[Show NC] Voola - Stream live or recorded videos, watch hyperlapse replays



  • Hi,

    I am the creator of this Streaming app called Voola. Its currently only for Android 4.3+. Downloadable here - https://play.google.com/store/apps/details?id=in.voola.android

    It allows to stream live or record and stream recorded videos. Once streaming is over, the streams are hyperlapsed and available in feed. We also have a blur screen, which blurs the screen while streaming and the viewers can only see a blurred screen and the streamer's identity is hidden. There are other filters which we have currently taken out. We will bring them in in our V1 release.

    Initially it started as a pure live streaming app. Something like Instagram for Live streams. We even had Square videos and only live streaming.
    But as we worked with more users, we realized that especially in India, relying on a stable connection for streaming a moment is too much to ask. So if bad internet connection, then you either lose the moment or switch to some other app to capture the video. Hence we are coming up with offline support, ability to record and later stream recorded videos. Its due in 2 weeks from now.
    We also couldnt let go the fact that Vertical videos might be the future. So we revamped our entire UX/UI for supporting easy consumption of Vertical videos.

    We believe Streaming videos is a whole lot different experience than uploading videos. Its lot more interactive, where just like live streaming, users can join, chat, share unlimited loves.

    There are other things coming up, but above is the crux of the product. We completed 1 month at Playstore yesterday and because we are fixing bugs (a whole lot of them) we haven't really told many about this app. But from past two updates, the stability has increased tremendously.

    So what do you think?



  • Congrats on completing one month! How many users you currently have?

    Did you know notCRUD has cool signatures? Use them by going to your User Profile > Edit.

    Personal blog: paraschopra.com/blog/

    1


  • @paraschopra We are few short of 200. I believe it will grow at this rate till 15th. Post that we will have our public beta (android 4.1+ support, recording and streaming recorded videos) and then we can concentrate of letting more people know about us. I hope this is not a bad strategy. I also believe IOS might hold our key to a faster growth.

    We had to hold our urge to publish everywhere intially. Now reflecting back, we think we did a good job becase there were quite a few bugs/app crashes initially. Now that things are really smoothening out, we are getting more confident of letting others know. Another thing that I am proud off is that we have personally told only 5 people to install the app. Rest everyone has come from referral from our users.

    Actually this notCrud post is our second post on a public forum in last 1 month.



  • Good luck! Seems you have some very good ratings. Though I wonder how do you plan to compete with already established Periscope?

    Did you know notCRUD has cool signatures? Use them by going to your User Profile > Edit.

    Personal blog: paraschopra.com/blog/

    1


  • @paraschopra Thanks for the wishes. Yes its difficult but not impossible. But even periscope came after LiveStream/Ustream who had more than 10 million users.

    Coming out of the philosophy, we are actually short form streaming. We intentionally restrict to 4 mins of streaming. So as an analogy if Periscope is an Event, we are "behind the scenes" and "tidbits" of that event. Our mission is also about streaming moments, be it live or recorded. So basically its an app we are trying to build that users will always open when they need to capture anything video. If there's internet and user wants to stream, they stream, else they record to stream later.



  • @gouravd App looks very slick, I dont use android phone so couldnt give a try myself. Does app connect live stream to Youtube or its standalone? I also noticed that you are only support ARM processors devices, any specific reason behind it ?

    BTW, congratulation on app, its really well done.



  • @shardul Thanks for the encouraging words. Now I also know whom to reach out when our IOS version is ready :)

    No, our app does not connect the live stream to Youtube. We maintain our own backend (and control) with servers, videos, hyperlapse, facial processing systems. There are few reasons why we use our own infrastructure

    1. We want to manage/control the entire pipeline from Video Creation to Consumption. This is not because we are control freak :), but because of our long term plans and mission.
    2. We post process a lot. One of them is generation of hyperlapse replays which is not possible through Youtube.
    3. For using Youtube backend, the user has to go through a very tedious and frustrating multi-step login/verification process which we cannot allow at Voola

    In the latest version we support ARMEABI, ARMEABI-V7A, ARMEABI-V8 64. I personally do not have test devices to test on architectures like MIPS, X86. We had initially support for these (around 0.8.0) but there were no installs in those devices but the file size was large due to extra .so files for those architectures. So we decided to drop those for the time being as I do not see much value at this point.



  • @gouravd Can you elaborate on the technology stack of Voola?
    I also attempted a video streaming platform before. Following link helped me a lot, in coming up with an architecture for my application. Hope you find it useful too.
    http://highscalability.com/blog/2010/3/16/justintvs-live-video-broadcasting-architecture.html.



  • Hello Gaurav!
    Must say I love the user interface! I feel like you've got the colours spot on, and the user flow is great as well. However when I replayed the live stream I was broadcasting , I found the video/audio to be very poor in quality, as it was occasionally getting stuck. Having said this I have a few questions to ask regarding your product and how it plans to fit into the market.
    Live-streaming is highly dependent on a stable internet connection and you seem to agree with it as well. But even if that problem is solved, I believe consumer behaviour is something that is going to be very difficult to tackle. Right now I feel live streaming is a glamourised feature( in the mobile space), and has the inability to get someone hooked. What will make people want to come back for more?
    Currently, if I want to know what's happening with my friends around I could use snap chat ( 10s videos/picture) with stories lasting upto 24 hours before self destructing. This is also has the added advantage people being selective in what they upload doesn't it? If I record something and it sucks, I can always choose not to upload it. Now what if as a viewer I watch these "tidbits" for 4 whole minutes and find nothing of interest? People are more impatient than ever and this can be seen with other applications as well with Instagram videos, Vine, 9Gag videos. We just love to scroll through them in our free time.
    Where would you say Voola can fit into the market place, or rather where do you see it fitting in the near future? Who would you say are your competitors?
    Eager to hear your response :)



  • @p_chhonker

    1. Android 4.3+'s exposed MediaCodec is used to redirect the camera input to PreviewSurface. Audio running in separate thread is then muxed using FFMpeg that is interfaced via JNI.
    2. We then send those muxed streams to our Media servers
    3. Our backend is Nginx webserver on Linux combined with customized Nginx rtmp media server. We also use a very less known opensource media server platform called SRS as failback.
    4. After a stream is over, http post is used to request Post processing on another server, powered by our Hyperlapse algorightm and runs on ASP.NET 8
    5. Data is stored on Azure/Amazon and server and served over CDN.

    For mobile backend we are using Parse, but slowly moving to

    1. Neo4j for GraphDB
    2. SQL Server for User details
    3. Elasticsearch for meta and search
    4. Redis for caching.


  • @rahul047
    Firstly sorry. I know it must have been a terrible experience. Let me answer each question of yours one by one.

    ***"Must say I love the user interface! I feel like you've got the colours spot on, and the user flow is great as well." ***
    => Thanks a lot for that

    "However when I replayed the live stream I was broadcasting , I found the video/audio to be very poor in quality, as it was occasionally getting stuck"
    => Previously we built our SDK from scratch. The architecture was pretty good and allowed "Adaptive Bitrate". But it missed one big thing. "Network State". There was no way to know what was the state of the network and streaming end point. So even if the network was down the SDK continued streaming to a non-existent end point. So we re-wrote the sdk from scratch and mainly focussing on the "Network state" part. But this new architecture heavily was tied to FFMpeg for packaging of the video/audio as well. Because of that if we switched Bitrate in between the streaming had to stop and restart. This was not conducive for our backend.
    In the meanwhile we also notched the resolution to 480x864 at 512kbps @30 fps. So you can imagine, now we have a situation where the data from Camera and our sdk is being fed to a buffer in phone and the buffer is emptied by our SDK to our streaming endpoint. Because so much data is being sent and the bitrate is not getting adjusted, the buffer is getting full very fast and new frames are getting dropped as our sdk is not able to empty the buffers in time. Hence this issue. As you might have imagined we are currently working on bringing back our "Adaptive Bitrate" method that auto adjusts the quailty at realtime. The quality is pretty good (if you look at some our past replays) and video does not gets stuck.

    "Live-streaming is highly dependent on a stable internet connection and ....."
    Live streaming on mobile even in its 4th year, is still a very nascent stage. Especially for the publisher/producer side. There is no doubt that there is enough audience for consuming them. But the technology barrier is restricting large scale adoption from a creator side. But in last year or so, Internet has increased tremendously and so has number of users who live stream. Even in our app we have a good percentage of people who open the app to view streams/replays.

    You also pointed out the pshycological part that might restrict users to go live. We are well aware of it. We have a blur feature that blurs the screen and hides the streamer's identity. But to really solve the creation problem we will have to target specific areas of life and focus on making streaming those. We are currently stil trying to get the platform ready (read bug-free) so that we can then approach the relevant publishers or creators. There are many use cases like live cookery class, live exercise classes and many more. We are starting with two such specific areas later this month (hopefully).

    Also because of this content creation problem, we are launching our Video streaming feature shortly. Video streaming is just like live streaming but not from live camera. It can be a recorded video or existing video from your phone. The idea is to get many people get together simultaneously to view, chat while watching/streaming recorded videos. People anyways upload videos, we want them to stream. This is much more interactive experience than uploading a video. We call it "Interactive sharing". We also forecast that most of our users in the next 1 year will use Video streaming more than live streaming. But in due course they will get comfortable with videos that are not edited and they will live stream more and more. We are providing a smooth landing surface. From Videos Uploading Era -> to Video Streaming Era -> to live streaming era

    "Now what if as a viewer I watch these "tidbits" for 4 whole minutes and find nothing of interest?"
    Actually we generate hyperlapse replays of 20s. So actually you do not watch the 4 mins video unless you want. In your case, our hyperlapse algorightm failed because the video was malformed. So the hyperlapse was "not" generated and you simply got the full replay.. But all videos would have <=20s hyperlapse replay

    "Who would you say are your competitors?"
    In pure live streaming, I remember around 25 live streaming apps. These are of these genres

    1. Theoritically limitless streaming and replay available for certain time. (Like Periscope, MyEye)
    2. Theoritically limitless streaming and no replay (purposefully) (Like Meerkat, Outcaster)
    3. Combination of 1 & 2 but targeting certain backends (like Facebook, Youtube etc..)

    And we are :
    Periscope married Hyperlapse and gave birth to a baby like Vine, but taller (~20 secs)



  • This post is deleted!

    ga___jit

    0


  • @gouravd In streaming business, bandwidth costs can make or break a startup. I ended up choosing OVH's 'VPS SSD 1' plan for our application. We had to do some architectural changes to use this particular plan. Check it out at http://www.ovh.com/us/vps/vps-ssd.xml. In our case OVH's bandwidth turned out be 100X cheaper in some cases to the services like amazon etc. They delivered on what they promised on their website. However, we only tried them for a few months, so your mileage may vary. I suggest you try their offering and figure out for yourself.
    Would love to see you succeed. Cheers.



  • This post is deleted!

    ga___jit

    0


  • @p_chhonker we are currently using 2 octacore 2.8 Ghz , 32gb,1.6gbps machines . I couldnt find any machines near those configuration in OVH. The machines at OVH and other similar VPS providing sites are way too slow for our needs.



  • @khitchdee I beg to disagree about the point-point architecture. There are certain use cases which can leverage that(like live video chat) but in general when you want thousand s or even millions to view a stream it has to be server/client broadcast. In point to point you will mostly use up all the bandwidth of the client pretty quickly.


Log in to reply