Test photograph showing studio.

Sometime ago I met a lady who had an idea for a computerised fashion design business. She envisaged that she would have a number of standard patterns on her computer, and that each customer would choose one of these patterns. Then the operator would use a digital camera to photograph the customer from several angles and run an automated program to analyse the photographs, compute her vital statistics from them, adapt her chosen standard design to fit her individual figure, and finally print out a personalised pattern for the seamstress.

I had already been thinking about how I could photograph my body to see how well my underwear fitted, and how much effect it had, and I did some calculations to evaluate various possible arrangements.

I have been thinking about this on and off ever since, and recently I decided to mock up a photographic studio in a bungalow in the back garden, which had been my office when I started a business many years ago. I bought two large mirrors, which I attached to old doors, and arranged them in a Vee, with the centre line pointing at the far corner of the room, where the camera was mounted on a tripod. A smaller mirror behind the tripod enabled me to see that I was standing in the correct position, and a foot pedal and remote release allowed me to operate the camera. To take a photo I stand between the mirrors, side on to the camera, and look into the mirror in front of me. This, in conjunction with the mirror behind the camera, enables me to simultaneously see my front and side views, so I can stand in the correct position. The camera simultaneously photographs my side view directly, and my front and back views in the two mirrors.

The arrangement is shown in the above test photo. The two mirrors are on either side of the yellow strip, and each reflects the other one, so that you can see two images of me standing at the camera. As many of my costumes have a relatively low contrast, I cannot rely on the the camera to focus automatically, so I put the camera on manual focus, and focus on this wastepaper basket. The fine detail with relatively high contrast makes it easy to tell when the camera is focused.

The camera does not provide an accurate readout of the exact setting of the zoom lens, but the two horizontal black lines enable me to calibrate each set of photos.

When I tested the system, everything worked as planned, but the background was exceedingly messy, as the three mirrors produced an infinite series of reflections around the room. I managed to reduce this somewhat by arranging screens strategically, but the background is still too cluttered to permit the pictures to be analysed automatically, and even to make the photos attractive to look at I have to tediously remove the background in Photoshop.

The original proposal would only have used two mirrors, and this would have made it much easier to control the background, as it would not have produced an endless series of reflections.

Playtex ‘I Can’t Believe it’s a Girdle’

Purchased in 1974

When I started trying to compare photos, I found that minor variations in the way I was standing produced larger changes in the shape of my body than the garments I was evaluating. I also found that my hands and arms posed a considerable problem. If I just let them hang naturally, they tend to obscure the outline of the body in both the front and side views, so that I cannot make measurements, and if I try to hold them straight, and a bit away from the body, I get a rather unnatural pose, and additional variation in the shape of the body.

I also need a dresser. This photo reveals that I have difficulty with the back suspenders, as I am sure you will all understand. I am also hanging over the waistband at the back, but to do anything about this I have to raise my elbows so that I can get at my back, which does away with the problem, only to have it reappear immediately I resume my normal position.

Despite all these problems, I have got quite a number of interesting photos, which illustrate my collection of girdles, and will enable me to do some comparative studies. Some of these will appear here in the fullness of time.

The original purpose of this exercise was to determine whether a photographic method like this could provide accurate measurements of a model without having to resort to a tape measure. Analysis of my photographs indicates that an accuracy of about 5% can be obtained. This corresponds to about 1” in a waist measurement of 30”. There are four main factors contributing to this error.

  1. For the purposes of analysis I have assumed that the body has an elliptical cross-section, but in practice it is a rather irregular shape. With the technique used it is only possible to measure the width of the body from side to side, and from back-to-back, at the widest points.

  2. The shape of the body varies considerably with minor changes in posture, and probably from time to time. I did not make any measurements when I took the photos, so I am using measurements taken at a later date to calibrate them. This may introduce significant errors.

  3. It is actually quite difficult to decide what is the smallest waist measurement, or the largest hip measurement, because of the irregular shape of the body and the fact that the narrowest (or widest) dimensions in the front view do not correspond to the same dimensions in the side view. This makes it hard to tell just what is the largest or smallest measurement, both when you are measuring someone with a tape measure, and when you are taking measurements from the photos.

  4. Because of the limited size of the studio the camera is relatively close to the model, and parts of the body closer to the camera will appear to be larger than those further away. In my setup the camera is 4700mm from the model, and an error of 100 millimetres in the position of the model will cause a 2% error in the measurements. In principle I could determine the exact position of the body from the photos, and use this to correct for this error, but I have not attempted to do this.

The girdle in this example is a Playtex ‘I Can’t Believe it’s a Girdle’, from 1974. It is worn over a Triumph ‘Something Else’ brief. These early models had a rigid panel of almost transparent net in front, with a window in the elastic, as can be seen. Sadly the window was omitted shortly after this. I have always felt that this was a particularly attractive design. This garment is one of the first girdles I bought, before I had worked out my size, and it is a size small, so it is very effective, but I cannot wear it for any length of time. I wore the cuff-top version for several years, but I never found it really comfortable. The stockings are Kayser ‘Intensive Support’ stockings, which were discontinued about five years ago.

The ‘Tweave’ fabric used in these girdles is of excellent quality. I have several garments of similar age, and all are in perfect condition. I also tried on a 20 year old Playtex‘Fits Beautifully’, which is a similar but cheaper line, but the photo showed bulges all over my body, where the elastic had perished.

 

Home