Birds by Binocular

Which binoculars should I use? 7x42? 8x30? 8.5x44? 10x40? I buy binoculars in the hope of seeing more birds. Having bought a variety of binoculars, I decided to note down the birds I actually saw and look to see which of my binoculars suited me, and my usual haunts, best. Actually, if choice of binoculars was a big factor in deciding what birds you saw, I would expect to see it mentioned a lot more in the instructions for volunteer bird surveys I have taken part in, and in books such as "Bird Census Techniques" (Bibby, Burgess, Hiil, and Mustoe). Nevertheless, I always like to make sure even of the obvious, as there is always a small chance of finding out that what was obvious isn't so.

Designing an Experiment

As far as possible, I have made this a properly designed experiment, albeit a very simple one. I have four mid-priced binoculars available, of the specifications hinted at above and described more fully below. I decided to make a total of 32 birdwatching trips, divided into 8 blocks of four trips each. Each block is a set of 4 visits to the same place, taking, as far as possible, the same route. Having decided this much, I sat down and used dice to decide at random which binoculars to allocate to the 4 visits in each block; in fact I ended up with two 4x4 tables, which I shall write here, one above the other. (I wanted to continue regular visits to Devizes while seeing a variety of other sites).

The final column in the table shows the place visited. You can see that each row includes A, B, C, and D, arranged in a randomly chosen order. This tells me which binoculars to use for which visit. The fact that this order was chosen at random is crucial. As the year goes on, I expect to see different birds. Things happen to the sites, too; my last visit to Slimbridge was different because they closed the path out towards the estuary for winter, and my last visit to Devizes was different because they blocked access to a large stretch of towpath. (British Waterways had previously said they would provide alternative access. What this turned out to amount to was putting a big arrow on the fence that pointed in no particularly useful direction, together with a notice saying "please use the public footpath"). But because the order in which I used the different binoculars was chosen at random, no binocular was any more likely than any other to be favoured or penalised by these changing circumstances. If - to take an extreme example - one binocular always provided more birds for a particular site than the other three, the chance of this happening at random would be the chance of winning 8 games of chance in a row, with a total of 4 competitors: 1/4⁸ = 1/65536, because the use of dice to chose a random order has changed this from a question about whether different days provide the same selection of birds to a question about how likely you are to repeatedly pick the winner in a succession of games of dice, assuming that there is nothing special about any particular binocular. Despite the trials and tribulations of daily birdwatching, if any one of my binoculars was uniformly better than the others, I would come back with strong evidence far that fact.

Binoculars

The binoculars I used were as follows:

Bausch & Lomb Discoverer 7x42

Field of view 140m at 1000m. There is plenty of eye relief, and this is pretty much what I see (measured at only 5m, I'm afraid). My only quibble about the field of view with this binocular is that the outer edge of the field of view focuses at a different place than the centre. This is the only roof prism of the four, and has the lowest resolution, even after allowing for the lower magnification. It is generally agreed not to be in the running as a high-resolution binocular, but it is compact, robust, has a wide field of view (for a roof prism binocular), and is easy and comfortable to look through. The one I have was bought ex-demo from an binocular repairer. It seems in good condition, except that to get the dioptre adjustment balanced, you need to move the adjusting control right over to one side. I know this isn't just my eyesight, because when I have it in this extreme position, I can turn the binoculars upside down, look through the eyepieces (so swapping eyepieces to use the left one with my right eye and vice versa), and still have the eye adjustment OK. I have seen complaints that the dioptre adjustment is easily displaced from the central position, so it is just possible that it has been deliberately adjusted in this way, so that the 'balanced' position is one that is less likely to slip. I believe that only relatively new Discoverers have phase coating. I don't know whether mine has it or not. (Phase coating corrects a problem suffered by roof prism binoculars, but not by porro prism binoculars. In general, it is easier to make an optically good porro prism binocular than an optically good roof prism binocular).

Nikon 8x30 EII

The field of view is stated as 8.8 degrees, which I reckon is over 153m at 1000m. In fact it says 8x30 8.8^o WF, and I reckon it has earned the WF!. I seem to be able to see about 150m of this, with some straining and adjusting of eye position; I wonder if this isn't wider than anybody can sensibly take in at once. This is the 'little brother' of the famous Nikon SE 8x32. It is cheaper, and has a rubber covering that tends to come unstuck. It has a wider field of view, but less eye relief, and reportedly the area of highest definition is not quite as large as in the SE. However, I do not believe that my eyes could spot any increase in resolution beyond the EII. The SE has a reputation of being rather hard to look through; if you don't hold in the correct position the view 'blacks out'. The EII has this feature to a lesser extent, although I find I can restore the view quickly by moving the binoculars slightly away from my eyes to line them up. After a while, your accuracy increases and the problem disappears, but it is not a relaxing a binocular as the 7x42 or 8.5x44.

Swift Audubon 8.5x44 820ED

Advertised field of view 144m at 1000m. Perhaps because of my eyesight, and the eye relief, what I see at 5m distance is almost exactly the same as with the stated 140m of the 7x42 B&Ls: something under 70cm at this distance, which scales to 140m at 1000m. This is the porro version of the 820, and has ED glass. I have both the ED and non-ED versions. The ED version is superior, but to be sure of the difference, I have to resort to taking pictures through them with a digital camera and then enlarging them greatly on the computer. These binoculars appear to have the same very high resolution as the EIIs, but are far more relaxing to look through; in fact they give me my favourite view of all my binoculars. Their sole optical drawback is that the eye relief that may be limiting my view in this case. They are bulky, even for a porro prism, and, like the other porro prisms, not really waterproof. (They are marked and advertised as waterproof, but that turns out not to be the case in practice. Swift use a particular definition of the word waterproof that does not cover, for instance, using them in the rain for a few hours). Despite all that, these are my favourites, at least until I find something that clearly shows me more birds.

Optolyth Alpin 10x40

No markings to show field of view, but it seems to be equivalent to about 104m at 1000m (so the apparent field of view is in fact wider than that of the 7x42s). These are the old-style Alpins; I bought them second hand on ebay. Apart from a few areas where the paint has been rubbed off, they seem in excellent condition. Optically, and taking account of their greater magnification, I would say that they are not quite on a par with the Nikon and Swift Audubons, but they are probably still good enough for my eyes not to be able to tell the difference. In their heyday, the Optolyth Alpins had a very strong following: they are quite compact for a porro prism, and unusually light, even when compared to a roof with the same magnification and objective. I notice that in the US, the old Alpin design is now being offered again, as an alternative to the new Alpin NGs.

Digi-bin

As a rough indication of the resolving power of the binoculars, I used a Contax SL300RT* to take pictures through them, with the lens set on full zoom, which is roughly 3X. I enlarged and cropped the centre portion. Each picture is taken from the same position, and enlarged by roughly the same extent; what you are seeing is a very small part of a resolution chart downloaded from the internet and printed out with an ink-jet printer. The detail here is so fine that a perfect image would show visibly jagged lines produced by the ink-jet printer, but the border between dark ink and light paper (coloured because I didn't border compensating for the room lighting) would be sharp.

The images are in order of magnification, from left to right, so we have the 7x42 Bausch & Lomb Discoverer, the 8x30 Nikon E2, the 8.5x44 Swift Audubon 820ED, and the 10x40 Optolyth Alpin. To show where my eyesight comes in this, consider the lower set of four lines, which you can resolve in all except the 7x42 shot. Even looking through the 10x40 Alpins, I can see that this is not a single solid bar, but I cannot tell how many lines make it up. Through the other binoculars, I cannot tell whether this group of lines is in fact one solid line or not.

These two pictures show what the camera might do with a perfect pair of binoculars. I have simply walked in close enough to get roughly the same size picture as I did at 7X and 10X.

My Eyesight

I should warn the reader that my eyesight is unlikely to be the same as theirs. I have unusually strong short sight and some astigmatism as well. I always use my binculars with glasses on. Apart from anything else, my short sight is so extreme that most binoculars won't focus for me without glasses. They are fairly close-fitting, though, so I usually get the full, or nearly the full, field of view with modern binoculars. Even with glasses on, my distance vision is not very good. I routinely use binoculars to scan for birds even when I can't see any sign of them, because I know I can find, and identify, birds easily with binoculars that I can't even see with just glasses. This is one reason why I prefer binoculars with a wide field of view.

Results

The main conclusion is that no binocular showed itself as convincingly better at showing me a variety of birds than any other binocular. The only reliable way of seeing more birds is to change where you go birdwatching!

Here is the output from an ANOVA of the figures, using Minitab. The basic measurement is the number of different birds I saw on a single visit. So the mean of 16.6 birds for C1 shows that, on average, I saw 16.6 different birds with binocular C1 (the 7x42 Discoverers). A Site corresponds to a set of 4 visits to the same place, which I might have called a block. So the 16 times I looked for birds in Devizes end up as 4 different 'sites', according to this - which isn't unreasonable, as the first and last visits were at quite different times of year.

Two-way ANOVA: BirdsSeen versus Binocular, Site


Analysis of Variance for BirdsSee
Source        DF        SS        MS        F        P
Binocula       3      8.50      2.83     0.34    0.795
Site           7    328.00     46.86     5.67    0.001
Error         21    173.50      8.26
Total         31    510.00

                       Individual 95% CI
Binocula        Mean   --+---------+---------+---------+---------
C1              16.6            (-------------*-------------)
C2              16.4          (-------------*-------------)
C3              15.4   (--------------*-------------)
C4              15.6     (-------------*-------------)
                       --+---------+---------+---------+---------
                      13.5      15.0      16.5      18.0

                       Individual 95% CI
Site            Mean   -----+---------+---------+---------+------
1               15.0        (------*-------)
2               18.3                (-------*------)
3               14.5       (------*-------)
4               14.5       (------*-------)
5               15.8          (------*-------)
6               13.3    (------*-------)
7               23.5                             (-------*------)
8               13.3    (------*-------)
                       -----+---------+---------+---------+------
                         12.0      16.0      20.0      24.0

The column marked 'P' is an indication of whether there is any evidence that choice of Binocular, or choice of Site, makes a difference. P=0.795 shows that no evidence has been found that choice of binocular makes any difference whatsoever; I imagine that it does make some difference, but such a small one in comparion to all the other effects, including dumb luck, that we can't tell what it is. P=0.001 means that it is very unlikely that the effects that appear to be due to differences between Sites are just the result of chance.

The ascii art is an attempt to outline a range of values in a way that makes it likely (probability 95%) that we will end up including the true value in the range outlined. The binocular codes C1-C4 are again in order of magnification, and you can see from these drawings, or from the mean number of birds seen, that the B&L 7x42 Discoverers, whatever their optical shortcomings, in fact found me more birds than the others. However, we also know that the difference is so small as to be indistinguishable from the effects of chance; one token of this is the extent to which the lines for different binoculars overlap. My favourite Swift Audubon 8.5x44 820ED in fact came last, but again there is no evidence that this is anything except chance. I have just taken them to my latest RSPB outing, knowing full well the results here, and intend to continue using them as my binoculars of choice for pure birdwatching (I keep thinking that it might be a good idea to use the Optolyth Alpins next time I know we are going to be doing a lot of long-distance stuff, but I haven't actually persuaded myself to do this yet).

The spread and overlapping of the confidence intervals doesn't just show that no difference has been found between binoculars. It is also an indication that very large differences between binoculars can be ruled out. Unfortunately, you can't tell that much from individual 95% confidence intervals as given here, and there isn't enough data to really narrow things down, but we can probably rule out any difference of 6 birds or more between the typical scores for any single binocular and those of the others (based on 99.75% confidence intervals for the 4 exact tests between the score for a single binocular and the merged scores for the other 3 binoculars).

The lines for choice of site show that one site provided a lot more birds than the others. In fact, if I were to edit the data to remove this site from the figures, I would find that there was no proveable difference between the sites, either. This site is WWT Slimbridge.

The calculations behind the ANOVA result make some assumptions about the variability of the figures which are, strictly speaking, not true. In practice, it is known that ANOVA is likely to produce a very close approximation to the true situation in these sorts of circumstances. It is possible to phrase questions that can be answered exactly, by considering the possible allocations of binocular to visit produced by the dice. I have done this, and this also shows that there is no evidence that any binocular is any better than any other.

The Data

Here are the figures themselves. The numbers for any given day are the number of different species of birds seen. Most routes amounted to a couple of hours walking with binoculars. I only used hides in Slimbridge, and then I stopped off in them as I passed through, rather than waiting for any noticeable period of time. I chose a route that I happened to like walking through, rather than putting a lot of effort into searching for birds. In Coate Water this meant that I didn't visit the conservation areas, which are away from the main footpaths, and surrounded by signs that I didn't find particularly welcoming. I did stop periodically to scan with my binoculars.

In conclusion

My first reaction to this data was that I was actually telling everybody to spend their money not on binoculars, but on travel. There is a problem with that. Not only will you spend a lot of your time driving or hanging around when you could at least be reading bird books, but if everybody produces CO2, and generally pollutes the environment, travelling, there will be fewer birds left to see when they get there. Notice that the one site worth travelling to was WWT Slimbridge, where they have put in a lot of work making sure they have an area birds will be attracted to, and will do well in. Perhaps the conclusion of the survey is that the best way to see more birds is to work through conservation to ensure that there are more birds out there to see!