Universal Scale of Roughness

Hi Christian,Wow that would

Hi Christian,

Wow that would be quite an undertaking, working out the relative roughness of the surface your running on.

By it's nature it is going to be very subjective.

I guess you could find the extremes at each end then interpolate from there. I guess the only criteria is that it must be a real world surface ie not a 'bed of nails'.

Also you may want to exclude 'forgiving' surfaces such as sand and grass.

So, at one end you might have smooth concrete with a 1, at the other sharp stones with a hard undersurface at 10.

A skilled barefoot runner might be able to get to a Ten, I couldn't do that. I can run on gravel with a more sandy undersurface that moves around and can be compressed a bit more, that might be an 8.

It should be something that we should try to come up with a scale for so we can rate various courses and races. I guess there are so few barefoot runners in races that a database would take a while to build a comprehensive list.

I was contemplating running in a race in the UK when I'm there is a few weeks, I was using Google street view to see if I could work out what the surface was like on the course, not an easy thing to work out.

Regards



neil

As a barefoot runner and

As a barefoot runner and researcher, I am interested in this, too, especially as surface roughness has been my biggest challenge in running barefoot.

The problem is, as you note Neil, that we need to have multiple people rate the same surface at the same location to determine the level of agreement and amount of variation in perceptions. This just isn't a practical strategy.

Photos of surfaces can be very deceptive or unrevealing based on the color of the surface, lighting, and other factors. Ken Bob's examples may be about as good as you can reasonably get with photos, but they don't really do justice to the surfaces. For example, the black asphalt surface next to his rating of 4 probably is pretty rough, but from the picture, it looks like to me that it should be closer to a 5 or 6 (compared to his other pictures).

I did a little looking online, and as we might expect, transportation engineers have spent decades measuring road roughness, as roughness affects many aspects of pavement performance (longevity, noise, ride comfort, etc.). Here are some links to some overviews of how engineers measure road roughness:

http://training.ce.washington.edu/wsdot/Modules/09_pavement_evaluation/09-2_body.htm

http://www.umtri.umich.edu/content/LittleBook98R.pdf

http://docs.lib.purdue.edu/cgi/viewcontent.cgi?article=1859&context=jtrp&sei-redir=1#search=%22surface%20roughness%20measurement%20road%22

(this last one is very old, but useful to see how even early methods are far beyond our level of sophistication)

As these links show, we don't have the equipment to measure roughness objectively to the level of precision that engineers do.

The low-tech idea I came up with is to place a blank sheet of white paper on the surface and rub the length of a crayon (with wrapper removed) on the paper. I've tried this on several surfaces, from smoothest to roughest – hardwood floor inside my house (extremely smooth), concrete paver (mostly equivalent to new concrete sidewalk), approximately 10 year old chip seal road (rough), and 60 year old highly eroded concrete (even rougher). I'll post links to the rubbings if others are interested.

The rubbings of the rough surfaces have much more white space than the rubbings of the smooth surfaces. The distances between colored spots are also much greater for the rough surfaces, and there are fewer colored spots for the rough surfaces. It might be possible to measure with some sort of software the amount of whitespace in these images. It also might be possible to sample within a rubbing to get a manual count of spots per unit of area. Measuring distances between spots might be more difficult to do objectively. I'll have to think more about how this could be done, and maybe others have ideas, too.

These different aspects of the rubbings (% whitespace, number of colored spots, average distance beteween colored spots) probably could be combined to arrive at a summary measure of roughness. If this worked out, people could make their own rubbings, and rate the roughness of the surface. By collecting a lot of rubbings and ratings from a lot of runners (along with information about the runners and their barefoot experience), we might be derive an acceptable objective measure of roughness as well as learn something about how people perceive roughness.

If others think this might be worth trying, I would be happy to coordinate some testing of this approach (and other approaches, if others have workable ideas). If the testing showed that it might work, I could also set up a site to collect runners' rubbings and ratings, and perform the analyses.

Of course, there are challenges to my low tech approach. People may apply different pressure in their rubbings and that would probably affect the results. Also, we would need to sample the same surface at different places, as undoubtedly there is variability (some places on the same surface in a route definitely feel rougher than others). This approach probably isn't sensitive to the "give" of the surface, either (especially important for gravel and the surface underlying it). But even with these possible shortcomings, there still might be something we can learn with this approach.

And you thought people look

And you thought people look at you strange now for running barefoot. Wait til they see you out on the road with a piece of paper on it and rubbing a crayon on it.

When I go out for a run, I run on several different roads and on just one of those roads, it could have different levels of roughness. I think the best thing would be to use a mapping system like google or mapmyrun and do an overlay for it. You could rate how a certain section of road feels. Then when more than one barefoot runner runs on it, you could take the average of those runners. This would also help with races and that is what more people are interested in. Again, you would want to do the road in sections and not just say what the scale is for the whole marathon. The whole course could be smooth except for 0.1 mile where they decided to patch it with chip seal.

I find this to be over

I find this to be over analyzing, and really over the top. Then you have to put in the human factor of each individuals tolerance to these conditions, pretty much impossible to put on paper.

I did some more experimenting

I did some more experimenting today with the crayon rubbing method.

It doesn't work on river rocks (rounded, 2-4 inches in diameter) or grass, and I would guess wouldn't work on gravel (if bigger than half an inch). It doesn't work on the river rock because the surface is so uneven. And on grass, there was nothing to "grab" the paper and crayon.

I finally found a software tool (GIMP) to compute the % of whitespace (parts of the paper that have no crayon marking). This is the only quantity I could measure objectively (at this point), but it seemed also to be the most intuitive component of roughness. To translate whitespace into roughness, simply subtract whitespace % from 100% (0% whitespace would mean a perfectly smooth surface, all points of equal height). A large whitespace % means the foot essentially rests on a small number of points on the surface, which contributes to the surface feeling rough or poky.

It turns out that very smooth surfaces (hardwood floor, polished concrete, and a plastic office chair mat) give very different whitespace percentages, ranging from 44% to 67%. All of these surfaces seem to me to be essentially equally smooth, and I think I would have perceived them the same way before I started running and going barefoot. So this method doesn't work very well in on the smooth end of the scale.

Here are the whitespace percentages I found for other surfaces, in order of increasing roughness as I perceive it:

- concrete paver (like fairly new concrete sidewalk) = 55% (within the same range as the very smooth surfaces, although noticeably rougher, in my opinion)

- packed dirt = 71%

- sport court (plastic grid made up 1 cm squares with hollow interiors) = 78%

- 60 year-old highly eroded concrete = 81%

- chipseal = 81%

So the method can capture differences in perceived roughness, but within a very narrow range of measured whitespace. I did a few retests of the same surface (same spots) and was able to reproduce the whitespace values within 1-2% (for example, eroded concrete values of 80% and 81%).

These results suggest that at least for me, perceived roughness takes off once the whitespace hits 70-80%, with each additional small amount of whitespace increasing the perceived roughness dramatically. For instance, the difference to me between the concrete paver and packed dirt is very small, but the difference between packed dirt and eroded concrete or chipseal is very large, even though the difference in whitespace for the latter comparison is actually smaller.

As a practical matter, I think this method would not be useful for our purposes. Most ideas in science end up being dead ends, so I don't regret trying this.

The best way forward might be to elaborate Ken Bob's scale, perhaps with taking pictures of the surface from different angles. If we had a much larger set of examples, it might be more broadly useful. Also, whenever there are BRS meetups, groups of runners could rate the same surface to see how much agreement there is about this kind of thing.

I agree that the degree of

I agree that the degree of homogeneity of a run would vary hugely, and that subjectivity will still make any universality a problem. Still, there might be room for a way to describe a given single surface that has some objective meaning.

Now how would you objectively describe the trail that NakedSoleNate ran last spring that gave him such a hard time? Maybe there's room for negative numbers on Ken Bob's picture scale.

Thanks, TJ.JT, you're right

Thanks, TJ.

JT, you're right that Nate's experience probably would rate off the charts by almost anyone's measure!

I think the roughness factor is important to consider, particularly for those just starting out and those who have not spent much of their lives barefoot.

Eight months ago, I had been running barefoot for about 9 months and had become committed to barefoot running, but I seriously thought it was not possible for me to run on chipseal. Earlier on, I almost gave up the thought of running barefoot because of my reactions to what felt to me (then) as unacceptably rough surfaces. (With injuries in minimal shoes and a long history of injuries in regular shoes, I was painted into a corner and decided to try as hard as I could to follow Ken Bob's suggestions.) If we want others to join us, surface roughness is something major that must be addressed. Glib comments from those who have mastered rougher surfaces that "you'll get used to it" , "it's not so bad", or words to similar effects actually are discouraging, because it leaves those trying to learn to run on such surfaces feeling inadequate, not hopeful. That's one of many reasons I'm so glad that Ken Bob took such a positive tone in his book.

As I've thought about the roughness factor, it occurred to me that we in industrialized countries don't really have a good idea of what our human (perhistoric) ancestors faced in terms of surfaces. Almost everywhere we have radically transformed the surface -- we've paved it, graveled it (especially trails for hiking and running!), built on it, plowed it (and especially removed the rocks) for farmland, cut the timber (disturbing the soil, surface, and undergrowth), dumped sand or rocks on it (beaches and coastlines), or grown something new on it (like manicured grass). An archaeologist friend of mine says that he has to routinely ignore the top layer of sediment as hopelessly messed up by modern humans.

Why is this relevant? Well, we often argue for barefoot running because this is how we ran throughout most of our evolutionary history. But what surfaces did our ancestors typically have to contend with? Would they have to run barefoot miles on end on something equivalent to gravel or even chipseal? My guess is that they generally didn't, although they might have occasionally encountered rough patches. For over a year, I have been searching museums and ethnographic materials about what aboriginal people had on their feet (if anything) in rocky deserts that would surely have been challenging to run on barefoot. People living in such environments today seem always to be wearing at least sandals. "The Great Hunt" movie (http://www.youtube.com/watch?v=yfeNcsQZmZY) about persistence hunting in Kalahari is noteworthy because it looks like the hunters are running (barefoot) on soft and sandy surfaces.

If we knew what the human foot generally had to face in the past, we could get a better grip on what we might hope the _average_ human foot (and brain!) can reasonably handle in terms of our modern surfaces.

I think that was a brilliant

I think that was a brilliant idea, DB. I just have a quick question, do you think that a carpenters pencil will work better than a crayon? Crayons, from my experience, tend to leave a lot of white space when coloring anyway. Silly wax and it's clumpy-clumpitude. (I have since switched to markers for all my coloring needs.)

Do you think that your results would change if you used something that put a more reliable mark?

Vickers,Your suggestion is

Vickers,

Your suggestion is good. It gets at part of the problem, which is that the length of a crayon works against detection of roughness or smoothness if the surface is uneven. For example, if one end of the crayon is on a high point and the other end is on a low point, it won't grab (or color) much of the surface between the ends, because it is essentially lifted up by the high point. I apologize if this doesn't make sense.

I tried using just a quarter the length of a crayon and came up with mostly similar results. When I used the flat end (tip) of a crayon for rubbing, there's much less whitespace for both smooth and rough surfaces. For the rough surfaces, the smaller width of the end (vs. the whole length of the crayon) lets the crayon go into depressions and so they get colored when they shouldn't (according to the "logic" of this method, where only the high points should get colored).

I've thought of other low-tech ideas (small, wheeled vehicles [like a toy] rigged to record up and down movements; inking or painting a surface and then pressing a stiff piece of paper on it [essentially an upside down stamp]), but each has bigger flaws than the crayon method. I'm basically at a standstill for simple, mechanical ways for measuring roughness. Maybe we have among us some engineers, tradespeople, inventors, artists, or others who work expertly in 3-D for a living who might have better ideas.

TJ, if we could establish agreement among different people about the roughness of different surfaces (or could derive a formula to map variation in perceptions according to runners' characteristics, such as amount of barefoot experience, foot size, etc.), then we could get away with not measuring roughness directly. We could just use perceptions and be o.k. The only other piece that might need to be fixed is an accurate description of a surface, and that's where better photos or photos of the same surface from more angles might be needed. Too bad we can't transmit a Braille-like physical texture over the Internet! These descriptions would be for the purpose of letting runners determine how the particular surface they are interested in compares to those illustrated on the grand scale.

A good chunk of my scientific career has focused on measuring agreement when there is no necessarily "correct" answer, and perceptions of surface roughness would seem to fit in this category.

If we were to do a little study, the way I'd suggest going about it would be to have a group of runners varying in running experience, barefoot experience, physical characteristics (foot size, weight, height), and perhaps demographics (males and females, young and old) judge the roughness of the same set of different surfaces (the greater number of surfaces judged, the better). These runners would have to be at the same physical location (like at a meetup). It might be a stretch to get many and especially diverse runners together, but even 5 or 6 might be enough to get a sense of how much people agree, especially if they judge a number of different surfaces.

If we could get a set of people who would be willing to do this, I could offer some more suggestions about how to collect perceptions (a simple "rate on a scale of 1 to 10" probably would _not_ be the best way to go). Of course, I realize some people think this is a waste of time, so maybe we might not get many volunteers. And we'd need one or two people to coordinate everything "on the ground", so to speak!

I hope to attend the Seattle Barefoot Mile event on August 28. I could set something up there if the organizers are interested. The park where it will be held has a number of different surfaces in fairly close proximity, so it might be a good location.

TJ, since we will be doing

TJ, since we will be doing something entirely new, I think it's best to start small and then make corrections afterwards as necessary. So I'll contact the folks organizing the Seattle run and see what we can do there. After that, we can see what the next step would be.

You're right that the kind of judgments we want people to make are relative, not absolute. That is, we do want people to say which surface is rougher than another. Ideally we would be looking for rankings of the surfaces they are asked to judge.

The biggest challenge is logistical. At this particular park in Seattle, about 200 meters from the start of the run, there is a spot that has about 6-8 different surfaces within a radius of about 30 meters. Since it will be difficult to describe surfaces with words effectively and unreasonable for people to remember what a surface felt like a few minutes ago far away, it will be important that people can go back and forth between the surfaces as much as needed so they can give thoughtful judgments. So this park might be the ideal place to test this out.

I will keep you all posted on progress, and of course, if any of you have suggestions or comments, please let me know.

TJ, you and I are thinking

TJ, you and I are thinking along similar lines. The pictures will have to be taken days in advance and printed (in color) on the questionnaires. My plan was to have signs by each surface, designating the spot to evaluate. The questionnaires would also have a simplified map of the area, showing where each surface is. At this particular part of the park (blessed with so many different surfaces in close proximity), each of the surface locations should be visible from every other location. Also, it should only take a few minutes to visit all of them at a walking pace.

Your idea about asking volunteers to write their own description of what the surface felt like is great. These responses will not only serve as reminders to the volunteer filling out the questionnaire but also provide some useful data.

Thanks for all your thoughts

Thanks for all your thoughts on this, DB. I have often thought (while running on gravel roads) "I'm pretty sure my distant ancestors didn't have to deal with miles of this!"

I think the survey ideas are good. I agree with the OP that the experience of different surfaces has a large subjective component -- and I think surveys could help you grok those feelings and identify the key physical factors related to comfort.

I don't mean to throw a wrench into your plans, but I wonder if delving deep into data acquisition on surface roughness is the right place to focus energy right now? Personally I think that surface difficulty/discomfort is strongly related to softness as well as roughness. Maybe I missed something (I have not read every single word of this thread) but is there some place that you account for the softness of the underlying surface, in addition to the roughness? For example, say an equal load of 1/4" crushed gravel was spread over concrete, and over soft dirt. As I understand it the same roughness metric would apply to each. However, I'm pretty sure the stuff over concrete would be less pleasant to run on. If the rough stuff has room to MOVE, then it's more comfortable on the feet.

(However if you do get into measuring whitespace there should be lots of analytical tools out there beyond GIMP. All sorts of graphical analysis subroutines have been written for GIS, forest ecoogy, etc.)

I'm dying to know what you find out. What's the whitespace on a road made of 3" diameter crushed gravel? (95%??)

---

ps DB and OP, may I suggest an alternate, objective metric? The Runner's Speed.

I think what most runners really want to know is if the experience will be difficult -- uncomfortable, exhausting, slow? I'm not sure a simple index of surface roughness really expresses that comfort factor. Also a more complete description of surface characteristics would probably take something like a multivariate ordination to summarize. You could develop a roughness "index" from there, but it'd be a lot of work and it still wouldn't express what runners are really interested in, which is what THEIR experience is going to be.

Instead, you could focus on a very practical and measurable quantity: the runner's speed. This, I submit, is a decent proxy for the comfort of the running experience; fast speeds are likely to be on comfortable surfaces, or at least efficient ones; slow speeds are likely to be on uncomfortable or inefficient surfaces (deep mud is comfortable but slow). The data gathering here would focus on the actual experience of runners. You could measure the speed of N runners on J different surfaces, normalized the speeds by using something (concrete) as the runner's "standard" surface, then simply compare speeds between surfaces using posthoc multiple comparisons. Then you could rank surfaces in terms of speed.

You would quickly discover which surfaces are really different in terms of performance. Then the ones that are especially fast or slow could be singled out for actual physical examination to see what makes them that way.

Just my 2 cents. -stomper