Screenly Q&A: Optimization Crucial For Raspberry Pi Digital Signage

screens_add

I caught up by email this weekend with Viktor Petersson, the guy behind Screenly,which is arguably the most commercial digital signage management platform on the market that’s been built specifically around Raspberry Pi micro PCs.

Certainly, there are other options, but in some instances Pi devices are supposed along with others, or the efforts are open-sourced, or in the case of the UK’s Silver Curve, it’s not a platform but a graphics engine that supports other CMS platforms.

It’s been about a year since I first wrote about the San Francisco-based company, which also has people in Sweden, and I wanted to get a sense of where things are at.

Q – There is a marketplace perception that using Raspberry Pis for digital signage is just about cost, and the user base is all academia and geeks. I gather that’s not what you are seeing. Can you describe your range of customers?

A – With Screenly Pro, we have customers ranging from small shops to retail chains and international companies. While we do also have a fair amount of customers in academia, it’s not a majority of our customers.

With Screenly OSE (Open Source Edition) on the other hand, the users do tend to be more technical. Even so, we see plenty of posts on the Raspberry Pi forum from users who are unlikely to consider themselves geeks.

The use cases for Screenly range from the traditional digital signage applications (public and in-store ads), to office deployments to display Business Intelligence (BI) or other performance metrics.

For us, the choice of using the Raspberry Pi was more about how easy it is to work with than cost. Before the Raspberry Pi, you were more or less forced to buy a large number of units (or very expensive SDK) to even get
started. The Raspberry Pi gave us everything we needed to build a digital signage player, without having to design our own hardware.

Q – Everybody sees the $45 micro-PC reference to Raspberry Pi and assumes that’s the hardware cost. To properly kit out a Pi with an enclosure, memory, power, what are you more realistically looking at as the capital cost for a box suitable to use in digital signage applications?

A – Right, the Raspberry Pi (Model B) costs $35, but as you correctly point out, you need additional equipment, such as a case, SD card (Class 10), power and cabling. Normally, you’d be looking $60 to $70 for the actual required hardware.

screens

Q – You did a presentation recently for a meet-up in San Francisco, and talked about what you and your developers have learned in the past year. Can you walk through some of those findings?

A – Certainly. We’ve learned a great deal while developing Screenly.

One thing that may seem obvious to hardware hackers, but is easy to forget if you come from a modern software background, is that the Raspberry Pi has very limited resources compared to your desktop computer. A simple operation that takes less than a second to run on your laptop when you write the code may take a 30 seconds or a minute to run on the Raspberry Pi. Hence, optimization is crucial.

Another thing that have caused us a great amount of headache is related to how unreliable SD cards are (or at least consumer-grade ones). While we do require Class 10 cards for speed (it makes a huge difference), we’ve founds that the lifespan of SD cards (in particular for certain vendors) is very short. We’ve worked diligently to minimize the amount of activity (in particular writes) to the SD cards, which has improved the lifespan significantly.

While I probably could do an entire presentation about this alone, the final thing I want to cover is overscan. While it’s somewhat of a Raspberry Pi-specific issue, we’ve learned a lot about this.

In short, it is very hard to automatically probe for auto-scan settings. While most modern TVs are intelligent enough to work out-of-the-box, many older (and cheaper) monitors might require a bit of tweaking. To make matters worse, the HDMI identification string, which is more or less the only data point you have to work with when trying to automate this, is often time shared among a fleet of different models, but they still have different overscan settings.

Q – How are you pricing Screenly and how do you position the product? Entry-level for small business and higher ed? Or what?

A – The way we try to position Screenly’s pricing is to make it reasonable for all types of business. We give you the first screen for free, and then we scale up with a volume discount to incentivize large installations.

Q – Products evolve and undoubtedly both the Pi and your product will evolve. How do you see things evolving over the next 12-24 months?

A – We have some very exciting things to announce later this year, but I can’t talk about that yet. We prefer to announce new features when they are ready for our customers to use.

What I can talk about is our over-all goal. We want to make digital signage accessible for the vast majority of the businesses out there. Regular businesses do not want to have to deal with overly complex software that requires consultants to even get started with. They want a turn-key solution that just works. We want to be the ones providing that.

As for how the Raspberry Pi will evolve, we’re watching the development of open source drivers for the Broadcom VideoCore IV chipset in the Pi very closely. With better low level control of the hardware, it should be possible to further improve performance and really take the hardware to its maximum potential.

Q – Pi gets all the attention, but it’s not the only low-cost ARM micro PC out there. Are you looking at anything else?

A – We keep our eyes open. The landscape has changed significantly since the Raspberry Pi first came out.

You now have a plethora of Android-based micro systems, as well as other Linux-based ones. We are getting requests from customers to expand Screenly to support other platforms too. Down the road, we might do so but for the time being, we’re happy with the Raspberry Pi and the community around it.

At the end of the day, we want to deliver the best digital signage platform to our customers, and what our customers love is how easy it is to get up and running and using Screenly.

Q – Are you satisfied where your company is at with Screenly, in terms of business growth, at this point?

A – Absolutely. When I wrote the first prototype of Screenly as a fun side-project, I had no idea that it would turn into something this big.

1 thought on “Screenly Q&A: Optimization Crucial For Raspberry Pi Digital Signage”

  1. Hi Dave,

    Thanks for the coverage.

    Just to clarify: I wrote the proof-of-concept of Screenly. I have since handed over the development to our developers and my co-founder (Alexander Ljungberg), who are all infinitely better developers than I am. 🙂

Comments are closed.