Navigate / search

Computational Sound: the next frontier in hardware+software+internet revolution

Lets state the obvious upfront: we are witnessing a hardware tsunami. Cost of developing hardware has gone down, computational power accessible even on mobile platforms has gone up, connectivity is widespread, and amazing things are being done at the intersection of hardware + software + internet.

One of the areas that has benefitted tremendously from the this revolution is the field of imaging. Over a few short years we have gone from traditional 2D photography to advanced digital imaging, 3D imaging, augmented reality and virtual reality. Cameras are now able to capture a tremendous amount of information at very low cost leading to amazing innovations in both image production/processing as well as image mining for data. At Lux, we have invested heavily in the computational imaging space with our investments in MatterPort (3D scanning), Planet Labs (satellite imaging), CyPhy (aerial/drone imaging), LensBricks (computational imaging), Orbital Insight (image analytics), AltSpace (virtual reality).

But one area that I feel we are just barely starting to recognize as almost equally important is sound. I am really interested in what I have been calling ‘Computational Sound’. Sound is present and adds color/data to wherever we capture images (for example consider all the examples I listed above and how they would benefit by capturing the right sound for their applications). But we have taken it for granted, or have had very complicated and expensive ways of working with sound. I am interested in how we capture sound, manipulate it, analyze it, play it and share it. As we try to capture smart sound in an otherwise noisy environment, we need better sound collection devices. We need better microphone technology, but we also need to be able to triangulate sound from multiple sources to paint a more holistic 3D view of sound, or only capture sound from a certain direction or form a certain distance away. We should be able to create a 3D sound stage using computational techniques (and perhaps with use of simple hardware accessories if needed). We should be able to project sound in certain directions, selectively cancel sound in certain applications, and provide immersive experiences in sound not dissimilar form immersive experiences in virtual or augmented reality. Let me give just some examples:

  • canceling noise of drone in your ‘follow-me’ drone video
  • capturing low threshold sound from an interesting subject on camera despite noise from other sound sources
  • projection of surround, 3D or holographic sound in spaces. Create elaborate sound stages across devices and environments
  • optimizing sound in cars without using 6-7 speakers. Canceling ‘highway noise’, better hands-free phone experience
  • immersive sound experience in VR, including dynamic movement of sound with scene in view
  • new music instruments, synthetic sound
  • simplified/accessible sound engineering and encoding etc for consumer apps, games across devices, platforms
  • others…

As I have dug into this over the last few weeks and months, I have realized there are amazing scientists and engineers innovating in the space…but what we need now is entrepreneurs to focus on the space and bring exciting companies to life. I would love to hear from anyone working in this space. And hopefully become a partner in building a great company in this space. Contact me at

Optometry: a field ready for innovation

I have been wearing eye glasses since I was probably 10 years old. I don’t enjoy wearing them the least bit and don’t look forward to going through eye check-ups every 2-3 years. But unfortunately I have to. As a result I went in today to get my eyes checked, especially to make sure my prescription had not changed in the past few years.

I went to a local optometrist, who was extremely nice, professional, and qualified. However, I walked out of her office thinking this business is going to die. Or at the least get transformed in a very significant manner in the next few years. We just need more entrepreneurs focused on the space, and also put up the kind of fight Uber and Tesla have had to put up against regulatory fiefdoms.

A few observations:

  • American Optometric Association represents roughly 36,000 optometrists across the USA. They work out out of stand-alone offices, doctor’s offices or retail locations. They are the primary eye-care providers but it is clear that quite a bit of what they do on a regular visit should no longer be a part of their offering. They take up an hour for a regular eye check-up when that entire process could potentially be automated and done in a few minutes, either at home or at something resembling a minute-clinic. As health providers beyond eyesight prescription checks, they obviously have a role to play, but their offering and workflow needs to be altered and brought into the 21st century so they can provide better service at a lower cost to the system.
  • The equipment most optometrists use has practically not changed for 25 years. The entire check-up process still seems so manual, and frankly is no different than the process a local optometrist used with me ~25 years ago when I got my first set of eye glasses in Pakistan. It includes putting a variety of lenses in front of my eyes and asking if I see better with 1 or 2, 3 or 4, 5 or 6?…and so on and so forth. Could there not be a more automated, quantitative and less subjective way of doing this. I sit in that seat always stressed if I may be straining too much to read, or squinting, or not focusing enough.
  • Given such a significant part of the test is simply changing focal lengths and refraction angles etc, why can’t this be done on a mobile device that I could use at home myself ? I am hoping Eye-Netra may have a solution for this.
  • The optometrist showed disdain for my online purchased WarbyParker eye glasses by informing me that the lenses were off-centered by 1mm, and if I had bought them at a store this would not have been the case. She may be right, but I think that is also a problem that could be easily fixed with a digital tool. Her solution was itself a tool that was probably invented in the 1950s and I have to believe this can be replicated a dozen ways using my cell phone and/or Oculus type device.
  •  As a part of the test, my eyes were dilated. Well, here’s the fun after-effect if you haven’t had this done before: you see really blurred for a few hours. So I left the store unable to drive, and unable to read anything on my phone or laptop. In fact had to cancel all my afternoon meetings, and 5 hours later I am still waiting for it to become safe enough for me to drive home.
  • Many optometrists also try to sell you eye glasses. These can cost anything from $150->$500. Through the founders of WarbyParker I know that they really should not and its all monopoly pricing from one company. In fact the actual cost is likely not more than $10-20. No surprise that when I did not show any interest in a purchase, the store brought up “high quality Asia-made” frames that they could sell me at very low pricing. They are kept in a separate drawer and not displayed publicly.
  • Since I am significantly near-sighted, when I try new frames at the store I have no idea what I actually look like! I am blind enough that I can’t see myself in the mirror when I don’t have my glasses on. Unfortunately there is still no solution to this (what I do? I take pictures of myself with frames on and see them later…alternatively, my wife picks for me :)

It is estimated that nearly three-quarters of the US population needs some type of vision correction. This is a very large market. And is desperately calling for attention from innovative entrepreneurs. I hope to learn more about them.

Startup sausage-making: from idea to company

Great companies are built around brilliant ideas, great people and strong execution. It takes a lot of work to take a brilliant idea and turn it into a fundable company, and we are proud to invest in founders who put their blood, sweat and tears into doing so. At Lux Capital we invest in seed stage, and sometimes even pre-seed stage, companies and hence often find ourselves as partners with the founders in doing the messy work that is involved in moving from idea to a company. We sometimes politely call it startup sausage-making.

This kind of startup sausage-making is messy. It is not always fun. And it is not always successful. Trust me we have lots of debates internally how much of our efforts should focus on such super early stage opportunities…but as entrepreneurs in our previous lives, some of us cannot resist the temptation to take the risk and put in the hard work in bringing such companies to life. We fall in love with ideas and we fall in love with entrepreneurs. We also don’t always succeed but for sure its always a team effort at Lux, and we give it our best.

I was recently discussing some of the things involved in such startup sausage-making with a friend and thought it was worth sharing so founders we have yet to invest in would know where we have some experience and can help:

  • Completing the founder team — this is the toughest, and frankly most frustrating. You can’t really hire a search firm to help find the missing founder (trust me, I have tried). Yet, we regularly meet founder teams that are incomplete. Either all technical with nobody understanding the business/market/customer needs, or all business and nobody really having the technical depth and breadth to address issues quickly. At Lux we maintain a wide network of experienced entrepreneurs across all kinds of spaces who can be pinged as opportunities arise where they might be able to play a critical role. But there is no magic formula to it. The process of dating takes a long time, and even then there is no easy fit. Arranged marriages have their advantages and disadvantages. Most important for us at Lux is to have an honest dialogue with existing founder(s) on what we think is missing, and then getting on the same page vis-a-vis what we would like in an ideal co-founder(s).
  • Technology to product — Technology doesn’t sell, product does. Yes all companies go through this process but in some industries this step can be too time consuming, expensive and/or unpredictable. It is important to understand that early. Supplementing in-house competences and capabilities with knowledge gained from experienced consultants maybe the best way to triangulate on this at early stages. Understanding this also helps determine how much funding would be required to take company through to exit velocity.
  • Customer discovery — Companies get built around customers. Early conversations with customers help build the right product, the right team, and the find the right investors. This is very tough if the business team is incomplete or does not exist. Investor team can help pitch in, and we open lots of doors, but somebody is needed to own the problem and build a vision for the product with customer input in mind.
  • Financing — Very early stage companies have to delicately balance their pitch between being bold/visionary/change the world and here’s what we will do in concrete/focused/viable fashion over next 12-18 months. We spend time helping companies understand not only the art of pitching to investors, but realizing that a good pitch is also a document that can be used to help with recruiting and in organizing execution.
  • Ugly details — Depending on the nature of the startup there are often lots of loose ends that need to be tied up for the company to succeed in the long run. For example company registration issues, IP licenses, non-compete agreements, no-hire agreements, visa issues, payroll, HR policies etc. Oddly these these are things that pester entrepreneurs most (esp first-time entrepreneurs) but fortunately these are areas that investors understand well and typically have lots of vendors and partners who can help companies gets et up the right way.

SBIR/STTR grants are great. ‘SBIR shops’ are not.

As a partner at Lux Capital I search for, and invest in, technologies that are often based on fundamental innovations in science and engineering. Our team is not turned off when we encounter deep scientific facts or publications. In fact, we thoroughly enjoy learning about them – a lot of our internal communications are around discovering/sharing amazing inventions in all kinds of domains. We regularly spend time with academic researchers and small companies who are recipients of non-dilutive SBIR/STTR grants. As such we are supporters of government funding programs to sponsor translational research and invention-to-commercialization efforts.

However, it is unfortunate that along with many amazing entities that truly benefit from SBIR grants, I also routinely see entities that are frankly called ‘SBIR shops’ in the industry jargon. These are life-style organizations that frankly feel like leeches feeding off these government programs without any real interest in commercialization, or in bringing their technology into the real world. They go from one SBIR grant to another for years, some times decades, and their teams have professional grant writers who are paid to do nothing else but submit successful grant applications into multiple agencies. My problem with these organizations is not that they are not accomplishing the explicit goals of the SBIR/STTR programs, but that they are sucking away precious resources from the system that would otherwise be available to enterprising technical entrepreneurs, especially younger researchers, who are genuinely interested in commercializing their inventions. They are hurting innovation in the long run.

SBIR/STTR programs are sponsored by several large government agencies (such as Departments of defense, agriculture, commerce, energy, homeland security, EPA, NIH, NASA and NSF). Approximately $2.5-3B are awarded in such grants every year. These are not small dollars, and if utilized properly, they can provide much needed capital to support commercially relevant technical innovations in new biology, materials, devices, systems, and software. SBIR awardees produce >2500 patents per year and more than 15% receive outside funding to continue with their ventures. A lot of good has come from SBIR grants in the past, but much can be done to improve the system, and to get rid of the ‘SBIR shops’.

Formally none of the awardees can receive a Phase III grant without external funding towards commercialization, but these ‘SBIR shops’ have figured out how to perfectly game the system. They move from agency to agency, and from one set of grants to a different one. When I meet young researchers who are working on securing their first SBIR grants, they appear nervous and unsure if they will get it…In contrast, these ‘SBIR shops’ are easily identified as they exhibit complete confidence in their ability to secure multiple grants worth millions of dollars per year to continue to operate for years to come. They abuse the system, and can be easily identified when they pitch to venture investors, but unfortunately the SBIR management system itself doesn’t have enough resources to put the right checks and balances in place to stop them in a timely fashion.

I am not sure what the solution is…maybe angel/VC community can be more actively involved in awarding and management of SBIR grants? Maybe there should be a limit on the total number of grants (not dollars) that can be given to any single entity/organization? Maybe there can be an easier and faster way for people to identify ‘SBIR shops’ so award administrators can investigate abuse of the system? Other ideas? Lets improve the system so our resources can be used to more efficiently foster technical innovation.

As an aside: if you are a technical entrepreneur int’d in securing SBIR funding, or an investors looking to unearth gems among previous recipients, SBIR Source can be one of the more helpful resources.

Celebrating Mario Molina, Nobel laureate in Chemistry, and my Ph.D. advisor.

Mario next to the equipment I used to understand phase transitions in atmospheric aerosols.
Mario next to the equipment I used to understand phase transitions in atmospheric aerosols.

I am headed to UCSD tomorrow to join generations of previous graduate students, post-docs and research collaborators of Mario J. Molina to celebrate his 71st birthday, and his receiving the Presidential Medal of Freedom in 2013 (Mario Molina Symposium).

Mario was my Ph.D. advisor at MIT from 1998-2003. Mario’s contributions to science earned him the Nobel Prize in 1995. He, along with Sherry Rowland and Paul Crutzen, discovered the fast chemical reactions in the stratosphere that were destroying the ozone layer. And then worked tirelessly to prohibit the use of CFCs by industry, leading to the passing of Montreal Protocol that banned the use of ozone depleting substances. He was the first Mexican-born citizen to receive the Nobel Prize.

My Ph.D. work with Mario focused on phase transitions in atmospheric particles, and heterogeneous chemistry that affected the troposphere. Climate models (that are used to predict global warming) carry uncertainty due to their inability to predict indirect effects of aerosols in the atmosphere (cloud albedo etc), and my work focused on understanding when, where and how cirrus clouds form, and how global warming inducing carbon soot particles are washed out of the atmosphere.

Some of the work we published together:

My time spent with Mario is among the most memorable in my life. I didn’t just learn chemistry from him…he was a role model on how science (and engineering) was critical to solving some of the most important problems facing the world. He was a strong supporter of environmental causes, but his views were not based on philosophy – he was a student of science and rational thought. He was a lead author on the IPCC reports, and we frequently discussed the benefit of such outreach beyond the scientific community. He took his role as an ambassador of science and environment very seriously, and worked tirelessly (including sometimes sleeping overnight in his office). We published technical papers together, I read and commented on drafts of his book on pollution in megacities, debated whether governments or industry would be first to take action on environmental causes, discussed science and engineering more broadly for developing countries, and traveled to Mexico together for several months on an international field campaign to understand air pollution in Mexico City. He was also instrumental and very supportive of my decision to not purse academia as a career after my Ph.D. and convinced me to spend some time at a DOE national lab (PNNL) before I began my career outside the academia.

Mexico issued a stamp in Mario’s honor.
Mexico issued a stamp in Mario’s honor.

There is a funny side-story about me becoming a member of his lab. After admission in to MIT’s graduate program in chemistry, I reached out to several faculty members to learn about their research interests. I was entering MIT after an undergrad degree at a liberal arts college, and my grasp of chemistry was rather basic. MIT faculty profiles were not just impressive, they were intimidating. Some faculty members  had distinguished memberships into national academies, others were founders of companies, and some were advisors to major corporations, institutions and Presidents of various countries. Back in 1998, Mario had a short web profile that probably some IT guy had dug up and put online. It had no mention of any of his accomplishments, but just his research interests in aerosols and chemical kinetics. I thought “here is this nice guy from Mexico, another developing country. He doesn’t seem to have many awards etc either. I should join him because unlike other faculty, he might actually have time and attention to give it to a lowly student like me”. I sent him an email indicating my interest in joining his group and asked if he would please point me to some of his publications so I could read up. Mario’s reply was essentially “I apologize I am not very internet savvy. I don’t have a website. But somebody has posted some of my publications online and you can go here to find them.” And the link he sent was something like…. I was flabbergasted when I got that mail. In my search for a not-yet-as-accomplished advisor, I had somehow landed in the lab of a Nobel laureate! He was a superstar in the chemistry department. I guess Nobel laureates don’t have to advertise the awesome work they do. He was humble, awesome…and loved hosting gatherings at his house. He was generous with me, his family met my wife before my own family did, and he even wrote a letter to the US State Dept to encourage them to give me a residency permit so I could stay in this country and start my company. In fact my startup conducted our early experiments in his lab with his students’ help.

Thank you Mario, for continuing to be an inspiration. We follow in our teachers’ footsteps.

Mario, my wife and I at his favorite eatery near the MIT campus.
Mario, my wife and I at his favorite eatery near the MIT campus.