Blocking members who records and uploads on other websites

[Shane]

I'm not a model. I'm a helper to a model.

The model only does nude shows on group and privates except flashes in public chat.

Some members record the show and upload it to the other sites without the model's permission. And she don't likes it.

Is it possible to display the names of the members on the webcam screen as a base62 string for groups and privates?

If the recorded video is posted somewhere, it can be identified who were on the show, if there is like only one person on the group
show then he might be the one.

 

[blayze_noir]

I'm not a model. I'm a helper to a model.

The model only does nude shows on group and privates except flashes in public chat.

Some members record the show and upload it to the other sites without the model's permission. And she don't likes it.

Is it possible to display the names of the members on the webcam screen as a base62 string for groups and privates?

If the recorded video is posted somewhere, it can be identified who were on the show, if there is like only one person on the group
show then he might be the one.

Nope. That is a feature request to make for each of the cam sites

 

[Shane]

Nope. That is a feature request to make for each of the cam sites

If implemented in large scale, everyone will know what the code on the corner of the screen means and the uploader might just hide/crop/cut it.

 

[blayze_noir]

If implemented in large scale, everyone will know what the code on the corner of the screen means and the uploader might just hide/crop/cut it.

Or record while not logged in. But, a little deterrent goes a long way. It's amazing how even a little inconvenience makes people stop doing something.

 

[maskedsissyslutforyou]

Providing a slightly different video stream to each and every webcam viewer would require a lot of re-encoding and is probably not viable in the stream itself.

It would really need to be done on the browser end so that it is a single video stream with adjustments made on the browser end.

If implemented in large scale, everyone will know what the code on the corner of the screen means and the uploader might just hide/crop/cut it.

Generally when someone does this sort of watermark it just adjusts pixels in a less than obvious way on the screen and you need software to see/decode it. Ie if you had the original stream and differenced the same frame then the "difference/error" would be the watermark. I believe each and every hollywood screener has been doing this for a few years so if it gets released they know which copy it was and they know who to blame or sue or give a copy to ever again.

 

[ericlafayet]

Their server could do transcode the stream adding the watermark in random positions for every model. So the watermark is not always.

Doing it per user instead of per model doesn't makes sense. Lot of resources to process the transcode.

But, my opinion is that the watermark should be added on the source, that's on each model's computer. Less work for SC servers, less costs for the company.

My 2 cents.

 

[maskedsissyslutforyou]

Doing it per model means you do not know which user did it, which is not really useful for much of anything but knowing it came from this or that streamer (which pretty much is already probably already known).

There should be easy ways to watermark a compressed stream without doing a full decode/re-encode with a decent understanding of the underlying encoding algorithm.

 

[ericlafayet]

There should be easy ways to watermark a compressed stream without doing a full decode/re-encode with a decent understanding of the underlying encoding algorithm.

As far as I know, nope. You need to transcode the rtmp stream (likely with ffmpeg) to be able to modify that stream.

 

[cbhours]

I'm not a model. I'm a helper to a model.

The model only does nude shows on group and privates except flashes in public chat.

Some members record the show and upload it to the other sites without the model's permission. And she don't likes it.

Is it possible to display the names of the members on the webcam screen as a base62 string for groups and privates?

If the recorded video is posted somewhere, it can be identified who were on the show, if there is like only one person on the group
show then he might be the one.

could put a very faint date/time stamp on the stream using OBS and then enable it during those shows.

 

[ericlafayet]

could put a very faint date/time stamp on the stream using OBS and then enable it during those shows.

Exactly, watermark, model name, date/time, ... but this will just show the same for all users.

@maskedsissyslutforyou's point is to send unique streams per user. Doing it server side is something that doesn't make sense.

But maybe browsers could do that. https://blog.scottlogic.com/2020/11/23/ffmpeg-webassembly.html

The thing is: should the camsites do an overuse of the user's CPU ? I wouldn't do that.

My 2 cents.

 

[blayze_noir]

I don't know about the other sites, but Chaturbate overlays their logo in the top right corner of the video. It would not be difficult to overlay the user name as a watermark. Again, this only works if the user is logged in. Which, in this case, you would need to be logged in in order to spy on a private. Yes, someone could blur the user name using a video editing app, but that takes time which effort that could stop a great deal of recording.

 

[Shane]

So, my idea doesn't work?

It looks like around 3 guys, one in particular joining groups with the model I know and uploading the show everyday to some other site.

 

[ericlafayet]

I don't know about the other sites, but Chaturbate overlays their logo in the top right corner of the video. It would not be difficult to overlay the user name as a watermark. Again, this only works if the user is logged in. Which, in this case, you would need to be logged in in order to spy on a private. Yes, someone could blur the user name using a video editing app, but that takes time which effort that could stop a great deal of recording.

I'm afraid it doesn't work like that.

The stream goes from OBS/web via RTMP to site's servers. They receive the incoming RTMP stream, add a layer with an image (or whatever) trnascode, and publish the outcoming RTMP ready to be fetched by web clients.

Servers manage 1 stream per model.

If you want that each user receives a stream its unique stream with its name, sites will have to do the same process 1 per user. If a room have 400 users, this means 400 outcoming RTMPs. And that means lots of resources (CPU basically), and that means lots of money.

What I posted in my previous message was about do process/Transcode the rtmp stream at client-side. But I don't think that's a good idea, it will use lots of resources on user's computers.

AFAIK there's no way to do that.

 

[Shane]

could put a very faint date/time stamp on the stream using OBS and then enable it during those shows.

Yeah, this will help! https://obsproject.com/forum/resources/time-and-date-overlay-updated.1461/

 

[ericlafayet]

Yeah, this will help! https://obsproject.com/forum/resources/time-and-date-overlay-updated.1461/

Yep, that's the only option I see to correlate captures with users.

 

[maskedsissyslutforyou]

If the uploader is capturing the RTMP stream directly (I know there used to be software to login and do exactly that--not sure if that still works or not) then the browser adding a watermark would not help.

If the uploader is recording the screen then the browser adding the mark would still work.

I think the key is to put the watermark/info on the video such that it is obviously visible to anyone watching the video.

I am betting stream from model -> chaturbate encoder location-> encode to several resolutions/bandwidths -> distribute the set of streams to each distribution node and then there the correct stream to each viewer.

I would not think (cost-wise) the distribution nodes would have hardware encoders/decoders to for the full stream, so any scheme used would need to rely on simple modification of the x264 stream that does not rely on a full decode/encode.

You might also be able to hide it in the audio and that would be fairly easy to do on distribution nodes. But audio is not always recorded, and having it in a video frame would show on the screen caps if done large enough.

 

[NotYou]

The fact that NOWHERE does this must tell you that it can't be done or not so easily that it's "worth it".

 

[ericlafayet]

I would not think (cost-wise) the distribution nodes would have hardware encoders/decoders to for the full stream, so any scheme used would need to rely on simple modification of the x264 stream that does not rely on a full decode/encode.

I don't know what you mean with hardware encoders, but definetll no. More likely just linux servers with ffmpeg doing the transcode. Whatever you want to do to the stream, it has to happend during the transcode, so yes, relays totally on it.

If you want to generate one stream per user, for 100 users will need 100 ffmpeg doing the job. There's no other way to do that. If you know how to, please, tell me how yo udo that magic.

 

[ericlafayet]

The fact that NOWHERE does this must tell you that it can't be done or not so easily that it's "worth it".

Taking into count that the this can be done on OBS, sites do not need to spend time and money on that.

But, it was nice to have those ideas and discuss. thanks all

 

[maskedsissyslutforyou]

I don't know what you mean with hardware encoders, but definetll no. More likely just linux servers with ffmpeg doing the transcode. Whatever you want to do to the stream, it has to happend during the transcode, so yes, relays totally on it.

If you want to generate one stream per user, for 100 users will need 100 ffmpeg doing the job. There's no other way to do that. If you know how to, please, tell me how yo udo that magic.

No. They aren't going to be using ffmpeg with simple linux servers without hardware encoders.

I can put a huge amount of money and power in a big multi-core cpu and will be able to only transcode a few x264 1080i streams at a time, or I can buy (for example) a much cheaper machine with a nvidia card that most recent models have 1 or 2 hardware encoders. For around $150 the right card will easily do around 10 1080i streams real time at a lot less power and cost. My 6 core ryzen will do roughly real time (ffmpeg/software only) with decent quality on a 1080i stream, my several generations old nvidia card (in a old slow machine) will do 2.3x on the same stream. My ancient nvidia card is rated at 8x1080i with less than good quality, so that $60 card is 2.3x faster than my general purpose ryzen that cost way more. Note ffmpeg supports using nvidia and intel hardware encoders. Read about "nvenc" on wikipedia for info about them. Note a $150 Tesla family nvidia card is will do 10 1080p streams at good quality, ryzen cores needed to do that are around 50-60 cores (and that 60 core chip alone is $8kUS). And with several 8x PCIe slots I can put several of those $150 Tesla cards in and really my limit is then getting the data in and out of the machine and cards. And the various cloud providers do provide VM with access to GPUs for various tasks that GPUs are much better at handling. And I bet that dedicated hardware encoders probably exist for all of the video being distributed.

 

[ericlafayet]

AWS with EC2 instances on autoscaling with ffmpeg on containers, and lot of development.

I was working on a streaming platform (not modeling) and we were running that. It's not about the cost of 1 resource, but the cost of scalability depending on the traffic. A business will reduce the cost as much as possible, as a Sysadmin, I t's part of my job to evaluate what and how to do in order to reduce cost.

On platofms with hundreds of thousands of concurrent connections, scalability is key on the business performance.

But anyway, I'm just explaining based on my experinece on those tecnologies. That's al.