[0:00] hi there in today's video I'm once again
[0:02] working on my capacity project which is
[0:05] an automatic YouTube short caption
[0:08] generator now to recap this project
[0:11] which has now been going on for several
[0:13] videos it started out with this short
[0:16] trity project which is an AI generated
[0:19] YouTube short generator so basically it
[0:22] takes in a text input and then it
[0:24] generates a YouTube short based on that
[0:28] and I can actually show you a demo so we
[0:30] can create a new file let's call it
[0:33] prompt. text and here we put the prompt
[0:36] for the short this can be either just a
[0:38] description of the YouTube short you
[0:40] want to create or it can be just a copy
[0:42] pasted news article or whatever text
[0:45] really let's try something like HTML 5
[0:49] features you've never heard of so we
[0:53] will save this and then we can run our
[0:56] main.py and we can give it this prompt.
[0:59] text and then it will just generate the
[1:01] whole thing and if we then open this
[1:04] thing in VLC for example then it will
[1:08] look like this HTML 5 a revamp of the
[1:12] classic HTML has brought many
[1:14] advancements to web development but some
[1:16] fascinating features might have slipped
[1:18] under your radar features like the micro
[1:20] dat API a method to integrate metadata
[1:23] into your sites or content security
[1:26] policy enhancing protection against
[1:28] excess attacks and webq L database
[1:30] although somewhat controversial can be a
[1:32] useful tool for web-based applications
[1:35] several Hidden Treasures lurk within
[1:36] HTML 5 enhancing the prowess and
[1:39] potential of web developers everywhere
[1:41] now as you might have noticed we have a
[1:43] couple of issues with this the first
[1:45] issue is that the captions are not quite
[1:48] synced to the video or to the narration
[1:51] and that is because we use kind of a
[1:52] dumb way of syncing the narration with
[1:55] the captions so basically we generate
[1:57] the narration one sentence at the time
[2:00] and then we check the duration of that
[2:03] narration audio and we calculate the
[2:05] duration per character and then we show
[2:08] every word in the video one by one based
[2:11] on how many characters there are and it
[2:14] is surprisingly accurate but still not
[2:17] that accurate so hence I created this
[2:20] capacity project and this one takes an
[2:24] existing video with a narration and then
[2:27] it transcribes the audio with open AI
[2:30] whisper and gets the timestamps of every
[2:32] word in the narration and that way we
[2:35] can sync the narration pretty much
[2:37] exactly correctly so if I go quickly
[2:40] back to the short trity project and I go
[2:43] to my text module and I disable the
[2:47] drawing of the captions and then I
[2:49] generate this short again then now we're
[2:52] going to have a short without the
[2:54] captions so let's move this file into
[2:58] the capacity project
[3:01] and then let's go into the capacity
[3:03] project and add the captions to this so
[3:06] here we can run again main.py and we
[3:09] give it the video file which in this
[3:11] case is short. Avi then it will
[3:14] transcribe this and it will add the
[3:16] captions to it and if we now play this
[3:20] with transcript. Avi which is the output
[3:23] of this script it will look like this
[3:27] ever wonder what HTML 5 might be hiding
[3:29] and its depths there's more to it than
[3:31] just to facilitate webpage display did
[3:33] you know it features geolocation
[3:35] abilities allowing sites to know your
[3:37] location or how about its video and
[3:40] audio elements making for easy
[3:42] multimedia integration perhaps the most
[3:44] fascinating is the offline storage
[3:46] feature enabling the browser to store
[3:48] your data diving into HTML FES talents
[3:51] indeed gives A New Perspective into the
[3:53] world of web development so now the
[3:55] captions are synced better and we can
[3:58] actually see multiple work words at a
[4:00] time and it in fact checks how many
[4:03] words fit into the video on two lines
[4:06] and that's how many it's going to use
[4:08] which brings us to the latest video and
[4:12] the subject of today's video so the
[4:15] captions you just saw were added to the
[4:18] video with open CV and open CV is not
[4:22] that great at styling text so we are in
[4:25] the process of moving this to movie pie
[4:28] and in the previous video I started
[4:31] already this movie pie test.py script to
[4:34] try and style the text in the videos and
[4:38] this script takes in a video file let's
[4:40] change this to the short. Avi that we
[4:43] have here and it adds some styled text
[4:46] to it so we have this create text
[4:49] function into which we can pass in the
[4:51] text and a font size and a color and a
[4:54] font and a background color and a blur
[4:57] radius which is used for generating
[5:00] shadows and an opacity and a stroke
[5:02] color and stroke width and cerning and
[5:05] most of these things are just passed
[5:08] into this text clip class but because
[5:11] blurring is not supported in movie pie
[5:13] then we add this blur radius parameter
[5:17] here and if we have a blur radius then
[5:19] we take the movie pie text clip and we
[5:23] convert it into the pillow format so
[5:25] we're using the pillow Library which can
[5:28] handle images and then we add just a
[5:30] blur effect to it and then we convert it
[5:33] back into a movie pie clip and basically
[5:36] that allows us to create Shadows for the
[5:38] text so here we are defining the text we
[5:41] want to draw and the font we want to use
[5:44] and then we create the text clip of the
[5:46] actual text and then we create the
[5:49] shadow clip and then we create this
[5:51] composite video clip into which we add
[5:53] the video the original video and then
[5:56] the shadow clip and the shadow clip
[5:58] again to make the shadow or stronger and
[6:01] actually three times and then we add the
[6:04] text to it so these are basically layers
[6:07] in the video and right now we are just
[6:10] saving the first frame of this just to
[6:12] test it out but we can also write the
[6:14] full video so if I run this now Python 3
[6:19] movie test.py then this should create
[6:22] one frame of our short with the text
[6:26] subscribe in it so if I open output. PN
[6:29] G then this is what it looks like so now
[6:33] after 30 minutes of recap which
[6:35] hopefully I can condense down to a
[6:37] couple of minutes we can get to the
[6:39] point of today's video which is to
[6:42] actually integrate this styling of the
[6:45] text into capacity so let's do it and
[6:49] I'm in fact thinking about making this
[6:51] another a third project just for writing
[6:54] styled text on video with python because
[6:57] there doesn't seem to be a very good
[6:59] solution solution for that at the moment
[7:01] but for now I will simply move this
[7:04] function into a separate module so let's
[7:07] create a new module here and let's call
[7:10] this I don't know text drawer. PI in
[7:14] lack of a better name and we just want
[7:17] to add this function here and this one
[7:21] but we don't want to do that we only
[7:24] want to have these functions in there
[7:26] and we in fact don't need to import the
[7:29] video file clip we have to import this
[7:32] in the main application and I guess we
[7:35] don't need the composite video clip we
[7:37] need that again in the main application
[7:40] and this one is actually just a
[7:42] hallucination by jat GPT so we don't
[7:45] need that either and this comment does
[7:49] not belong here either and we don't have
[7:51] to save this test. PNG in here by the
[7:55] way if you know how to do this without
[7:57] having to save theame frame from movie
[8:00] pie and then opening the file with
[8:02] pillow then feel free to make a pull
[8:05] request because I'm not really happy
[8:06] about how I did it so basically this
[8:09] converts from the movie pie clip format
[8:11] into the pillow image format anyway
[8:14] let's save this now and let's go to our
[8:17] main function and let's find out where
[8:20] we are actually drawing the text here we
[8:24] have a right line function which takes
[8:28] the text and a frame and a text y a font
[8:32] font scale white color black color
[8:35] thickness and a border and then it
[8:38] calculates the size of the text to make
[8:41] it centered and then it just puts the
[8:44] text into the frame and here we're kind
[8:47] of doing a shadow again so we are
[8:49] drawing it twice and this is more like a
[8:51] stroke than a shadow because we draw it
[8:55] first in black and the thickness is
[8:58] going to be the regular thick
[9:00] plus border * 2 so in CV2 this results
[9:03] in kind of a stroke and where do we use
[9:06] this right line function we use it here
[9:10] so this is the loop where we read every
[9:14] frame of the original video and then we
[9:17] Loop through all of the captions and
[9:19] then we get the line data from calculate
[9:22] lines which takes the text and the frame
[9:26] width and from here we are going to get
[9:29] the height and the lines so the height
[9:33] is apparently the height of one line and
[9:36] lines is a list of the texts in the
[9:39] lines okay and then we write all the
[9:42] lines on the frame so we of course have
[9:46] to rewrite this calculate lines function
[9:49] because we are not using CV2 anymore
[9:52] let's see what is calculate lines it
[9:55] takes the text and the frame width and
[9:57] basically we calculate the size of the
[10:00] text so the width and height and this is
[10:03] done with CV2 so we have to figure out
[10:06] how to do this in movie pie so let's go
[10:09] to our text drawer and let's create that
[10:12] function let's define get text size and
[10:18] let's see what copilot says it wants the
[10:21] text the font size the font and a stroke
[10:24] width which I guess makes sense because
[10:27] those are all the things that matter in
[10:29] the calculation of the width and the
[10:32] height of the text so let's do that and
[10:35] we are going to create the text clip
[10:37] into which we pass the text and the font
[10:40] size and the color which color doesn't
[10:42] really matter so here we use White and
[10:45] the font and the stroke width okay and
[10:48] then we convert that into the pillow
[10:50] image and then we get the size of the
[10:53] image which actually makes sense
[10:55] although it just seems like kind of a
[10:58] wasteful operation again because we are
[11:00] saving it into a file and we might be
[11:04] doing this quite a few times in the code
[11:07] but let's use this to begin with and
[11:09] let's see what the size actually is so
[11:12] if this returns an image what is the
[11:14] definition of an image we have the image
[11:17] class here and it has a size which I
[11:21] guess is going to be width and height
[11:24] okay and in our
[11:27] main we also have width and height the
[11:30] same way so that will work the same so
[11:33] this has to be get text width which we
[11:37] have to import from our text drawer so
[11:40] from text drawer import get text size
[11:45] which actually yeah it's not text width
[11:48] it is text size and we don't have to
[11:51] take any zero element from there but we
[11:55] need to change this so we have font font
[11:57] scale and thickness but we want font
[12:01] size font and stroke width okay so we
[12:04] just move these around let's rename font
[12:09] scale to font size and let's use that
[12:14] 120 was I using that one what was the
[12:18] what was in the example here 120 okay
[12:21] let's use that and we pass in the line
[12:25] the font size and the font and stroke
[12:28] with WID so let's get rid of thickness I
[12:32] just call it stroke width and what were
[12:36] we using here two so let's use two
[12:40] stroke width is going to be two and by
[12:43] the way these are right now global
[12:45] variables which is not that great but we
[12:48] have to refactor that at some point so
[12:50] that should in fact work we get the text
[12:52] size and the rest of this should just
[12:55] work the same because we just check the
[12:56] size and then if the width is greater
[12:59] than the frame width then we are going
[13:02] to add to the lines to write the current
[13:06] line to draw if we already have some
[13:09] text that we have drawn already or that
[13:12] we tried to draw but it didn't fit
[13:14] anymore so then we put it there okay
[13:16] that makes sense they should work and
[13:18] then the right line function now
[13:21] currently we are using the frame which
[13:24] is a CV2 frame but in movie pie this
[13:27] works a little bit differently
[13:29] now we might at first just literally
[13:33] create the image of the text and then
[13:35] just put that in the frame with CV2
[13:39] which will be very slow I'm guessing but
[13:42] let's start with that so again we have
[13:44] to get the text size which we will do by
[13:47] get text size and we pass in the text
[13:51] and the font size and then the font and
[13:55] the stroke width okay and no zero and
[13:58] then we do the same thing here so we
[14:00] take the frame shape one which is the
[14:03] width of the frame of the video frame
[14:06] and then we subtract the width of the
[14:10] text and then we make this origin point
[14:14] which is text X text y but then we want
[14:18] to actually get the image and presumably
[14:22] there is some sort of put image is there
[14:25] such a thing let's see how this works um
[14:29] image we can't find image here so let's
[14:32] actually ask chat gbt how do I put an
[14:37] image in the frame in a specific
[14:41] position in CV2 we are going to say I am
[14:47] read so we read the frame we read the
[14:51] image and we are going to say frame y
[14:55] offset y offset Plus image height Okay
[14:58] so so I think what happens here is that
[15:03] it is actually going to replace that
[15:05] part of the frame with the image which
[15:07] is not what we want because we want to
[15:10] have it on top so we can't really do
[15:13] that so it's probably easier to just use
[15:16] movie pie so let's abandon this idea of
[15:18] using CV2 here and we have to then
[15:23] refactor this part as well so let's open
[15:25] the video file which we do like this so
[15:30] this we have to do in main over here and
[15:33] none of this stuff and where do we
[15:35] actually Define the name of the file it
[15:39] is video file let's still put it here
[15:43] and let's come back to these later we
[15:45] still like get the frame rate and the
[15:47] frame width and height with CV2 but let
[15:50] that be like that for now and we have to
[15:52] import this video file clip so let's do
[15:55] that from movie pie editor
[15:59] import video file clip and here we are
[16:02] reading the frames with CV2 again and
[16:06] for every frame we get the time which is
[16:11] calculated by adding to it one over
[16:14] frame rate and what do we do with the
[16:16] time we just see if the captions start
[16:21] and end is within that time and what is
[16:24] actually in captions where do we get
[16:26] this okay we get this from the segment
[16:28] part
[16:29] which takes the segments which are parts
[16:32] of the narration transcribed by Whisper
[16:35] and we pass this fit function which
[16:38] basically checks how much text fits in
[16:41] the frame so these captions are already
[16:45] the captions we are going to put on the
[16:47] video and since movie pie Works
[16:50] differently we can probably implement
[16:52] this in a completely different way so we
[16:54] are not going to go frame by frame over
[16:58] the video
[16:59] we are just going to Loop through all of
[17:02] the captions and in movie pie we had
[17:06] this thing called set duration so we set
[17:11] the duration of a clip but how do we set
[17:14] the start time let's ask chat gbt how
[17:17] does that work how do I set the start
[17:21] time of a clip in movie pie okay set
[17:26] start that makes sense so this is
[17:29] actually going to be very easy we just
[17:32] go through all the captions and we
[17:35] create the text so text equals create
[17:39] text was that what I call it create text
[17:43] yes and we pass in all of this stuff
[17:46] like this and this was called actually
[17:48] font size like this and we don't have a
[17:51] color let's call this font color and set
[17:55] it up here right now font color let's
[17:59] just do I don't know white just to test
[18:02] it out I mean I guess we can use the
[18:04] color that we tried in this file which
[18:08] is that so font color will be this and
[18:12] the font is not actually the CV2 font
[18:15] let's again use this font and we create
[18:20] the text which we have to import now so
[18:23] create text and this I believe returns
[18:29] an image clip or a text clip which is
[18:33] clip let's see how this is defined it is
[18:36] a video clip so let's actually Define it
[18:39] here that this is going to return a
[18:41] video clip which we have to import from
[18:44] here so this is the video clip and the
[18:49] text we pass in here is actually caption
[18:52] text font size is fine font color is
[18:54] fine font is fine background color let's
[18:58] say the transparent now blur radius zero
[19:01] opacity one stroke color okay that's
[19:03] fine at the moment did we actually set
[19:06] the stroke width somewhere we did so
[19:08] let's set that to the stroke width
[19:11] stroke width and we have to set the
[19:14] stroke color let's do stroke color and
[19:18] actually background color is transparent
[19:20] by default so we don't have to set that
[19:23] and blur radius is zero by default we
[19:25] don't have to set that opacity is one by
[19:27] default we don't have to set that that
[19:29] either so we just pass in the stroke
[19:30] color and the stroke width and curing is
[19:33] zero as well so let's just Define stroke
[19:37] color and stroke color is black so this
[19:40] is going to be our text clip and then we
[19:44] can say text equals text set duration
[19:48] and set start why doesn't it suggest
[19:52] text equals text. set start caption
[19:56] start wait a minute did chat gity lie to
[20:00] me let's actually ask about a text clip
[20:04] how do I set the start time of a text
[20:07] clip in mipie what time the text clip
[20:12] should start in a composite video that
[20:18] is what I mean to set the start time of
[20:20] a text clip in movie you can use a set
[20:22] start method on the text clip but it
[20:25] doesn't seem like we have that thing or
[20:29] H let's go to the definition of this set
[20:33] duration do we have set start we have
[20:37] set start so vs code is just being
[20:39] annoying so we can set the start and we
[20:42] can set the duration I want to set the
[20:44] start first okay what so now set
[20:47] duration is not defined set start
[20:50] returns any that is the reason it
[20:53] returns any so vs code doesn't know
[20:55] anymore what happens fine so this will
[20:58] will now write the text in the correct
[21:02] position in the video but it will write
[21:05] the whole text on one single line so we
[21:09] do need this calculate lines function
[21:11] but we don't need this part we just do
[21:14] that so basically this has to be moved
[21:18] inside of this for line in line data
[21:22] lines we are going to do that and we
[21:25] have to do text equals text do set
[21:30] position and we can Center it
[21:32] horizontally easily but then we need to
[21:34] position it in the y direction because
[21:37] we want multiple lines now I wonder does
[21:40] movie pie support new lines can we add a
[21:43] new line here that would simplify this a
[21:45] lot let's try if we now Run movie pie
[21:49] test will it actually write two lines
[21:53] that would be pretty amazing if we now
[21:55] check our output then it in fact
[21:59] does support new lines so this is going
[22:02] to be super easy let's go back to the
[22:05] text drawer sorry to the main and we are
[22:09] just going to do this we calculate the
[22:11] lines and we only draw The Thing Once
[22:16] down here and I'm not sure what this
[22:19] break does here we don't need this so we
[22:21] draw the thing once but here we set like
[22:25] lines to be an empty string and for
[22:29] every line we just say that lines plus
[22:33] equals line plus a new line which I
[22:37] guess we can just say that lines equals
[22:41] new line join line data lines I think
[22:45] that's how it works and then we can just
[22:48] write the lines directly here so that
[22:51] simplifies it now I wonder is there even
[22:55] an automatic line break can I set the
[22:59] width of a text clip let's see the
[23:01] definition here and is there like set
[23:04] width or set bounding box
[23:07] bounding there is no bounding box I
[23:10] think we still have to like do our own
[23:13] new lines so now we calculate the lines
[23:16] we get the text Y which we actually
[23:18] don't need so this can be now just
[23:21] Center and we don't need this stuff and
[23:24] we don't even need the line height we
[23:27] just want the line
[23:29] okay and we have to gather these text
[23:32] clips into some list so let's put like a
[23:35] Clips list in which we have the video
[23:39] and then we're going to add the text
[23:42] into the clips and then we need to do
[23:45] this stuff we have to do composite video
[23:48] clip which conveniently takes a list so
[23:51] we can do this after that video with
[23:55] text is composite video clip with all of
[23:58] the clips and we have to import
[24:01] composite video clip from here and what
[24:04] else currently we are using FFM Peg to
[24:09] combine the video and the audio and this
[24:11] is because CV2 doesn't support audio but
[24:15] I guess now we can just do this so we
[24:20] call on the video with text right video
[24:23] file output video and let's put 30 FPS I
[24:27] think or or original video is 30 FPS and
[24:31] will this work uh let me check the fits
[24:34] frame function how does this work so
[24:37] basically it uses the calculate lines
[24:39] with the text and the frame width and it
[24:41] just checks if there are fewer than or
[24:44] equal to two lines then the text will
[24:47] fit because that is our rule right now
[24:49] we have to fit two lines of text maximum
[24:53] okay that should work and do we even use
[24:56] right line we don't use right line we
[24:57] can re move right line and calculate
[25:01] lines should now work because we get the
[25:03] text size and then based on that we
[25:06] calculate how many lines fit on the
[25:09] screen and we don't need to do this
[25:12] extract audio on video because movie pie
[25:14] supports video or audio automatically
[25:18] and this is just a whisper
[25:20] transcription and we don't use the oh
[25:23] actually we want to use the frame rate
[25:25] which we probably don't want to use CV2
[25:27] to get but let's just put it there FPS
[25:30] is the frame rate and frame height we
[25:32] don't use at all right now so let's
[25:35] remove that and I'm almost ready to try
[25:38] this out we go through all the captions
[25:41] we calculate lines which is kind of
[25:43] redundant because we do this already
[25:45] when we calculate the captions we might
[25:47] want to save that over there already but
[25:50] basically this just splits the caption
[25:53] into lines and then we draw the lines
[25:56] okay I am ready to test this let's run
[26:01] main.py Python 3
[26:03] main.py and did we give it the file how
[26:08] did this work sorry I'm an idiot of
[26:11] course we need to extract the audio
[26:12] because the whisper needs the audio so
[26:17] let me go back a little bit and take
[26:19] this extract audio and let's put that
[26:23] back here and what this video file okay
[26:27] this Arc V1 so let's do it main.py and
[26:31] short. AI which is the one without text
[26:34] and then let's fix all the errors that
[26:36] happen here of which there were zero
[26:40] zero errors and now we're writing the
[26:43] video file so let's see what happens
[26:47] okay it is now done that was way slower
[26:50] than CV2 kind of surprisingly and the
[26:54] actual slow part was the rendering of
[26:55] the video so none of my code was was
[26:58] really slow but the rendering part now
[27:01] my code might have something to do with
[27:02] that but let's see what this video looks
[27:05] like let's run with VLC output video.
[27:10] MP4 ever wonder what HTML 5 might be
[27:13] hiding in its depths there's more to it
[27:15] than just to facilitate webpage display
[27:17] did you know it features geolocation
[27:19] abilities allowing sites to know your
[27:21] location or how about its video and
[27:24] audio elements making for easy
[27:25] multimedia integration perhaps the most
[27:28] fascinating is the offline storage
[27:30] feature enabling the browser to store
[27:32] your data diving into HTML 5es talents
[27:35] indeed gives A New Perspective into the
[27:37] world of web development okay it
[27:40] actually worked first try I can't
[27:43] believe it something weird happened in
[27:45] the end I'm not sure if that was a bug
[27:47] in VLC or what happened with the audio
[27:51] but yeah it actually worked interesting
[27:54] now let's see if we can modify it what
[27:57] if we want to have just one line of text
[28:00] is it still going to work if I change
[28:02] here just one and then we run it again
[28:06] so now we should have only one line
[28:09] visible at a time in the video but it
[28:12] should still show all of the
[28:14] transcription and let's take a look at
[28:17] this ever wonder what HTML 5 might be
[28:20] hiding in its depths there's more to it
[28:22] than just to facilitate webpage display
[28:24] did you know it features geolocation
[28:26] abilities allowing sites to know your
[28:28] location or how about its video and
[28:30] audio elements making for easy
[28:32] multimedia integration perhaps the most
[28:35] fascinating is the offline storage
[28:36] feature enabling the browser to store
[28:39] your data diving into HTML 5es talents
[28:42] indeed gives A New Perspective into the
[28:44] world of web development yeah it works
[28:47] just as expected now one thing I would
[28:51] like to do is have some sort of padding
[28:54] there because sometimes the text goes a
[28:57] bit too close to the borders so let's
[29:00] add a padding and let's call it I don't
[29:04] know what we should call it is it in
[29:05] pixels I guess and the video is pretty
[29:08] big so let's do 50 pixels of padding and
[29:11] then our fits frame function will
[29:14] calculate the lines with frame width
[29:17] minus padding does that make sense maybe
[29:21] we want to just pass it in here so when
[29:24] we call fits frame we pass here frame
[29:27] withth
[29:28] minus padding like this but we also have
[29:32] to pass it in here so maybe we want to
[29:35] say something like text bounding box
[29:41] with uh that's so long so this would be
[29:44] frame width minus padding let's just
[29:47] call it bbox text bbox width that's not
[29:51] too long so then we can pass those in
[29:53] here and let's actually check out our
[29:56] competitor sub
[30:00] magic.co what kind of size do they use
[30:04] for these
[30:06] videos so well they have different kind
[30:08] of sizes but how about in the Mr Beast
[30:12] video you can basically fit algorithm
[30:17] rewards so how many characters is that
[30:21] algorithm rewards and in
[30:26] our we can fit ever wonder what HTML 5
[30:30] ever wonder what HTML 5 and we should
[30:36] have some padding so maybe two
[30:39] characters from each side we should drop
[30:42] so drop these two and these two so it's
[30:45] basically the same size we can make it a
[30:48] little bit bigger so we can say over
[30:51] here that our font size is 130 and I
[30:56] think we can add some stroke let's put
[30:58] like four to the stroke and what is in
[31:01] fact margin we don't use it so this was
[31:04] used with the CV2 for the line like
[31:08] margin between the lines but now it's
[31:10] just the new line so I guess that is
[31:14] something that we can't now change
[31:16] unless we can change it directly in
[31:18] movie pie and Border five what is that
[31:21] we don't use it and of course white
[31:23] color and black color we don't use
[31:25] anymore because now they are in this
[31:27] different format okay and I do want to
[31:31] get rid of CV2 now I don't want to
[31:34] import CV2 un necessarily and we are not
[31:37] using math so where do we use CV2 we get
[31:41] the frame rate from it and the frame
[31:44] width so how do we do that in movie pie
[31:47] how do I get the frame rate and width of
[31:53] a video with movie Pi we can use
[31:58] video. FPS and video. W okay so this is
[32:03] going to be video. W and frame rate we
[32:08] are using just here it is video. FPS and
[32:13] we don't have to destroy CB2 windows and
[32:16] we don't have to get the FPS here and we
[32:19] don't have to get the cap and we have to
[32:22] do this after the video which means we
[32:26] have to do this after after that so this
[32:29] is opening the video and then we
[32:31] calculate this bounding box wi and then
[32:34] we set the clips to the video and then
[32:37] we get the captions so now we are not
[32:40] using CV2 anymore we are just using
[32:43] movie pie which means our requirements
[32:47] shall not have CV2 anymore only movie
[32:50] pie and pillow and Whisper great so
[32:54] let's run it one more time are we
[32:58] getting rid of this temp audio file
[33:00] right now um yeah so it is actually a
[33:04] temp file okay how about temp video file
[33:08] are we even using it we are not using it
[33:10] and we actually Define output file but
[33:12] we don't use it so let's say that this
[33:15] is the output file and do we need to do
[33:19] anything else that's basically it and it
[33:21] works now we did have this position here
[33:25] um caption y pause so in a previous
[33:29] video I added this kind of like
[33:31] multiplier for the position which is
[33:33] kind of a stupid idea anyway so we might
[33:36] just say position and we can put here
[33:40] Center Center and then we can pass it
[33:43] directly here like this so then you can
[33:47] change the position if you want okay so
[33:50] now we can change all the settings here
[33:53] which we might want to move in some sort
[33:55] of configuration file or something and
[33:57] actually pass them in as parameters to
[34:00] some I don't know a class or some
[34:02] function at least but let's try it again
[34:04] and see if it works nicely ever wonder
[34:08] what HTML 5 might be hiding in its
[34:10] depths there's more to it than just to
[34:12] facilitate webpage display did you know
[34:14] sorry this is the wrong one because now
[34:17] I Chang it to a different file so it is
[34:20] now with transcript. Avi ever wonder
[34:24] what HTML 5 might be hiding in its
[34:26] depths there's more more to it than just
[34:28] to facilitate webpage display did you
[34:30] know it features geolocation abilities
[34:33] allowing sites to know your location or
[34:35] how about its video and audio elements
[34:37] making for easy multimedia integration
[34:40] perhaps the most fascinating is the
[34:41] offline storage feature enabling the
[34:44] browser to store your data diving into
[34:46] HTML 5es talents indeed gives A New
[34:49] Perspective into the world of web
[34:51] development okay it seems to work pretty
[34:54] well but there's again in the end I
[34:56] heard some weirdness in there and I
[35:00] think we can add a little bit more
[35:03] padding because at some point it went a
[35:06] bit too far in my opinion let me try to
[35:10] see the place like here it is a little
[35:13] bit too close to the edges so maybe even
[35:15] 100 pixels of padding and well yeah I
[35:19] guess we should do times two because I
[35:22] want 50 pixels on each side and I really
[35:25] find this very annoying that movie pie
[35:28] doesn't really support a proper stroke
[35:31] because the stroke is in the middle so
[35:34] it will make the text thinner which I
[35:37] don't like but I don't have a solution
[35:39] for that right now as I tried in the
[35:41] previous video it doesn't help if I draw
[35:43] the text twice because the thickness
[35:46] doesn't work in the same way as in CV2
[35:49] in movie pie so let's do that as the
[35:52] last thing I will put a stroke with
[35:55] three and I will make the padding times
[35:59] two so then it should look pretty nice
[36:02] and in the next video I am going to
[36:05] implement the highlighting of the
[36:07] current word so that the current word
[36:09] will show in a different color and maybe
[36:12] even a different size maybe it will
[36:14] become bigger for a while when it's like
[36:16] highlighted and I probably want to like
[36:18] zoom in and out the background image or
[36:21] pan it somehow or do something like that
[36:24] but that will be in the next video so
[36:27] make sure to subscribe if you want to
[36:29] see that one but let's now try this one
[36:32] more time and let's do a little bit of
[36:34] modifications to the style so what if I
[36:38] want let's say a red color and then
[36:43] let's try one other font that I have we
[36:47] have poetsen one which I can't view here
[36:51] but let's see how that looks like
[36:53] poetsen one regular and if we we run
[36:59] this thing with those settings let's see
[37:02] what it will look like ever wonder what
[37:04] HTML 5 might be hiding in its depths
[37:06] there's more to it than just to
[37:08] facilitate webpage dis did you know it
[37:10] features geolocation abilities allowing
[37:12] sites to know your location or how about
[37:15] its video and audio elements making for
[37:17] easy multimedia integration perhaps the
[37:19] most fascinating is the offline storage
[37:21] feature enabling the browser to store
[37:24] your data diving into HTML fives talents
[37:27] in indeed gives A New Perspective into
[37:29] the world of web development okay that
[37:31] wasn't the greatest font or the greatest
[37:34] color but at least we can modify the
[37:38] colors and the fonts so hopefully you
[37:41] like this video and if you did then make
[37:43] sure to subscribe and hit the like
[37:46] button and also give me a comment down
[37:48] below and let me know what video should
[37:51] I do next and if you want to try out
[37:54] capacity then you can find it on my
[37:56] GitHub page p and by the way it is not
[37:59] open CV anymore now it is movie pie
[38:03] anyway thanks for watching and I will
[38:04] see you in the next one