst.caps_from_string(codecfinder.codecmap[c]))
for c in self.audiocodecs[x]: # Use codec descriptions from GStreamer
self.audiorows[x].append_text(GstPbutils.pb_utils_get_codec_description(c))
+
+ if self.audiopass[x]==True:
+ self.audiorows[x].append_text(_("Audio passthrough"))
+ self.audiocodecs[x].append("pass")
+ self.audiopassmenuno=(len(self.audiocodecs[x]))-1 #FIXME
+
self.audiorows[x].set_sensitive(True)
self.audiorows[x].set_active(0)
x=x+1
@@ -1080,11 +1088,6 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.videorows[0].append_text(_("Video passthrough"))
self.videocodecs.append("pass")
self.videopassmenuno=(len(self.videocodecs))-1
-
- if self.audiopass==True:
- self.audiorows[0].append_text(_("Audio passthrough"))
- self.audiocodecs[x].append("pass")
- self.audiopassmenuno=(len(self.audiocodecs[x]))-1
def on_containerchoice_changed(self, widget):
self.CodecBox.set_sensitive(True)
@@ -1099,7 +1102,7 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.codeccontainer = self.builder.get_object ("containerchoice").get_active_text()
# self.check_for_elements()
if self.discover_done == True:
- # self.check_for_passthrough(self.codeccontainer)
+ self.check_for_passthrough(self.codeccontainer)
self.populate_menu_choices()
self.transcodebutton.set_sensitive(True)
@@ -1137,14 +1140,22 @@ class TransmageddonUI(Gtk.ApplicationWindow):
self.rotationvalue = self.rotationchoice.get_active()
def on_audiocodec_changed(self, widget):
+ name=widget.get_name()
+ if name.startswith("audiorow"):
+ x=name[8:]
+ x=int(x)
self.audiopasstoggle=False
if (self.houseclean == False and self.usingpreset==False):
- self.AudioCodec = self.audiocodecs[self.audiorows[0].get_active()]
- if self.codeccontainer != False:
- if self.audiorows[0].get_active() == self.audiopassmenuno:
- self.audiopasstoggle=True
- elif self.usingpreset==True:
- self.AudioCodec = self.presetaudiocodec
+ no=self.audiorows[x].get_active()
+ self.audiodata[x]['outputaudiocaps'] = self.audiocodecs[x][no]
+ if self.codeccontainer != False:
+ if self.audiorows[x].get_active() == self.audiopassmenuno:
+ self.audiodata[x]['passthrough']= True
+ else:
+ self.audiodata[x]['passthrough']= False
+ # print(self.audiodata)
+ elif self.usingpreset==True:
+ self.audiodata[x]['outputaudiocaps'] = self.presetaudiocodec
def on_videocodec_changed(self, widget):
self.videopasstoggle=False
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]
