omenti

do bats echolocate in color?

Tonight I waited by our pond hoping to see a bat.  Spring is finally here and there are more than enough flying insects to tempt a bat, but nothing.  The past few years have been terrible for bat watching - I'm afraid white nose syndrome has taken its toll.  Still, there were a few last year including a fearless little brown bat that would swoop within a few feet.

 

About fifteen years ago I built a few simple bat detectors with simple being the operative word.  A microphone that worked up to about 100 kHz, more than enough for many bat calls, and a  bit of electronics to sample and downshift the signal's frequency by a factor of 32, bringing it to a range where it could amplified and sent to headphones.  You learn a lot about their feeding habits watching and listening, but serious students of the critters were using sophisticated spectral analysis and high speed photography.  In a decade the field was revolutionized.

 

The richness of the signals is wonderful - you can move around in the time and frequency domains and "see" some of what they are up to.  Not only can they measure distance and relative velocity of their prey, but they can determine size and even what it might be to some extent.  You find yourself wondering how the bat experiences this richness.  We perceive rich visual and acoustic worlds, dogs and ferrets have fantastic senses of smell.  Do bats have something like color to sort out the different frequencies of their echoes?  They clearly use doppler shifts - just how do they perceive that?

 

It is easy to imagine what lower levels of perception are, but going the other way is frustrating.  As a teen I came across Edwin A Abbott's wonderful Flatland - a world with sentient beings all living in two dimensions.  I didn't appreciate that it was a commentary on Victorian hierarchy, but imagining what three dimensions must be like to a two dimensional being held me for some time.  Of course it is easy for us to imagine as beings of a higher dimensionality.

 

Most mammals only have two cones to sense color.  Your dog doesn't see the world in black and white, but his color vision is greatly limited from ours.  Although we only have three sensors, our brain sorts out the relative strength of the signals and creates something we call color.  We are sensitive to wavelengths from about 400 to 700nm and can perceive differences of about 1nm in the blue-green and yellow region and about 10nm in blue and red areas.  So although we can only distinguish a few hundred frequencies, our brain mixes these together to give a very large number.¹  But we have an idea what our vision is like, what about our dogs?²

 

Using a few papers on dog vision I cobbled a simple filter together.³  Here are some common colors seen by us and the same objects as seen by a dog. Of course I have no idea what "colors" the dog would imagine, the relative difference is what is important.  Although color vision is limited, you shouldn't feel sorry as your dog has a much better sense of smell than you and probably can't imagine how debilitating it must be to live in such a deprived world.

 

As rich as it seems our view of the world is extremely limited.  We tend to live in the time domain and can only discriminate between events a few tens of milliseconds apart.  Likewise we live more or less in the present and the notion of geological time is foreign. The wavelength range we see corresponds nicely with the output of our Sun, but it it is a tiny bit of the electromagnetic spectrum.  To make matters more interesting we live slightly in the past - our brain synchronizes some signal inputs moving us some number of milliseconds backwards in time and much of our visual perception is averaged over very long periods - we live in a mixture of the fairly recent stretching out to the past ten or fifteen minutes back.  

 

But our brains let us build sensors that give views into the richer universe

 

In order to keep the post short I'll stay away from multispectral, hyperspectral and ultraspectral imaging - at least not here as it is extremely important and deserves attention.  Basically you look at a lot of wavelengths to find any number of interesting things.  My guess is this is why Google was interested in a drone company.  The richness of Google maps is very sparse when you consider what is really available.  There are some rather impressive opportunities to understand the natural world - and one can imagine a number of business opportunities.  Timing and practicality are issues, but it is clear where this is going.  But back to bats and the present.

 

Smartphones give you some sensors, a computer and a connection to the Internet.  The computational piece is now sufficient to do sophisticated signal processing.  If you build certain types of sensors into the phone - or plug them in -  impressive work can be done.  Serious amateur, citizen and real science.  Environmental monitoring could quickly reach a state that has a major impact on pollution law.  There are the beginnings of that in China with very primitive tools.  The legal landscape for polluters may see a large shift in the next decade.  Atmospheric particulate, NO_x, CO, CO₂, and various water tests should be possible, although there are issues of calibration and repeatability that need to be addressed. Carbon dioxide measurements in the office may have a real impact on your own working conditions and even office productivity.

 

It would be fun to watch bats more seriously as an amateur, but the $5k and up for tools is prohibitive.  But there are smartphones...  All you would need is a good ultrasonic microphone and some programming.  The A7 chip in current iPhones has serious number crunching ability in its graphic cores.  It turns out someone has done it - the  price is about a tenth that of serious last generation kit and it is just as capable - potentially more capable.⁴

 

Developments like this are only the beginning of the directions are phones will take us.  Sensors are really important.

 

But what does acoustic imagining seem like to a bat?  Is the the similar to echolocation in whales or ...?

 

___________

¹ This varies a lot in humans and some of it is sex-linked.  Women are much better than men in the reds.  

² We sort of have a handle on it - there is still much to learn.  I have a neurological disorder that mixes color and sound, so I have no idea what orthogonal senses of hearing and seeing are.  This causes some issues that were probably selected against thousands of years ago in human evolution, but are benign after agriculture was invented.

³ Miller PE, Murphy CJ (1995) Vision in dogs. J. Am. Vet. Med. Assoc. 207:1623-34.  The filter is really simple and a rough hack, but it gives an idea.  Write if you want the mapping.  Here is a rough visual map, normal spectrum on top and filtered on the bottom:

 ⁴ In large quantities this could be very cheap, but this is low volume and they need to recover their costs.   

 

__________

Recipe Corner

 

Roasted Cabbage Salad

 

Ingredients

 

° half a green cabbage cut into chunks

° one diced apple - I used a Fuji, but almost anything should work

° a handful of pecan halves

 

vinaigrette

° 1 tbl cider vinegar

° 1 tbl balsamic vinegar

° 1 tbl dijon mustard

° 1 tsp maple syrup

° kosher salt and freshly ground black pepper

° 2 tbl extra virgin olive oil

 

Technique

 

° put the cabbage on a baking sheet, sprinkle with a few tsp of the olive oil, add some salt and pepper and spread into a single layer.  Roast at 400°F for about 15 minutes 'til the cabbage starts to brown.

° cool and transfer to a cutting board and dice.  Put into a bowl with the diced apple

° put the vinaigrette ingredients into a jar with a lid and shake to emulsify. Pour over the cabbage/apple and toss. 

° toast the pecans in a dry frying pan over medium heat until you can smell them, but *before* they brown.  Add to the salad and toss.

° I served with dried cranberries.  It was suggested that crumbled goat cheese would be amazing.

 

personal numbers

Early August marks the State Fair in Great Falls, Montana.  The latest farm implements, horse racing, a lame midway, deep fried everything, homemade lemonade, 4H and, of course, cow chip tossing.

 

As a high school student I remember watching a rather large and enthusiastic teenager, perhaps inspired by the Summer Olympics, decided emulating a discus thrower made sense.  Perhaps so, but his release was a bit late.  It made a lovely arc and, to the delight of everyone under about twenty and away from ground zero,  an unlucky spectator served to spare the ground the full force of its impact.

 

Apart from having a good throwing arm and a bit of style the trick to doing well in the event is to select one with the right moisture content.  Too dry and they break too easily - too wet and - well - they're disgusting.  The year before watching the discus thrower I was curious about a chip's energy density - after all - early prairie settlers supposedly burned them as fuel as the prairie is mostly treeless - and found dried dung provided about 15 megajoules per kilogram - about a third the energy density of gasoline and about 20 times that of a modern lithium-ion battery.¹

 

Food is not completely metabolized .  You can completely combust food in a bomb calorimeter and learn its energy density to good accuracy, but human metabolism is very complex.  It surprises people to learn the basic rules are still based on the work of Wilbur Atwater over a century ago measuring, as carefully as he could, energy going in and that which came out unused.  But life is more complicated.   Consider the a medium potato that weighs about seven ounces - or 200 grams.  Adding up the components.

168g water    0 calories

36g carbohydrates (4cal/g) = 144 cal

4.5g protein (4cal/g) = 18 cal

2.4g fiber (2 cal/g) = 5 cal

0.2 g fat (9 cal/g) = 2 cal

for a total of about 169 calories

But that's a raw potato.  If you cook it some of the work of digestion is already done for you.  Normal cooking to a soft texture add about 38 calories taking it to 207 calories.  But your body still needs about 6 to 10 calories to digest it and you need to feed the bacteria in your gut for 2 to 5 more calories leaving you with something just under 200 calories.  Getting numbers to better than ten percent accuracy is difficult and will change person to person.  Sorting out how much energy we need as a basic couch potato is possible but not easy.  Finding the energy expended by any kind of exercise requires specialized laboratory equipment and can be very specific to an individual.  Calorie counting seems like a hopeless exercise.³

 

The bottom line is many aspects of our biology and physiology are not well known and are difficult to measure.  For a variety of reasons there is a lot of confusion about something as "simple" as what kind of a diet should we eat for the best shot at a long disease-free life.   Even measuring activity levels accurately isn't easy, but can be useful.  While they are usually dramatically off in accurately measuring the amount of work you've done, they can give a number that can be used to develop activity goals - generally not a bad thing.   People have been doing this sort of goal setting for a long time and there is evidence the current wearables are being used by the same people - not a large segment of the population.

 

But there are many measurables that relate to health that can be measured and the costs are dropping quickly.  Companies like Zephyr have wearables that can store or transmit fairly rich physiological information - information that can be very useful if it can be understood in the context of your own body and what you are doing.⁴  Motion capture combined with wearable accelerometers is making a big impact in athletics and physical therapy - it isn't a big leap to imagine it going into amateur sports and the homes of the elderly and disabled. 

 

The major issues are those of understanding the information flow - sensor accuracy, calibration, data analysis - and what it means for each person.  Some types of measurements will be more meaningful and easier than others.  Importantly who gets to use the information and for what purpose - your doctor, your insurance company, your employer, Google...? Are you processing the information locally or trusting someone else?  How secure is it?  What does HIPPA compliance mean? Are there liability issues for problems caused by low sub-par information creation? (some sensor attachments are prescription-only).  And for locally consumed information what is the user interface and experience?  Is this BJ Fogg territory or ...?

 

Just beyond our bodies is the local environment.  How good is the air you're breathing and the water you're drinking?  You probably have a smoke detector and a carbon monoxide alarm, but what about the airborne chemical soup in your home and office?⁵  In the next five to ten years we're going to be able to measure these in our homes, schools and offices with our phones and aggregate the measurements on the Net.  There are serious issues about accuracy and analysis, but it strikes me this could be an enormous driver for environmental quality.

 

There are an interesting class of environmental impacts on our physiology, health and mental state that aren't easy to quantify - beyond the easily quantified self.  The right music at the right time, high quality paper and the right pencil have an impact on thinking about a math or physics problem for me.  Closer to physiology is the intersection of my schedule and the fact that short wavelength light increases melatonin production.  The last thing I need just before going to bed is extra melatonin, so I use f.lux on my MacBook to drop the screen's color temperature starting about two hours before my normal bedtime. This is done without my interaction and I've noticed I'm getting an extra hour of solid sleep - for me that is a big deal and has increased my quality of life.

 

A decade from now it will probably seem quaint that people once thought of mobile devices as limiting as what a smartphone currently is...  of course the form factor may not be that of a smartphone.  Perhaps distributed sensors with something like a phone acting as a concentrating hub.  

__________

¹ If someone tells you the energy density of modern batteries is crap, ask them if it is just wishful thinking on their part.

I couldn't find a photo of the event, but did find a video of a Montanan doing it just for sport.

² The cooking, digestion and gut metabolism numbers can vary widely.

³ While it is nearly impossible to use calorie counting to accurately determine your energy balance, it can be useful if you are estimating relative calorie consumption and exercise levels 

⁴ This is being used in professional, Olympic and some Division I level sports, medical monitoring and the military.  The most difficult piece in interpretation.

5 At Bell Labs I measured the carbon dioxide level in my office at about 900ppm.  While CO₂ is not toxic in low concentrations, too much can be dangerous and it has been shown that concentrations greater than about 700ppm interfere with creativity.  I had the HVAC people readjust the airflow to keep the room under 500ppm.  A very creative company I've worked with has a strict 500ppm limit on carbon dioxide in their building and stays away from treated floors and carpeting.

__________

Recipe corner

 

This was for vegan friend who also worries about gluten ..  it was pretty good.  For bonus points you could throw some walnuts or pecans on top.  It would be healthier without the frosting, but carrot cake needs frosting.

 

Mostly Raw Carrot Cake

 

Ingredients

 

° 2-1/2 cups shredded carrots

° 1 cup pitte dates

° 1 cup walnuts

° 1/2 cup shredded coconut

° 1/2 tsp nutmeg

° 1 tsp cinnamon

° pinch salt

___

frosting

 

° 1 cup cashews soaked overnight and drained

° juice of 1/2 lemon

° 3 tbl maple syrup (or simple sugar, agave ...)

° 1 tsp vanilla extract

° pinch salt

° 1/3 cup coconut oil (melted)

 

Technique

 

°  Blend walnuts and dates in a food processor and then everything else blending until it is an even consistency

° press into an 8x8 pan lined with parchment paper (the goo is sticky!)

° pop it into the fridge for an hour or more

° Blend the frosting ingredients into a blender 'til smooth, pour over the cake and freeze for a few hours.

° Let the cake thaw for ten minutes and slice with a knife warmed in hot water

 

 

 

 

choice and bias

I'm finally up to date with the first four episodes of Cosmos - enough that I have a sense of the show's approach.  The fact that Seth MacFarlane, Neil deGrasse Tyson, Ann Druyan and others were able to pull this together and to have it aired on network television is remarkable going way beyond their considerable talents as producers and story tellers.   I'm not the intended audience so my comments don't really matter, but I'll offer a few and step back to a larger question.

 

The science is up to date and mostly accurate with the inaccuracies being no more than nits that, if properly handled, would probably interfere with the story telling. Some of the special effects strike me as a bit on the cheesy side, but the cosmic calendar is a great device that can be used over and over and will probably make a few memorable points.  The second episode on evolution struck me as the strongest and most compelling of the lot and I'm happy to see the human element brought in as science is, of course, a fundamentally human enterprise.  And there are those magical moments of poetry that take remind you of Sagan's poetry in the original series.

 

The fourth episode struck me as the weakest.  Special relatively was presented almost mystically with far too much handwaving.  Better introductions exist and it would have been nice to have a clear story that excite a few enough to learn the basics on their own.¹  The special effects - buffeting around a black hole and the non-physical speculation - seemed way out of place and way over the top.

 

But choices have to be made.  The aim of a show like this is engaging story telling and making at most two or three points per episode.  While I would have done the relatively episode differently, I'm sure I wouldn't have come up with a flowing story and, left to my own devices, I probably would have ruined the other episodes trying to shoe-horn in too much explanation and I have a bias towards deeper than necessary explanation.  I'd love to know how 11 year olds are reacting to the show and what kinds are inspired and why.

 

A larger point is that storytelling education content choices need to be made.  The content and context of each element is not neutral and compromises are required.  This is an area where the communication of science often dies. It is particularly difficult to create content that engages a general audience while maintaining scientific accuracy.   Neil has been doing great work as a science communicator for the past decade and others, Alan Alda comes to mind,  have been addressing issues central to good communication.  

 

We often assume bits of science or technology are neutral when, in fact, they aren't.

 

It is often said that technology is neutral - what we do with it is where issues arise.  But this isn't so.  Take E = mc² - a lovely result that easily falls out from special relativity.   It is usually connected with nuclear energy - specifically fission or fusion.  In both cases energy densities are far greater than our benchmark of fossil fuels.  The release of energy is accompanied by the release of ionizing radiation - radiation that happens to be harmful to biological organisms like ourselves.  Background radiation - what we are constantly bathed in - is low enough given our biology and lifespans that we mostly don't notice its effects.  The technology that makes radioactive elements like Uranium greatly increases the amount of radiation - often to very dangerous levels and care must be taken.  Nuclear technology is far from being neutral.  Understanding a technology and its impact is a necessary element for wise use.

 

The internal combustion engine is far from being a neutral technology.  While the technology has positive and negative impacts,  some of the negative impacts were not readily apparently until the technology scaled.  The positive aspects of a technology may far outweigh the negatives in the short term, but we should be aware the balance can change over time.  The cotton gin is a cautionary example.  With its invention cotton production at scale became possible and, if you could exploit a very cheap labor source, extremely profitable.  Around the time of the invention slavery was dying out in America, but suddenly slavery made great economic sense in the South and ignoring the negative aspects of slavery was common among mill owners in the North.

 

I was intrigued to see a piece by danah boyd a few days ago - Is Oculus Rift Sexist?  Oculus Rift is a virtual reality interface to 3d virtual worlds.   You wear a head-mounted display that forms a computer generated image for each eye giving a pseudo 3d display.  Sensors note orientation and a computer builds an artificial reality that you can navigate and even manipulate.  The concept has been the subject of research since the 80s and is very sexy among much of the tech community.  Not many have used it as it has been prohibitively expensive until now - Oculus Rift exploits inexpensive graphics chips used for gaming and small color displays used in smartphones.  While the tech isn't novel, the comparatively low cost of entry is and there is a chance this may have an impact ... or it may prove to be a Segway for your face.  danah made an interesting observation based on her use of a VR world and I recommend taking a bit of time to read what she has to say.

 

There can be an enormous amount of variation across the human population.  An obvious difference is height.   The plot shows a  good fit to the distribution of human height from a large sample of several hundred thousand men and women between the ages of 18 and 45.  The double hump comes from the fact that height is sexually dimorphic - men and women are different on average. Of course there are women who are much taller than the average man and men who are much shorter than the average woman, but on whole the difference is real and obvious.   Given a large sample the distribution for each sex alone is bell-shaped - a gaussian distribution.²   Many technologies are size dependent and may be said to be sexist of they can't accommodate a sufficiently wide range.  Kitchen design tends to assume that the user will be a woman and, since women not as tall on average, the technology has some built-in sexism.  

 

Some aspects of vision are sexually dimorphic.  As danah notes there are some areas where aspects of stereo vision differ and one of these may be the trick head-mounted VR displays used to generate the illusion of 3d. ³  Although unintended such displays would be sexist if the effects are real.

 

Another technological sexism many of you have experienced is color dying and quite possibly video conferencing.  

 

You're probably wondering now ...

 

It turns out color discrimination is one of the sexually dimorphic component of vision.  Women are much better at discriminating color - particularly reds.⁴  Women see differences that are invisible to men in many colors that are even partly red.  I realize this and only buy socks that are blue, black and white and staying away from anything reddish that would have to match something else.  

 

Recently we've learned this difference and a bit of biology impacts human communication in a fascinating way.  We have a lot of capillaries just under the skin on our faces.  Our faces display subtle color change, mostly in the red, as we communicate.  Women are sensitive to this signal which is mostly invisible to men.  It may be advantageous for women to have face to face communication.  This type of communication is generally lumped in with other unspoken modes like facial and hand expressions, changes in posture - even odor. The best commercial color cameras and displays don't render facial color differences properly - not only are the colors inaccurate, but the video can't render the differences.  The best VR in the world won't hack it.

 

When we understand the impact of the non-neutral bias of a technology and its evolution we can make better choices.  And this questioning must continue as the technology and the context it is used in evolve over time.

__________

¹ The basic concepts of special relatively are not difficult and a high school student with trigonometry and algebra under her belt should be able to work through the basics. If she has a high school physics background she can probe much further.  General relativity, on the other hand, requires a few years of University level math and physics and usually isn't taught until the third year as course for majors and then revisited in grad school at a much deeper level.

If you want to learn SR on your own I'd recommend Spacetime Physics by Taylor and Wheeler.  A bit quirky, but with clear examples and problem sets and the first half or so should be accessible to a smart high school student.

² Sort of that it.  There are many distributions in Nature that appear to have gaussian distributions, but in fact the distribution is only an approximation.  For about 99.99% of the population the distribution is an excellent fit.  For those in the tails it is a different story and the fit no longer holds in this example.

A curious thing about the readership of this blog is the height distribution is extremely non-representive of the human population.  I can state that accurately by knowing the readership is very low (about 30 look at any given piece within a week of its posting) and the fact that the female readership is extremely skewed by the presence of four who stand six feet or taller.  

³ There are many cues for stereo vision - parallax is only one.  It turns out 3d movies and 3d television don't exploit parallax the way human vision works.  It is a hack and can cause problems in some users.  I'm one of them.  These problems may have a sexual dimorphism to them.  It may be that a much larger percentage of women have problems than men.  Something that is very interesting and open for investigation.  

 

⁴ The effect is  pronounced for wavelengths longer than 590nm and very strong above 620nm.  

 

__________

 

Recipe Corner

 

Something very simple - baked olives based on a family recipe of a friend with Sicilian roots.  These are so good!

 

Baked Olives

 

Ingredients

 

° 2 cups kalamata olives

° ½ cup dry red wine

° 3 tbl olive oil

° 3 cloves garlic, 1 sliced, 2 coarsely chopped

° 2 tbl marjoram (oregano may work, but it is stronger, so use care)

° 1 tbl chopped parsley

° 1 bay leaf

° freshly ground pepper

° a few pinches red pepper flakes

 

Technique

 

° preheat oven to 375°F

° rinse olives if they were salted, place then in a single layer on a baking pan or dish you can cover

° add wine, half the oil, the sliced garlic and bay leaf

° cover and bake for about 45m until they are swollen and smell *wonderful*

° crush the chopped garlic in a mortar with the marjoram, parsley and a bit of ground pepper.  

° when the olives are out of the oven poke each with a fork and stir in the paste, extra olive oil and pepper flakes.

° I like to let them sit for a few hours, but it is difficult not to snack.