I have a Ruby on Rails application where users would be uploading videos and I'm looking for a system for converting videos uploaded by the users to FLV format.
Currently we are using FFMPEG and since video conversion is a heavy task it seems to be taking a lot of time and a lot of CPU resources..
We are looking if we can use map-reduce / Hadoop framework for implementing video conversion, as it is completely distributed.
Is it a good option to use map-reduce for video conversion in real time? If it is so, how can that be implemented?
Note: Each video file size is around 50 - 60 MB.
Your requirement is "Real Time" conversion. Keep in mind that Hadoop is a "Batch Processing Framework".
IMHO, I say Hadoop is a poor choice here. A better solution would be definitely to use something like Storm:
Apache Storm is a free and open source distributed realtime computation system. Storm makes it easy to reliably process unbounded streams of data, doing for realtime processing what Hadoop did for batch processing.
Personally, I implemented a project similar to yours using Storm and the result was amazing.
Another option is to use a distributed Actors model, such as Akka.io or Erlang. But since you are a Ruby shop, Storm or Akka would be easier for your team.
Related
I've looked everywhere and I still can't figure it out. I know of two associations you can make with streams:
Wrappers for backing data stores meant as an abstraction layer between consumers and suppliers
Data becoming available with time, not all at once
SIMD stands for Single Instruction, Multiple Data; in the literature the instructions are often said to come from a stream of instructions. This corresponds to the second association.
I don't exactly understand why the Streaming in Streaming SIMD Extensions (or in Streaming Multiprocessor either), however. The instructions are coming from a stream, but can they come from anywhere else? Do we or could we have just SIMD extensions or just multiprocessors?
Tl;dr: can CPU instructions be non-streaming, i.e. not come from a stream?
SSE was introduced as an instruction set to improve performance in multimedia applications. The aim for the instruction set was to quickly stream in some data (some bit of a DVD to decode for example), process it quickly (using SIMD), and then stream the result to an output (e.g. the graphics ram). (Almost) All SSE instructions have a variant that allows it to read 16bytes from memory. The instruction set also contains instructions to control the CPU cache and HW prefetcher. It's pretty much just a marketing term.
I am looking for the best tool to achieve something like this (this is Blender's game engine, no real reflections, etc.) in an webgl viewer.
http://youtu.be/9-n12ZH5O6k
The idea is to prepare several basic scenes like this and then for the user to upload his design and have it previewed on a car (or other far more basic objects).
While p3d is nice, I don't think it does the job. There's no API for these cases yet. What are some options to pull this off? The requirement would be to have a library that doesn't have a too large footprint, since the feature/product is planned for the Asian market, so internet speed has to be considered.
you should look into three.js/babylon.js maybe? But you surely won't achieve that app just by a fingersnap, so read the tutorials as well, but these libs will surely ease your task by much.
For developing a video content heavy website like youtube which language/framework might be a better option from performance and support for video conversion/compression plugins point of view. Some points worth considering may be.
CPU vs I/O time
Support for compression/conversion plugin (existing mods/gems/libs)
Ease of learning is not very important though inputs are welcome
I know the question sounds a bit subjective however my intention is to understand the technicalities involved from someone who has had experience developing similar kind of site(s).
Unfortunately there isn't one or two APIs/Libraries/Frameworks you can knit together to produce a video serving website.
Invariably this will require heavy involvement on all levels of the stack:
Server back-end will require the following problems to be solved:
Video Encoding
FFMPEG or MPlayer experience for encoding any number of video formats to either FLV or more recent h264 for HTML5 supported formats
A reliable mechanism to transcode video in a background process; initially on one server but eventually on multiple servers as your services scales
Video resizing
Bandwidth Management to throttle connection just enough so that the video trickles down to the user
Storing video files and a file sharding and naming mechanism
API Server - Something like Rails, Django or NodeJS Express to serve as a JSON service layer between web clients and the video encoding/serving service.
Front end will require the following issues to be solved:
Playing back the video reliably across multiple OSes (Windows, OSX, Linux, Tablets, Mobile) and Platforms (IE, Chrome/Safari, Firefox, Opera) with fallback support for older browsers
DRM - are your videos free or commercial? If the latter, this is another issue that needs to be addressed
I'd strongly recommend an Event Driven system on your back-end as it is much easier to develop code that supports concurrency. NodeJS would be a good pick. It is worth looking at node-fluent-ffmpeg module for NodeJS as a good starting point.
As for your front-end I'd recommend frameworks such as Backbone.js or AngularJS to develop you web-app.
It was a fun and challenging journey when I attempted something similar a few years ago. I wish you good fortune in your journey.
For a site like that, I guess will need to choose several tools to do the job.
For the web, you could use any framework, so rails would be OK, to deal with videos you'll need something like ffmpeg or transconding to convert the videos.
For streaming, if you can use HTML5 check this question otherwise you'll need a player whith flash fallback.
Remember that the heavy part in terms of storage and CPU is video compressing/conversion.
We are planning a Wep App for a Hackathon that's happening in about 2 weeks.
The app basic functions are:
The users are guided step-by-step to upload a video, audio and image.
The image is used as a cover for the audio. Making it into a video file.
The two video files are merged thus creating a single video from the initial three files.
So, my problem is:
How do you create a video from an audio with an image as "cover".
How do you merge this two videos.
We are thinking of using Heroku for deployment. Is there a way to do it using something like stremio?
What would be the best approach? A VPS running a C++ script? How's the easiest way to do it?
FFMPEG would be a good start as seen here
https://stackoverflow.com/a/6087453/1258001
FFMPEG can be found at http://ffmpeg.org/
Also another option that maybe over kill would be Blender 3d as it could also provide simular results and could be controlled via shell commands and maybe more flexible in terms of complexe needs for asset compositions.
In any case your gonna want a server that can run heavy rendering processes wich will require a large amount of ram and cpu processing. It maybe a good choice to go with a render farm that can run gpu as the main processor for rendering as that will give you more bang for your buck but could be very difficult to set up and kept running correctly. I would also say a VPS would not be a good choice for this. In any case the type of resources your gonna need also so happen to be the most expensive in terms of web server costs. Best of luck please update with your results.
I am starting to develop an automated webcam application. The goal is to automatically take pictures, do some image processing and then upload the results to a FTP site. All of these tasks seem simple.
However, I am having a hard time to find a decent camera. I don't want to use a simple webcam or hd-webcam because the image quality of still frames isn't very good.
I'm also having a hard time finding an affordable digital camera supporting USB snapshot or control.
My second concern is the development itself. I'm not quite sure which programming language to use. I have experience with AS3, Processing, Java and some simple C++ and Open CV.
Do you have a clue?
Regarding the camera, There are pretty good webcams that you can find, some with HD quality. look at the cameras on Logitech (I tested their API and it is quite good), A HD camera has a retail of $99 which is very cheap. If you are looking for something better I would go with Nikon as they also have a pretty good API for C#/C++. You can get a basic SLR with simple 28mm lens for $500. Don't use a PowerShot as Nikon stopped supporting their API. Whatever camera you decide to buy make sure a proper API is available, is being maintained and free.
Regarding development, I would go with C#/Java as they are easier than C++. There are quite allot of libraries for image processing for C#/Java, just make sure that the Camera comes with an API the fits your chosen language.
Good luck.
Generally (from experience) most USB cameras that show up as an imaging device through Windows can be used with JAI [Java Advanced Imaging]. Additionally [on the .net/c++ side], the same cameras can be used through DirectShow as a capture device. Java/C# will make development easier but expect to loose some performance [even with the best of optimizations]. Additionally you can only perform upto the speed of the camera and the data line running from the camera to the computer [USB1.0 will seriously limit a decent framerate]
first get the image in RAM:
If you are using CHDK, I suggest you get the image copied from camera memory to RAM by using supported scripting languages by CHDK - you can take help from the CHDK forum http://chdk.setepontos.com/index.php for this.
or if thats difficult you can continuously copy the image to hard disk and load in RAM from there. (you need to take care (delete) of massive images accumulated on hard disk in a short period of time !)
This sounds like a 'brute force' approach, but will get your work going while you are researching correct approach.
perform image processing:
once the image is in RAM, you can apply your image processing algorithms as usual e.g. using opencv library.
hope this helps you