I'm planning an alternative to Youtube, but aI need to know a few things first

[Dragonwinged]

As the title says, I'm planning on starting a website to compete with Youtube (as well as spotify, blogger and deviantart. All of those in one website.) I'd like to have a few features available at launch, and I'd like to know if there is pre-existing open-source code for those features.

 

1. Censor patches

 

On Youtubeb you are instantly demonetized for vulgar content. I'd rather that not be the case, but I also don't want the first thing browsers see is a lot of swearing, and since making a separate censored video for the web-page would waste a lot of space, I'd like to know if there are small size files that can be applied to content as the page loads. For text this at least sounds easy. A simple word library of naughty words, and few strings of code to auto-replace those words with illegible characters. Steam has already done this in their community forums, replacing swearing and slurs with heart emojis. (Not what I would've gone with. I think I'd make the default either a blackout text, of wingdings.)

 

For images, I believe there are already ways of layering multiple image files on top of one another. This could also be used to add things to the base image and have multiple different versions of an image simply by using layers. Not sure how it would affect file sizes though.

 

Video and audio would seem to be a bit trickier. Naturally as a platform that offers both, playing the audio version would just opt out of loading the video portions of the file. This would mean that there would only be the original audio and the censor patch loaded, which might have to be custom made for each video, as ai voice recognition just isn't that good yet. I would also allow uploaders to use either the default bleep, or custom censor sounds, which you can then monetize.

 

Video would ultimately have to compile five or more separate files to play one video: the video, the audio, the video censor, the audio censor, and the subtitles. Again, these patches would have to be pre-made for each video, and could use default graphics like blurring or black rectangles, or custom monetizable censors.

 

How the site actually decides to load these patches should be rather easy. If the user has an account, is logged in, and has censor options set to off, then no censor patch would be applied. To anyone not meeting that criteria, the patch is loaded in with the page.

 

Naturally I'd want to try and have all of this running onsite, however if that's not an option then I'll recommend optimal software for whatever hardware you're using.

 

2. Monetization

 

Naturally. Just like Youtube I'd like you to be able to monetize your content, either through advertising or through direct payment. I probably won't be using adsense. Instead, I'd like each user and advertiser to choose each other directly. You can message an advertiser and request ro advertise their product, or the advertiser will message you with an offer. This all happens either individually or blanketing across an entire video category. However, you will be prompted to confirm the deal in order for advertisements to show up on your content.

 

And you can diractly monetize through subscriptions, rental periods, or one-time purchases. And these can be applied to content individually or your entire content library. I'll also make it so that you can create group purchases, tying multiple channels' content under a single subscription or purchase. You'll just have to work it out between yourselves in terms of how much of a cut each channel gets.

 

And of courseb you can monetize custom censor sounds and images for your content, as stated in 1.

 

 

3. Private servers

 

I'm not some giant multi-billion dollar conapny, so I'll probably use peer-to-peer streaming to distribute content. However I'd like to give users the opportunity to host their own content, either with physical servers or a subscription to publically available server companies. And I'd lime to tie that in directly with the site as to just how you decide to distribute your content. And if you are paying a subscription for servers, you can set ypur channel to automatically subtract that cost from your monthly earnings.

 

4. Parent curating app

 

This will probably have to implement later on down the road. We'd probably do most things on an honor system, and getting reported for foul or vulgar content that you didn't report when you originally posted will put the content under review, hiding it from the lower rated areas of the site and probably fining you as punishment for a violation. Ultimately, I'd like to have dedicated applications for all devices for viewersb content creators, and curators who will check each content for a content rating system. As an extension of this, I'd like to have dedicated apps for parents and children so that parents can remotely curate what content their children can see on the site. Simply login both the parents account and child's account on both devices, set which is which, lockout settings using a password, and parents will be able to view content and decide whether or not they want their child seeing that content. I'll also enable parents to share curated content lists.

 

So already rhat makes my proposal very different from youtube. Do you know if this all sounds to be possible (code-wise)

[Kilrah]

Code-wise anything is possible, but business-wise it's another story.

 

Have you looked at LBRY?

[mariushm]

Well, anything is possible, actually implementing it is another thing. 

 

P2P for video.... imho we're not quite there yet. I doubt it can scale to support viral videos that do 1m views in a day, or even linus tech tip videos that do 100k+ views within 24h. 

You'll need your own download servers... if you keep it down to 720p and h264 it's a lot of bandwidth but manageable.

 

Censorship... I'm against censorship of any kind, it doesn't really work. Also, if you do it automatically, you'll start to get requests from different countries to censor automatically stuff they don't like (like China censoring anything supernatural or witchcraft or skeletons or other things which are perfectly fine in other countries,  Turkey censoring some anti muslim stuff or whatever)

Maybe add in their profile an option they can check (or can be force checked through parental controls) to censor swear words... but really, even that's not gonna work, you're just gonna force people to get inventive with the swear words like replacing letters with numbers or using unicode (  ℯ𝓍𝒶𝓂𝓅𝓁ℯ - this uses 26 bytes to write 6 characters... or 🄴🅇🄰🄼🄿🄻🄴 )

 

Monetization ... big content producers can already embed ads in their videos and have the "power" to make deals... see Linus and their sponsors.

IDEALLY, For the little guys, have maybe a pool of advertisers and their ad spots and maybe let content creators check what advertisers they're fine showing ads for. For example, if I have a channel making arduino tutorials I may have no problem accepting and showing ads for Fanta (or other drinks) or Ford and Tesla cars even though they're not related to arduino, but may not want to show eToro or crypto related ads on my videos, or ads for medicine.

Then, it's up to advertisers like Fanta or Tesla IF they want to run their ads on my videos as well... if they pay per thousands of views on channels / videos with specific keywords or themes, they may not want their ads to run on my arduino videos. 

 

edit: I'd also hope that you won't get as aggressive as Youtube interrupting videos every 6-8 minutes, nowadays mostly adding a couple of 5-15s no-skip ads every 8 minute or so in longer videos. (I suspect they're trying to push users to subscriptions by pushing too many ads) ... Or allowing ads that are more than 30s-1m long - i'm f..ing tired or being served 3-5 minute Indian music videos on Youtube.

Ideally, let users define "this is a good place to insert an ad, should you want to"  in their videos (on scene changes, where some chapter ends etc) but limit the number of ads.

Ideally, keep track across all open tabs how often the user got to see ads, and maybe serve an ad every 15-30 minutes (reward user if he doesn't skip it by resetting the timer or something like that)

 

 

What's gonna be a pain in the ass for you is gonna be

 

* royalties for encoding video, as you really can't get around to encoding to h264 or hevc... maybe in theory you could encode to vp9 which is sorta free or to av1 when it's gonna be finalized, but A LOT of viewers will have phones and computers that still only support hardware decoding of h264 or hevc 

On the audio side, you're good with Opus which is completely free, but you may have to pay royalties for AAC if you want that as backup (so that you can combine with h264 to stream mp4 to TVs and other crap). mp3 could be an option, as the patents for it expired... AAC patents should expire relatively soon, if they're not already expired.

 

* the pressure you're gonna get from music licensing groups to analyze audio, fingerprint, block videos if some music runs in background... which is kinda stupid - how would you know if that person didn't pay licensing fees for the music used in the video? 

 

 

Children stuff... careful with that... there's laws in US now about monetization and showing ads to kids, see youtube and how you have to check if the video is for kids or not. personally if I were to launch a video platform, I'd just save the hassle and require users to be 18 or older. 

I think too little of the actual content for kids is quality content - a lot of kids content on youtube is bad for kids (meme songs, product placement, advertising toys, spam videos etc) and too many "Creators" targeted this market because parents let kid with the tablet to watch cartoon videos and kid songs and kids don't know how to hit skip on ads so it was easy advertising money.

 

 

the private servers stuff ... companies already have options like that... see Vimeo - https://vimeo.com/upgrade - for example. You pay maybe up to $500-750 a year and you can post your videos and embed them with minimal/no vimeo branding, no ads, nothing. For a company, $500-750 a year is much cheaper that colocating or leasing a dedicated server, dealing with DDOS attacks, paying for bandwidth or a CDN etc etc...