Thursday, July 16, 2009

Sew A Mile In Our Shoes

La Traviata is in it's final orchestra dress rehearsal tonight. Glimmerglass has a tradition of putting full sets, costumes, and lights on stage much earlier in the process than most opera or theatre companies would. The advantage of this is that the designers can continue to refine their work to a much higher level and designers and director can better take inspiration from each other's work and react to it. From a costume point of view it also gives us even more time to custom fit the costumes we make for the singers.

The disadvantage to this schedule for us in the shop is that we have about half the "standard" amount of time from the point at which we do a first mock-up fitting to the point when we need to put a wearable costume on stage. This year we did that in about a week and a half instead of the customary four weeks. So we find ourselves needing to make design and fit changes to garments we're still scrambling to finish. The challenge is to continue to work to Glimmerglass's high standards while making these deadlines.

One of the most time consuming costume details my team has created this summer is a skirt made of layers and layers of tulle, covered in tiers of tulle ruffles. (If you've forgotten, tulle is a soft net, most available in nylon.)

The numbers involved in this project (and remember they all translate to many hours of labor) are truly impressive. The inner layers which support the skirt include 8 layers of tulle, each of them 10 yards (30 feet) around. That took most of a 50 yard bolt of tulle. The ruffles are gathered onto a tulle base. In most fabrics a ruffle twice as long as the base to which it's sewn would look plenty full. Tulle has an amazing ability to gather, so a traditional 2:1 fullness looks less than skimpy. We ended up using a double layer of tulle for the ruffle, and gathering it 8:1 at the bottom, slightly less as the layers move up the skirt.

I cut the tiers by neatly wrapping the full yardage around a long cardboard card. Then I cut the correct tier depths through all the thicknesses with a rotary cutter. This probably saved hours of time over marking and cutting 13" from the bottom edge of 40 yards of net, then going back and doing it again for each of the rest of the 5 tiers.

The cardboard in the middle definitely made this process easier.

As mentioned, the tiers get fuller as they go down the skirt, which you can see by the varying thicknesses of the wound and cut net. The topmost ruffle is on the left, the bottom on the right.

At least a half dozen costume technicians worked to gather, pin, and sew the tiers down to their base.

The finished skirt includes 220 yards of ruffle, all cut, gathered, and assembled in just a few days. That's on top of the 80 yards of tulle that hold the skirt out in its full shape. The hem underneath the bottom ruffle is held out to a circumference of over 3 yards by all of that tulle. I'm quite pleased with the final results, and my fantastic team!

3 Steps Forward, 2 Steps Back

It's been very busy here at the opera, but I have had at least a little knitting time. I've finally finished spinning the purple yarn I've been making for as long as Rachel has known me (she so kindly pointed out this summer) and have started knitting it into a sweater set. I often make my own patterns, which certainly carries it's own satisfaction, especially for those of us picky about fit, but for this set I had planned to use a pattern I already own in a 1950 knitting pamphlet. I think I had intended to spin the yarn to match the gauge given in the pattern (not sure, as my intention dates to about 2006) but while all of the balls of yarn are fairly consistent* with each other, they are not, in fact, the right grist for this pattern. So I had to do a very small rewrite of the existing pattern to match my actual gauge. Still faster and easier than writing the whole pattern, right?

This is how far I got under that plan:

Knitting in the round affords lots of chances to try on a garment as it's being constructed, and correct errors well before the end. This was my try-0n point, the point at which I instantly regretted second guessing my original instinct to knit 5" of ribbing. The pattern (authentic 1950 you recall) only called for 3" of ribbing, and although I knew better, I deferred to the pattern at 3" because I was sick of knitting 1x1 rib. Also, the body increases were too close together and the body itself a bit looser than I had hoped for (or indeed than they appeared on the model).

So instead of keeping 8" of the wrong thing, finishing it and hating it forever, I ripped out 5" of body back to the original 3" of ribbing.

Discouraging to do, but from this point I knit the additional 2" of ribbing I should have started with, and reworked the body shape and size to something much more flattering. At this point I've rewritten the pattern extensively, almost equivalent to just writing my own in the first place. Almost. I'm still trying to mimic the look and shaping of the original, but fit my non-1950-model-body.

* As a point of fact, I am incredibly proud of how consistent my grist is on 2 1/2 pounds of fine yarn spun here and there over a 3 year period.