Netflix considers mobile specific versions of original content
Netflix may make it even easier for its subscribers using iPhones to follow its original content in the future, an executive for the streaming service has revealed, advising that the firm is currently exploring the possibility of making mobile device-specific versions of its shows.
Due to having a smaller screen compared to a television or a notebook, the framing of some shots are sometimes harder to see on smartphones, making the experience less immersive or difficult to follow visually. A second version of the same show, possibly using alternate shots or different cropping, could help alleviate this issue for mobile users, while still keeping the rest of the content's components intact, such as speech and pacing.
Chief product officer Neil Hunt told The Verge the company is looking into using a different or reframed shots for content due to the increased usage of mobile devices to view the platform's shows and films.
"It's not inconceivable that you could take a master and make a different cut for mobile," said Hunt. While Netflix has yet to offer different cuts of content to its users as yet, Hunt said "it's something we will explore over the next few years."
Hunt's remarks were made during a two-day event at Dolby Laboratories and Netflix's headquarters in preparation for the release of Iron Fist, a new Netflix show that will be available to subscribers on Friday. Also during the event, executives from both companies talked about high dynamic range (HDR) content, and how it could help improve the experience for mobile users, as well as those with compatible televisions.
A mobile-friendly cut of content is not the only way Netflix is working to make its service more accessible to those with smaller screens. During Mobile World Congress, it demonstrated a "Dynamic Optimizer" that analyzes video to compress it as much as possible, while retaining the picture quality, in order to provide those with slower broadband or cellular connections a better quality of video stream.
In December last year, Netflix expanded its encoding capabilities with "chunk analysis," analyzing minutes of material to optimize the compression further, in order to minimize the amount of storage used for Netflix's offline mode.
Due to having a smaller screen compared to a television or a notebook, the framing of some shots are sometimes harder to see on smartphones, making the experience less immersive or difficult to follow visually. A second version of the same show, possibly using alternate shots or different cropping, could help alleviate this issue for mobile users, while still keeping the rest of the content's components intact, such as speech and pacing.
Chief product officer Neil Hunt told The Verge the company is looking into using a different or reframed shots for content due to the increased usage of mobile devices to view the platform's shows and films.
"It's not inconceivable that you could take a master and make a different cut for mobile," said Hunt. While Netflix has yet to offer different cuts of content to its users as yet, Hunt said "it's something we will explore over the next few years."
Hunt's remarks were made during a two-day event at Dolby Laboratories and Netflix's headquarters in preparation for the release of Iron Fist, a new Netflix show that will be available to subscribers on Friday. Also during the event, executives from both companies talked about high dynamic range (HDR) content, and how it could help improve the experience for mobile users, as well as those with compatible televisions.
A mobile-friendly cut of content is not the only way Netflix is working to make its service more accessible to those with smaller screens. During Mobile World Congress, it demonstrated a "Dynamic Optimizer" that analyzes video to compress it as much as possible, while retaining the picture quality, in order to provide those with slower broadband or cellular connections a better quality of video stream.
In December last year, Netflix expanded its encoding capabilities with "chunk analysis," analyzing minutes of material to optimize the compression further, in order to minimize the amount of storage used for Netflix's offline mode.
Comments
I've never thought while watching a Netflix show on iPhone that it was difficult to see...
Having said, that I don't want to watch a movie or TV show in portrait orientation, so I'm pretty sure this isn't what Netflix is talking about. I think they are just pointing out that like the transition from 4x3 TV screens to wide angle a generation ago, mobile devices are introducing new screen ratios.
And that’s just fucking hilarious. Some leftist woman cried about her show being garbage and now people can’t rate their content accurately in ANY respect. Honestly, there should be ratings out of 100, but we don’t even get out of 5 anymore. How long before Rotten Tomatoes stops aggregating and just gives a thumb up or thumb down?
http://www.cameradebate.com/2014/what-is-iphone-6-camera-image-sensor-size-rear-main-camera/
https://en.wikipedia.org/wiki/Image_sensor_format
https://petapixel.com/2011/11/04/why-dslrs-use-a-different-aspect-ratio-than-most-digital-cameras/
https://www.quora.com/Why-is-the-camera-lens-circular-but-the-image-sensor-is-rectangle-in-shape
Aspect ratio is 4:3 in digital cameras so in portrait the sensor is taller than wide. Forcing horizontal video just crops the data and would use about half the sensor so quite a lot of quality loss. There are apps that do this:
Square sensors would be better. It would still have to shrink the frame on an angle if it was maintaining the orientation but would be the same size in portrait as landscape.
http://theonlinephotographer.typepad.com/the_online_photographer/2012/01/why-not-square-sensors.html
http://www.earthboundlight.com/phototips/why-not-square-format-sensor.html
http://image-sensors-world.blogspot.com/2014/11/is-mobile-future-square.html
https://www.dpreview.com/forums/post/26988701
The issue is more about the framing though by coercing the users to frame the video in a way that suits widescreen aspects and cropping can be done in post-production.
Netflix having a crop for mobile likely wouldn't be about having a vertical format because our eyes are still horizontal and it will always be best to watch horizontal video, they'd mean things like zooming in some shots so the subjects in the frame were larger. Maybe they'd have a 4:3 crop for tablet displays. I wouldn't expect there to be much special consideration for phones at all as the screens are too small to watch a lot of video on. Larger TV screens are still most widely used for movies and TV viewing:
http://www.marketingcharts.com/television/are-young-people-watching-less-tv-24817/
http://www.adweek.com/digital/6-ways-netflix-mobile-habits-vary-around-world-169848/
People sign up to video services on devices but then migrate to larger screens for long-term viewing. It's best that most people view the same cut of video content where possible. The upcoming generations watch more content on PCs and mobile devices but that's probably in part due to parents watching something else on TV while they watch what they want in their bedroom and Youtube/Facebook videos are different from TV/movies. When they get their own place and own TV, they'd watch on a bigger screen. You can't have two or more people huddled round a tablet or laptop.
People like to feel that others share their opinion and the average ratings reflect that so they try to swing the average by voting to the extremes. This happened with the latest Call of Duty game where they spammed the downvotes on the trailer and filled the reviews with 0 ratings because they didn't like that it was set in space:
http://www.metacritic.com/game/pc/call-of-duty-infinite-warfare/user-reviews
http://www.metacritic.com/game/playstation-4/call-of-duty-infinite-warfare/user-reviews
http://www.metacritic.com/game/xbox-one/call-of-duty-infinite-warfare
The reviews jump between 0 and 10 ratings with people trying to pull the average one way or the other. Even people who played the game for over 200 hours gave it a negative vote on Steam. There's a point where the rating doesn't accurately reflect the user's experience and it's more about a sentiment they want to project. Sometimes rating systems are used to push an agenda like riding a wave of negative sentiment. People generally don't like to have unpopular opinions and online voting encourages people to move towards the extremes of their preferred groups and discussion topics because it's how they get the most likes.
The most accurate/honest voting is upvote alone because nobody can project an agenda or anger through it. If people don't like something, they just don't vote on it and the ratio of views to likes shows the popularity. Negative feedback is useful so on things like apps, they can leave feedback comments that get aggregated for publishers to deal with. Average mid-rated reviews are the least useful because they could be from either mid ratings or even amounts of polar opposite ratings. A binary vote eliminates the first option and they force a choice whether someone liked it or not.
The accuracy of ratings for an individual depends on whether the group leaving the ratings has on aggregate the same opinion as the person viewing the rating. It mostly works because the people who bother viewing certain content at all are going to have something in common with each other and the reactions follow that but there are times where it's completely meaningless. 1-ratings for issues that don't concern someone else make the rating useless.
The most useful setups would be ones that tailor to the individual sort of like how e-commerce sites work. App stores can have linked groups of apps e.g divide 2 million apps into groups of 50 or so and give each app a set of group ids that they relate to. Then when someone buys an app or puts an app on a wishlist, related groups can be recommended. In discovery queues, they only need to store the group id to know what apps people have seen, which is more efficient in the database.
- few million items, divide into groups of 10-50 e.g 40,000 groups
- associate each item with a few of the most related group ids
- show random mixed group to begin, people scroll through and choose the items they like and future recommendations come from the related groups
- store viewed group ids per user to avoid duplicating the recommendations. People wouldn't typically have enough time to go through more than 100 or so groups but the filtering should quickly narrow down to what people like based on their selections in the earlier groups.