Data Analysis & Stuff With Gautham Vasudev Menon 2
To be frank this blog has very little to do with its first part, Data Analysis & stuff with Gautham Vasudev Menon except for the fact that this blog revolves around something that goes synonymous with Gautham Menon, has some data analysis and I just wanted to let the new readers of the blog know that there is a prequel to it.
I remember writing the first one a couple of months post the release of Gautham’s last movie Achcham Enbathu Madamaiyada. Since then not a single Gautham movie came out. But during this period Gautham wasn’t idle, he had been very busy releasing teasers, trailers, songs, song teaser, song trailer, song lyrical, teasers of his friends’ movie, his assistant director’s movie, his productional movie, first look poster, motion poster, another teaser, another trailer, another song and the list just keeps growing. In simple words, he had just been teasing Tamilnadu with his teasers.
Meanwhile, Simbu has moved on with own his life and one of his movie AAA got released and ran very well. I meant running out of theatres. But this movie’s teasers garnered a good response. And in a recent singing reality show in which Simbu was called as a special guest, he mentioned how AR Rahman worked on the teaser of Achcham Enbathu Madamaiyada and released it and helped him while the time he was surrounded by the beep song controversy. I am one of those guys who doesn’t believe that Simbu is controversy’s child. I believe the other way round which is a thought that controversies are Simbu’s children.
Given the number of controversies that arise around Simbu, he needs a fairly large amount of teasers under his belt so that he can release one and cool the air around whenever it is needed. In a way what brings both Simbu and Gautham together is the common obsession they have with teasers. But everyone would not have the same interest. If will be unfair to knock AR Rahman’s and editor’s door on a midnight and ask for their help to create a teaser for the 213th time in the same month. So let’s help Simbu and Gautham to generate teasers for their movie with help of some program and some sampling techniques.
Before jumping into anything let’s take a look at the teaser of the last Gautham Menon movie which got released.
Such a nicely made teaser, which would raise the expectation level of anyone watching it. We will use this teaser as the base and try to learn to create teasers from the movie itself.
This is how the teaser was cut. With first 40 seconds of it being breezy, the next 20 seconds being bloody and ends with a motion poster of the movie. The ones in black dots represent the name cards of Simbu, Gautham, and the movie respectively. Let’s retain the black portion’s from the original teaser as that is not something that will change and replace the other parts alone.
With some trial and error understood that each scene cut must be at least 3 seconds long in order to give a good appeal to a teaser. When it is lesser than that it wasn’t visually appealing as it looked more like a picture rather than a moving image. So lets sample scenes of 3 seconds long from the movie file and try creating teasers. When it comes to sampling statistically there are 3 major techniques.
- Simple Random Sampling
- Systematic Random Sampling
- Stratified Random Sampling
Let’s take each of these methods one by one and try building a teaser.
Simple Random Sampling
This is a kind of sampling where samples are taken randomly from the whole available set. For us, the whole set starts after 30 seconds of movie runtime after Gutka Mukes makes a cameo and goes on till 6600 seconds. Let’s ignore the last 10–15 mins of the movie as contains something which makers believe is a plot twist and revealing it in our teaser may spoil it for the audience. So our sample set is 30 seconds to 6600 seconds of movie runtime. Let’s sample 18 scenes of 3 seconds each and stitch it with the title cards that we snipped from the original teaser in order to programmatically create a teaser that inherits the style of the original. Before moving any further, I promise that all these teasers were generated purely randomly, I did not try to tamper any of them and there was no foul play involved after all I am not a fan of the Aussie cricket team.
The scenes were cut into 3 seconds each randomly and spliced programmatically so expect some lag. So here is how the teaser looks.
Oh! wait doesn’t it look like a teaser of some Bala movie. Oh Shit! I forgot to add the secret sauce to the teaser which gives the feel of it. ARR’s BGM. Let’s watch the same teaser again but this time with BGM.
Ah, that looks much better! But still, there is something that this teaser lacks which mainly because it was made using simple random sampling from the whole set, the Breezy bloody mix from the original teaser wasn’t captured well. Dear Simple Random Sampling, you are way too simple for Simbu. Let’s move on.
Systematic Random Sampling
This a type of probability sampling method in which sample members from a larger population are selected according to a random starting point and a fixed periodic interval. I let the program choose the random starting point and the fixed periodic interval was set to 300 seconds.
Much much better than the first one. I am astonished how each and every frame of the movie goes well with the BGM. Not sure whose Magic it is ARR’s or Gautham’s or Simbu’s! But still I am not a big fan of this sampling technique as you only randomly get to choose the first sample and from there everything happens periodically. You give much information about the flow of the movie with this sampling technique. There isn’t much randomness after all.
Stratified Random Sampling
This method of sampling that involves the division of a population into smaller groups known as strata. In stratified random sampling or stratification, the strata are formed based on members’ shared attributes or characteristics. Based on the style of the movie and teaser let’s divide them into two buckets and pick a sample within each bucket randomly and then stitch it together to form a teaser.
30 Seconds to 3180 Seconds, which is the point where the movie is breezy forms a bucket and from that point to 6600 seconds form another bucket. Let’s pick samples randomly from this bucket maintaining the ratio from the original teaser.
And that is how it looks!
But wait, the blog doesn’t end yet. We have so far dealt only with creating teasers programmatically we have not answered the bigger question of when will the next Gautham Menon movie get released? How many teasers are yet to get released before we actually get to see the movie on the big screen?
Let’s try to answer that question with numbers we have from his last movie. Let’s assume that even the next movie of his follows the same pattern of story and runtime and he wishes to cut the teaser in the same style. Which tells us that between 30–3180 there are 3148 ways to cut a scene that is 3 seconds long.
Eg. If you want pick 3 numbers serially from 1,2,3,4,5 there are 3 ways to do it. 123,234, 345.
Since the breezy half is 39 seconds long with 3 seconds for the Simbu title card we still have to chose 12 scenes of 3 seconds long randomly. Which means the number of possible combinations that are possible is 3148C12.
Similarly the number scenes that has to be cut for second half is 6 ignoring the title cards. The sample set is 6600–3180 = 3420. Out of which there are 3418 ways to cut a scene which is 3 seconds long. Which gives us the total number of possible combinations in the second bucket to be.
So the total combinations that are possible for this teaser are the product of these two numbers which is 1936258457691910708420795771278953 * 2204920401297362854
Don’t sleep yet. Just googled and found some astonishing youtube stats. Which claims Almost 5 billion videos are watched on Youtube every single day. Being the users of youtube we know that each video has a suggestion list on their right-hand side. They are 20 in number.
Gautham Menon gets to know this. He decides to populate 20 different combinations of his teasers as suggestions for every youtube video that is being played on every single day. 20 different new combination of teaser for each video! get that around your head.
In that case, he will be able to survive just by releasing teasers for quite sometime before he runs out of combinations and actually gets to release his movie. Let’s make an estimate of that quite some time.
(1936258457691910708420795771278953 * 2204920401297362854 Combinations)/5,000,000,000,000 Videos per day * 20 Suggestions per video
42692957755494606347282799097166629951 Days
42692957755494606347282799097166629951/365 Years
Which is 116967007549300291362418627663470219 years before he runs out of combinations. Which in words can be read as
The Number that is written in blue alone beats the age of the Earth by around 3 billion years.
After all, we live in a world that is run by Gautham Vasudev Menon, where time is measured in teasers and everything else is just an illusion.
