In my last story I explained some of the issues of working with ‘free’ or near free software. This time I’m looking at video platforms.
Regular readers will know I am a video fan. I may not be the best film maker and I know I have a great voice for radio but it is a medium I find fascinating and satisfying. It forces you to compress ideas so that the essence is delivered, often in less than 5 minutes. It creates unique challenges that often require really fast but precise thinking. It is also one of the fastest growing mediums on the Internet. Everyone wants video. But…choosing a content delivery network (CDN) is a much more difficult challenge.
We knew that YouTube was a non starter. It may have a massive distribution system and we may still use it for viral distribution but for a website that is dedicated to video? Nah. Mostly. Video gets degraded when YouTube does its conversion and that sometimes means otherwise great content is almost unwatchable. The adverts are a deal breaker for what will be a professional site. As I said before, free has a price. The upside is that YouTube works well on many platforms and does not exhibit synch issues I come on to later.
Next we looked at Vimeo. I mostly like it but its colour palette is wrong for me and I wasn’t convinced I have enough control over the content for it to make the cut. Also, I noticed variable degradation in video quality which implies it is unstable. And then I discovered there are many known issues for mobile platforms that are languishing in their support forums. That ruled them out even though a plus account is only $60 a year.
Next on the list were Viddler and Brightcove. Both provide players that can be used on most platforms. That’s a big plus because it means coding for iPhone etc is kept to a minimum. Both provide extensive control over appearance and branding. Again, a plus because it removes video production considerations that would otherwise mean the inclusion of several steps. Both include distribution control. That means we get to say where our content goes rather than the viral method YouTube prefers in support of its advertising led business model. Viddler and Brightcove have a business model with which we can live but are very different. I’ll get to that later.
Xero uses Viddler to very good effect. This interview with Jason Blumer shows what can be achieved with some creative forethought, solid editing and a useful platform for HD delivery. However, we quickly found that when it transcodes, you can end up with audio synchronisation issues. A number of tests revealed this is a known and persistent issue going back THREE years. Check the forums. There is a workaround which requires the use of keyframes. That’s not a big deal and is good practice. What is a big deal though is that you have to be a premium user to discover some issues. That is because you have to pay in order to get access to additional video players for testing on mobile platforms. Support is not great. They are clearly swamped with questions and some of the answers are not exactly useful. However, it does have an excellent way to assess cost. It uses a combination of average video length plus anticipated number of views to establish your likely cost. I like that because it provides predictability plus a fair assessment of your likely commitment as you grow. Finally, it has a clean interface that is pleasing to the eye and easy to understand.
Brightcove also has synchronisation issues which are as yet unresolved but I suspect are tied up with the keyframe issue. They provide free access to most of the system on a 30 day trial but they don’t allow you to see all the players they have on offer. The players themself are marginally better than those offered by Viddler but that’s not enough to seal the deal. The interface was written by an SAP GUI developer. In other words it’s clunky and non-intuitive but with a bit of practice you can get used to it. Support is first rate. They get back to you quickly and follow up to ensure you are OK. The support offered is comprehensive. New users always make silly mistakes yet Brightcove ensures you get a readily understandable answer. I’m less happy with the business model which is capped for usage at different levels. I also noticed that there are some features I want that would require me to step up to a higher subscription level than I can readily justify at this stage.
So what to do?
I am going with Viddler for the time being. It will allow us to see how usage works out, assess peaks and troughs, get a handle on traffic and so let us have a means to easily predict costs. The niggling feature issues will get ironed out as we complete the test phase. I can live with the slower support because I know that most of our issues will arise as we go through testing and if it takes 12-18 hours to get a response then it is not a killer. We haven’t ruled Brightcove out completely but are keeping our options open.
Some will ask why we don’t go for self hosting on Amazon. We could do that and use open source components but that would mean sacrificing the analytics we need. We would likely save cost but then Amazon is not a dedicated CDN and that IS important because CDNs optimize their networks for rich media content delivery. Or at least they should do.
If you are interested in doing your own research on this topic then I suggest starting with this search.
- There is no substitute for research but you will find there are issues you’d not thought about which may trip you up.
- You get what you pay for. It sounds trite in a world where the cost of acquiring services is on a death march towards free but will you sacrifice for the inevitable need for serving Google Ads and the almost near certainty that service will be lousy? Many will and their sites will suffer accordingly.
- Video is an emerging trend that presents special problems for mobile platforms and devices. Make sure you understand the limitations of each.
Update: Viddler got back to me (nice touch) and advised the run 12 hour live service. Bearing in mind they’re based in the US there would be a time delay from my location, especially if I was wrangling the app during ‘normal’ GMT/CET times. I also mentioned something about an iPhone add-on as a cost plus item. That was factually incorrect and I have removed the offending sentence. Apologies to Viddler.