How many eyes with Suuntos?

[Extremeophile]

This is from the Carlsbad Caverns Management Plan:

For those using a Suuntos compass and inclinometer, be sure to use just ONE eye and move the
instrument back and forth or move your eye up and down to accurately line up the station point
and the line in the instrument. Using the two-eyed method always introduces errors in your
readings due to the way that the brain fuses the image of the target and compass scale which are
not at the same distance from each eye.

I wouldn't look at this as "it doesn't really matter" or "whatever works best for you". One technique will introduce less error into the data than the other. Using two eyes might be easier and how you've been conditioned, but it's easy enough to re-train yourself to do it with one eye. There are already enough sources of error on surveys without intentionally adding another.

Derek

[Anonymous_Coward]

I don't have any errors because I don't do backsights. I caved with these guys that do backsights, and they were finding errors all day long!

[Jeff Bartlett]

I don't have any errors because I don't do backsights. I caved with these guys that do backsights, and they were finding errors all day long!

+1 for Andy
(and one eye for Suunto-reading, period)

[Squirrel Girl]

So, an experiment is an ideal thing to do, to see which is better for any individual.

Sounds like the perfect thing for you to do at this years convention!!!

That would be fun. Though it wouldn't be a "you" thing, it would be an "us" thing. We could set up a course and encourage lots of people to survey. Each person should use their favored instruments, and we could set up easy and hard shots. We could encourage people to report how they surveyed (how many eyes, etc) and see what happens. It would be an interesting statistical experiment.

We could have people do it in the dark, just like if they were in a cave. We have to encourage a lot of people to bring their survey gear. We could have people use their tapes and their distos and see how that goes, too.

[rlboyce]

Small prizes are always a good motivator.

If the goal is to look at one eye vs. two, it may be best to have everyone use the same instrument or only have a couple of sets for everyone to use. Error among all the individual survey instruments may be a hard variable to quantify and might influence the data. Another way to put it, some of the error you see might be because one of the cavers' instruments was faulty.

However, if you want to make it into a full course, it probably would be best to let people bring their own instruments so multiple surveys can occur at the same time, making the process faster.

I think the biggest challenge of setting up a course will be precisely setting each station.

[Squirrel Girl]

Small prizes are always a good motivator.

If the goal is to look at one eye vs. two, it may be best to have everyone use the same instrument or only have a couple of sets for everyone to use. Error among all the individual survey instruments may be a hard variable to quantify and might influence the data. Another way to put it, some of the error you see might be because one of the cavers' instruments was faulty.

However, if you want to make it into a full course, it probably would be best to let people bring their own instruments so multiple surveys can occur at the same time, making the process faster.

I think the biggest challenge of setting up a course will be precisely setting each station.

There's certainly merit to an experiment to multiple people using the same instrument. But it might not answer a question of how well you do with your instruments (that you're familiar with). For the purpose of science, we should do both experiments, and would probably, as most people won't have their survey instruments along, and even if they did, we could trade off.

We'd have a whole week, and wouldn't have many stations. It wouldn't take long.

Am I being a smart ass, or do I not "get" your last statement. There's no problem setting up the stations, it's how we decide what's "real" as an answer to the measurements between each?

Hmm, small prizes. That's a thought. Of course that would mean there's a "win." Closest to correct (which means there's a known answer--anybody have a theodolite?), or fastest yet most correct?

Or a tiny token for just participating?

[Scott McCrea]

I might be convinced to donate some Sight Lights for prizes.

[rlboyce]

Whoa, those sight lights are pretty cool! Such a good idea, it almost makes you wonder why they don't come with the Suuntos in the first place.

I suppose knowing the exact position of each station in relation to another is not a total necessity, because if you take enough readings with enough instruments you will approach the actual value (which you will have to calculate after every last reading has been taken). If I were feeling industrious though with enough time on my hands, I would probably try to set stations as precisely as possible in the beginning. Knowing the precise value will give you a benchmark to easily compare your readings with. Also, it's very hard to know which side of the actual value your reading is on if you don't have that benchmark starting out. To complicate things, if you only have a few instruments, all those instruments may give readings that are consistently skewed to the one side of the actual value. That means if you took the averages of all your readings that were skewed to one side, this would not put you any closer to the actual value. For example, let's say you obtain a reading of 1° using one eye and then another of 2° using two eyes. Which reading is closer? Is the actual value 0.5°? 1.5°? 2.5°? With enough readings and enough instruments, you'll eventually have a pretty good idea what the actual value is, but if you already knew the value beforehand it will save you time and you won't be constrained to using many instruments. Of course, it takes time setting precise stations, so it's one of those "pick your poison" situations.

However you set it up or if you have the course at all, I wish you luck and hopefully some fun!

[Crockett]

Make loops. Least loop closure error "wins".

[Bob Thrun]

Yes, I have read this from the user manual as well. Actually, I was looking for data. I believe in numbers!

I've come across an older Compass & Tape article that claims you should use one eye for a compass, not two (page 16): http://www.caves.org/section/sacs/back/v06i2_21.pdf I was wondering if anyone else can back up this author's claims or disprove them with another error analysis. I am especially interested in inclinometer data since this author has not included it.

I have my comments on page 25 of the cited publcation.

There is an article, "The Effect of Binocular Vision Disorders on Cave Surveying Accuracy", in Vol. 33, No 2 of Cave & Karst Science, published by the British Cave Research Association. That article has no experimental data. My comments on that article appeaared in the next issue.

When a person looks at an object, his eyes automatically align themselves so that the images in each eye coincide. This is the reason, the ONLY reason, that two-eyed reading works. If there are no objects common to both eyes or the compass reader is concentrating on the compass, there will be errors.

Eric Hendrickson reported (Compass & Tape No. 45, p. 2-3) that some persons consistently have errors. I suspect they were using two eyes and concentrating on the compass, though I have absolutely no evidence for that suspicion.

There is a compass course every year at OTR. George Dasher did an analysis of reader errors a few years ago. There would be difficulties getting meaningful one vs. two-eye data from a compass course. I would get errors reading two-eyed unless I cheat by reading one-eyed with both eyes open, (Many pistol shooters aim with one eye while keeping both eyes open.) Others would cheat without realizing it. The only way I can think of to ensure two-eyed reading is puting blinders on the compass, pieces of cardboard sticking up and down so that the compass eye can't see the target or objects in the vicinity of the target.

[Extremeophile]

Sorry to nit pick, but I think the terminology is getting confused here. Precision relates to how fine a measurement can be taken (e.g. nearest 1 degree, 0.5 degree or 0.1 degree), and how repeatable those measurements can be taken. Accuracy is how close the measurement is to a "true" value. I've argued that using two eyes (if done the same way every time) can be just as precise as one eye, but is inaccurate. I also believe precise but inaccurate measurements can result in low loop closure error, but the loop may be rotated. I think it's important to have an independant true value to compare against. Theodolite with a measurement to Polaris should work as Barbara suggested.

[Squirrel Girl]

Derek, the whole POINT of this is to nit pick. If not, we wouldn't care one eye or two.

Thanks, Bob. George, where are you?

[Aaron Addison]

Bob Osburn and I were discussing this today at lunch and trying to design a test for this. The one eye method seems to be the de facto standard and there is little empirical evidence to support it. In theory it is because of parallax, but at what distance is this an issue? 50'? 10'? Are the errors consistent at all distances? Would it be better to have non-magnetic tripod mounted suuntos for the control course to eliminate the variable of how cavers hold a compass?

Another possibility we discussed that whatever the error is introduced via the two eye method was within the "noise" of the accuracy that most cavers are able to read the instrument anyway?

Just thoughts,

Aaron

[Bob Thrun]

Bob Osburn and I were discussing this today at lunch and trying to design a test for this. The one eye method seems to be the de facto standard and there is little empirical evidence to support it. In theory it is because of parallax,

The problem is not parallax. Parallax occurs when both eyes point in parallel directions. The problem with reading with two-eyes is that the eyes do not have to point in the same direction. When a person looks at something close, his eyes change focus and cross at the same time. I can focus and cross my eyes independently and at will. I am aware of my instinctive focussing and crossing. When I do the dominant eye test that Phil Winkler described, I see two fingers. I can choose to make either eye temporarily dominant and make one finger or the other disappear.

When I first read a Suunto compass with two eyes, I noticed that my eyes were crossing and changing the reading. I began reading with one eye and have been doing so ever since. I think it would be futile to try to collect test data with subjects who are aware of what is going on.

Another possibility we discussed that whatever the error is introduced via the two eye method was within the "noise" of the accuracy that most cavers are able to read the instrument anyway?

Jos Burgers gave a small set of data that shows reader and instrunent differences in Compass Points 30, http://www.chaos.org.uk/survex/cp/CP30/CP30.pdf But read my comments on the data in CP 31 http://www.chaos.org.uk/survex/cp/CP31/CP31.pdf .

DC Grotto ran a compass course in 1981 to study the reproducability of the readings. It involved more than 1600 readings. I finally wrote a 12-page report. I handed out copies at the 2009 ICS. Some of you may have gotten a copy.

[WVCaver2011]

To be honest, I started off using both eyes until I started having a hard time focusing on the station I was measuring to. It would seem to wonder about due to the fact that I was trying to focus on that station and then would focus on the numbers within the suunto. Try this, put a finger in front of your nose. Close one eye and look at the finger and close the other eye and look at the finger. Your finger appears to move... On a long distant shot this may not be all that significant but it may still give you an error with your azimuth of 1 degree. If you have a station thats closer, you will be looking at a larger degree of error.

Another example: A friend of mine always uses one eye while I used to use two. We were at stations that were roughly 15 feet away from one another. We went back and forth and could not get anything under 2 degrees. Eventually we gave up and blamed it on a magnetic anomaly. I later realized what was causing it because the next stations were further spaced and we were under our 2 degree error.

I had literally spent 20 minutes on a stations that were only 5 feet apart before contemplating why my error was 5 degrees or so off. All of you surveyors out there who use two eyes, how far are your stations apart on average?

So on a personal note, I would like to add that looking through a suunto with one eye and focusing on the station with the same eye proves to give you a straight, accurate measurement everytime while using 2 eyes will give you a slightly angled, inaccurate measurement everytime. I am not arguing for or against anyone, I would just like to share why I use one eye.