Surveillance Startup Entropix Wins Silicon Valley Competition

By John Honovich, Published Jun 02, 2016, 10:52am EDT

Industry people often complain about the lack of high tech startups over the past decade. With more and more former startups / independents (e.g., Exacq, Aimetis, Brivo, VideoInsight, Milestone, etc.) being sold off, there are fewer new ones rising to replace them.

Now, a new startup lead by former Arecont leaders, Entropix, is gaining ground in Silicon Valley.

In this note, we examine their recent progress, their upside and risks.

Competition ***

*********** *** *** **** Conference **** ******* *** ** ***********. SVOD, ** ******* ****** Open *****, ** ** organization / ***** **** showcases ******** ** ******** California. **** ** ********* with ******* *********.

**** ** *** ***** of ********'* ******* *****:

******** ** ******* ***** ****** A ***** *** **** win **** ********* **** them.

Upside ********

*** ****** ** **** there ** **** ********** here. ** ** *** another **-******* ****** ** hosted ***** ******** ** kitting *** *** ***** components **** * ******** / *********** ******.

**** ******** ** ******** is * ******** **** *********** magnifies ***** ********** ****** into **** ****** **** (e.g., ***** ** ***** becomes ** **** *****), using * *********** ** dual ******* ** ******* and ***** ********** / 'enhancement'. *** ******* ******** is *** ****** *** third ***** **** *** video *********** (*.*., *** ******* ***** *********** ************* *********** ******* ******** With *********).

**** ** * **** technology / ************ ******** that ***** ** **** widely.

Risks ********

*** *** **** ***** we *** *** ******** are:

  • ********** ********* ** **** to *** ***** **** increasingly **** ***** *** satisfied **** **** **** have. **** *** ******* at ~$*** ******** *********, at **** ***** ***** may ** ****** **** to ******* *******. ** the ********, ******** ****** that ***** *** ******** limits (** ******* *** imagers) ** ******** ** move ********** ** ******* their *********. *** *** one *** ** **** traditional ********* *** *** much **** ***** *** market ** * *** risk / ********** ** Entropix.
  • ******** ***** *** "******** *******". Even ** ** *****, like ********, ** ******** it ***** ******** / incorporated **** ******** ***** surveillance **********. *** **** challenging **** ********, ******** needs *** ******* *** integration **** ***** *** the *****. **** **** surely ****** ******** *** make ****** ********.

**** ********** *********** ** *** original ******** ****.

Broader *** ************ *******

******* **** ******* ** fail ** ***** ************, they ***** **** ***********, probably ****** ************* ******* of *** ************ *****. For *******, **** *** pitching ***** ********** *** incorporation **** ***********. ** could **** **** ***** for * ***** ******** to ******* *** ****** ****** hardware and *** ******* ********** software ******** ** **** consumers, ** ******, ******* pictures ** ****** ** friends *** ******.

** **** ** *********** to *** *** *** where ******** ***** **, and if **** *** ******** those ********* ****** *** video ************ ******.

Comments (12)

This is seriously impressive! There's no question that this technology is on the right track. GPU's are getting insanely powerful, and it's time that someone did something more than just affine transforms as the 'cool' add-on effects for video and pictures.

Cameras can't get much better on their own. The industry needs better tools, and better solutions to drive innovation. It seems that this technology can take hardware that has been commoditized to death, and breath new life into something that can create high margin revenue.

VMS companies are facing the same commoditization that us hardware folks have been feeling for years. Cameras are "pretty good" but really not THAT great when it comes to resolution - especially with the decline of really good lenses.

Everyone has video or pictures that they wished they could zoom in just a wee bit more.

I have a feeling it's going to be something you simply can't "un-see". Like the first time you see a retina display, or a truly good monitor... Everything else is going to suck in comparison.

Smaller companies that have traditionally found it impossible to break into the consumer markets (UAVs and other 'sports' cameras) would be able to wrap solutions around this type of engine, and bring highly differentiated (i.e. valuable) solutions to market.

I'm intrigued to see what effect this type of solution has on color fidelity, dynamic range, and other challenges that 'conventional' camera solutions have been plagued with.

Ian, good feedback. As someone who manufacturers cameras, how concerned are you about the dual imager / lens requirement? I mean in terms of device complexity, size and cost.

I'm personally fine with it, and would go as far to say that it's one of the more appealing features of it.

The complexity is a double edged sword for Entropix as it will limit adoption by all the 'massive' ODM players in Taiwan, etc, but it also creates opportunity for others.

ISD has been working with multiple imagers for many years now, and the chipsets to do this are getting more mainstream.

The timing is actually perfect for this type of thing to emerge. By Q2 2017, a conventional SOC will be available that can be used without having to fall back to expensive and hot FPGAs. Any earlier and it would have been a cool parlour trick, but would have been very difficult to commercialize.

The small furry rodents of the industry (like ISD) can sprint ahead with interesting products to market, while the larger dinosaurs are still hell bent on chasing a $29 1080P camera.

It's also helpful for Entropix, that Silicon Valley is engaged and noticing this technology. Lytro and other 'non conventional' cameras are highly interesting to the larger investors, which leads me to believe that Entropix will have no problem finding investment dollars (the usual bane of interesting startups).

The limitation is for that entire category, is exactly as you described however. All those products have been stuck in the 'science fair' camp, usually due to some serious hardware funkyness.

VCs love easy commercialization, and solutions that can be deployed to large markets. They hate risk, and love to create the 'illusion' of risk to keep other folks out.

That's the fundamental brilliance of Entropix, the hardware is largely commodity. The integration of those bits and pieces that will take some effort, but that's becoming easier by the second. At its heart, this really is a software play, but it masquerades as a tricky hardware play.

In terms of risk of software partner adoption, Nathan's other startup, Network Optix, proved that you really can boot up quickly and bring an entirely modern and great looking product to market.

Bravo on all fronts. I can't wait to build a camera for it!

Lytro and other 'non conventional' cameras are highly interesting to the larger investors, which leads me to believe that Entropix will have no problem finding investment dollars (the usual bane of interesting startups).

FWIW, Lytro has folded its consumer camera division and is pivoting to VR.

The first company that I thought about was Lytro as well. You are asking a lot for camera manufacturers (and VMS companies) to offer an ancillary cloud service when these companies are so used to selling boxes. In addition, you are doubling bandwidth (and storage) which will further raise the barrier of entry.

2, as an fyi, Entropix's pitch is that they are actually reducing bandwidth and storage since you can record video at lower resolutions and then 'cloud magically' enhance only the specific clips you need (e.g., record at 2MP 100% of the time, enhance to 18MP for the 1% of the time you need it). I actually think that part sounds reasonable (of course, assuming that it works / really does the enhancement which cannot be assessed since its not released).

That says, as I stated above, I do concur about the business structure concerns.

Effectively doubling the camera count by adding another 1080p (2MP) stream will be a challenge for many deployments.

Entropix's response:

"We can achieve 9x with roughly a 1.5x delta over whatever it would be with normal single stream camera."

It will be interesting to see what that is in practice, given their is so many factors involved with bandwidth, i.e., it might be 1.5 but it could be far more against some cameras that are very bandwidth efficient, cannot tell until it's shipping.

A 10x feature needed only 1/1000 of the time is not worth paying double per channel, at least for most people I would imagine.

btw, what's the ship date?

do they have a demo video?

Very interesting idea here. One question that comes to mind to me is how surveillance video processed through this technology would be viewed by the courts and the legal system?

Just curious how this might play out in a court case scenario, regardless I agree with the earlier posts that sometimes our client users just need a little bit more zoom to get the key detail.

As cool as it sounds, I don't believe it's a viable business in its current form because of the following reasons, any one of which would give pause, but taken together...

  1. Needs all new (proprietary) hardware, i.e. more expensive cameras
  2. Makes you record/store at double the bitrate, but only view half of that (without paying)
  3. Most video is recorded and deleted without ever being used, so why am I buying cameras twice as expensive and storing double the data for something rarely used?

And where are all the stunning demos anyway? Surely they could post some 1080p to 18MP videos by now. But I've only seen a few still frames.

Yes, the comparison to Lytro is apt. But what that showed is that as cool as refocusing after the fact is, people really weren't interested in paying much for it, leaving Lytro to exit the consumer business altogether.

Yet, the CEO estimates the market size for post-processing of video clips to be ~40 billion dollars in 4 years, (towards the end of the video).

Does that sound right?

While I think it's cool to see CSI and NCIS-type sci-fi actually coming to life, I also see a possible stumbling point.

From the standpoint of security, how could a jury possibly convict someone "captured" in a video where 89% of the data was artificially created (8 out of every 9 pixels)? Obviously it's more complicated than that, but ultimately that is a possible expectation of this technology.

Read this IPVM report for free.

This article is part of IPVM's 6,807 reports, 913 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports