I’ve had a few requests about the rest of the Baltic pickup style patterns I came up with. Like the previous, these are all woven on the same warp, and again, the patterns are designed so that you are always pushing threads down from the upper row and never lifting threads up from the lower row. This speeds up the weaving immensely.
So a friend of mine asked me about the trim in https://www.pinterest.com/pin/120400990016540518/, which is a picture (source unknown, because pinterest) of the neckline of a tunic that has been embroidered and embellished by trim.
Looking at the trim, you can see that the outermost three cards on each edge have been threaded with four threads of the same colour; 2 cards are tan next to the pattern, the outermost card is all black. Now when I first saw the picture I couldn’t see it up-close, so I didn’t notice that the inner pattern was actually three colours (each card has 2 white, 1 black and 1 dark blue thread), but the pattern is obviously based on the Egyptian Diagonals technique using 10 cards.
My first stab at the pattern; the 10 cards are each threaded identically with 2 dark and 2 light, and then arranged so that if all the cards were turned forwards you would get chevrons pointing up. You can see the starting position below the squares: the two circles represent the colours of the threads in the two top holes; bottom row is closest to the weaver, top row is furthest from the weaver, and the | or / indicates the direction the string travels through the card.
The pattern is read bottom to top, where each square is one column, each row of squares = 1 turn of the pattern. When the square’s background is grey, the card turn backwards towards the weaver; when it’s white the card turns forwards away from the weaver. As you can see, most of the time the cards turn together, but there is a central part of the pattern where half the cards turn in one direction and the other half turn in the opposite direction. It is because of these central 4 rows that the pattern is not twist neutral which is why – if using an inkle loom or similar to weave the trim – a weaver will often alternate between the two versions of the trim, since weaving one copy of each will result in a twist neutral pattern.
Note that the bottom and top 4 rows are not included in the repeat.
Still, looking at the picture, this pattern made the crossover point of the black X’s awfully narrow, and looking more closely at the picture I could see the crossover point should be wider. Which is an easy fix: by making the repeating portion of the pattern one row shorter top and bottom, the X’s crossover point becomes wider. It also fixes the width of the black dots at the sides, which was previously too wide.
(And here the top and bottom 5 rows are not part of the repeat.)
So I spent some time this weekend playing with the Baltic inkle pickup warp. The rules of the game were simple: two repeats of the simple diamond with the 4 interior diamonds, alternating with a diamond with some other free-styled pattern.
By keeping the edging pattern between the diamonds and the outer diamond’s dimensions consistent, the band feels cohesive, even though every third diamond is something else. Below are 16 possible variations I came up with, while enjoying a Columbo marathon.
Oh, and just one more thing … there are 25 places where the warp thread can be pushed down, or not, inside the diamond. That means that there are 2^25= 33554432 possible combinations, but I’m guessing that many of them lack pleasing symmetries.
Back of the envelop suggests that requiring symmetry along one diagonal reduces the possibilities to about 2^10+2^5=1024+32=1056 (the places below the diagonal line plus the diagonal line), but then the diagonal line can itself be flipped … But the “S” figures are flipped around a vertical or horizontal line; so there are 4 possible axis of symmetry, so 4*1056= 4224 possible combinations that are symmetric.
In short, back of the envelop suggests that I’m going to run out of warp before I run out of possibilities.
Baltic pickup style inkle is woven using background threads and pickup threads; the pickup threads are usually either thicker, or doubled up. In the threading diagram, you can see how each pickup thread is place between two threads in the opposite heddle; the white boxes representing the pickup threads. In the diagram one row represents the heddled threads and the other the open threads, and the rows aren’t labeled since it actually doesn’t matter which is which in this case, so why complicate things?
The patterns here use 17 background threads; the 5 leftmost and rightmost threads are woven over a green background, while the middle 7 pick up threads lie nestled between dark blue background threads. The plan is to eventually experiment with the Kostrup patterns, since those are 17 cards wide also, but I needed simple patterns to weave while out and about Monday.
The current patterns are simple diamonds, and by weaving them upside down I only ever have to drop threads from the current row, rather than having to pick up. In the pattern (Baltic Diamonds) squares with circles in them are the pickup threads that are high in each row. If a pickup thread is high but has a grey background, then it gets dropped, resulting in the front patterns we see on the left below.
This is what the pattern with the small diamonds looks like; the predominantly green/blue side is what you see while weaving, the white side is the underside which is much harder to photograph while on the loom.
Here I was working with larger diamonds, and experimenting a little with alternate patterns like the diamond in a diamond and the inner cross. Making the outside diamond a little smaller results in a 2×2 grid of the smaller diamonds, or a 3×3 grid of the larger, and then it’s just a matter of dropping a few extra threads here and there.
I draft a lot of patterns for tabletweaving, and the way I like to display the pattern includes both the turning directions and a guideline as to what the pattern should look like.
When I first started drafting patterns, I would work in pencil on gridded paper where each square represented one card in one row; first I would outline the main shapes, and then use a highlighter to add the turning directions for each card — highlighted for backwards, not highlighted for forwards. The biggest downside was that even at 5 squares per inch, a sheet of standard letter paper didn’t really hold room for a lot of cards. And heck, the squares could be smaller in any case.
I use computers a great deal, so casting around for solutions I experimented (briefly) with spreadsheets and discovered that this was too fiddly for me. Then I discovered Adobe Illustrator, which is a vector based graphics editor that has all the bells and whistles one could desire. Sadly, it also had a price tag to match, and I was at the time a graduate student, and student discounts are nice, but not *that* nice.
Which brought me to inkscape which is free, opensource, and available for Windows, Mac OS, and most importantly linux, aka my operating system of choice. Inkscape is a very powerful program, but I take advantage of a limited number of features.
- Grid lines – black grid lines 0.5 px wide
- White grid lines – white grid lines 3.5 px thick
- Pattern – one layer per colour to make colours easy to manipulate
- Colour 1 – line 3 px wide
- Colour 2 – line 3 px wide
- Rough Pattern
- Threading – the colour in the dots indicates the colour of thread in the two top holes.
- Greyscale – used to indicate the turning directions forward and back. Grey squares are turned towards the weaver, and white squares means the cards are turned away from the weaver.
Setting the layers up in this way allows me to take advantage of the order in which they’re drawn (bottom up) so that the diagonal lines I draw in the pattern colours become small individual diagonals.
In this case, I’m planning a very simple pattern where the cards turn continuously in one direction until twist builds up, and then they reverse direction. I’ve roughed it in and now it’s just a matter of adding the colours.
First off, I’ve been finding interesting articles all week, and collecting them, so without further adieu, a few links:
- Edyth Miller’s “Pattern Darning Tutorial” – while pattern darning is often assumed to be more modern (i.e. 18th/19th century) it can be dated back to ther 12th century. Here Edyth Miller provides a clear tutorial for the technique.
- An article on the origins of most popular tablet weaving patterns. – This article is posted on deviantart, but includes lovely examples and good references.
- Alfrun Ketta’s ” Viking Textiles – A deeper look at plaids, stripes and checks “ – Alfrun Ketta addresses the question of whether Vikings (and viking reenactors) can wear plaids. The answer is yes … ish. She includes a few examples of plaids that would or would not be appropriate and explains her reasons why.
Secondly, I’ve been (unsuccessfully) trying to design knotwork in the Snartemo 4 colour technique, but in the process ended up digging out a lot of my old 3/1 broken twill notes. So here a very brief explanation of how I weave 3/1 broken twill.
In 3/1 broken twill, patterns will have a light and a dark colour. Each of the pattern tablets is threaded with two light and two dark colours, just as one does for doubleface. But then, using Collingwood’s 2 pack method, the cards are divided into two groups: the odd numbered cards form one group, and the even numbered cards form the other. Each group turns as a whole; colour changes are created by flipping a card around its vertical axis – which reverse the threading – rather than by changing turning direction.
Setting up the two groups:
- In each group flip the cards as needed so that they alternate S and Z threading.
- In the odd group, turn the cards so that the background colour goes through the two holes closes to the weaver. This is the “vertical” position, because the two holes threaded with the background colour are one above the other (aka vertical if the warp is held horizontal).
- In the even group, turn the cards so that the background colour goes through the two top holes. This is the “horizontal” position, because the two holes threaded with the background colour are beside each other (aka horizontal if the warp is held horizontal)
The cards are turned in a few turn repeat. Each of the two groups moves 2F/2B, but there is an offset because of the differing starting positions. Repeating this turning should yield a solid-coloured band with a structural diagonal.
|Turn||Odd Group||Even Group|
I draw my patterns on a rectangular grid with a brick like pattern; Tree with Birds Pattern is a pattern that uses 32 cards, but because of the symmetry the setup is modified slightly so that the second set of 16 cards completely mirrors the setup of the first 16 cards
In the pattern, each column represents one card over all the many turns; each rectangle represents that card for two turns. Note how the odd and even rectangles are offset, because of the offset turning directions. In the right side of the pattern you can see the changes needed for the colour changes; if the next rectangle switches from white to grey or vice versa, flip the card.
But that would give you some ugly edges, so sometimes we also need to flip the card when the colour is horizontal rather than vertical. Those flips are represented in the left half of the pattern by the blue < in a rectangle.
I tend to spend a lot of time fiddling out new patterns, most often for tablet weaving, but also for brickstitch and occasionally blackwork. In fact, I’ve got blackwork on the brain at the moment after Kim Salazar posted Blackwork Inspiration, an article in which she talks about places to find inspiration.
I realized as I was reading her article that I like designing patterns in tightly constrained spaces. In tabletweaving, especially when designing patterns designed to be warped up quickly using speed warping, the design is primarily affected by how the tablets turn. You have essentially three options here; a tablet can turn forwards, backwards, or it can stay idle. I haven’t really started experimenting with tablet idling yet, so that reduces it to two turns, which can easily be plenty.
Consider for instance an early pattern I called Daggers. Back when I was graphing this I was using shorthand; each square represents two tablets and two quarter turns. The colour of the thread is actually indicated by whether the line is solid (colour 1) or dashed (colour 2), and the background colour (colour 3) isn’t shown at all; forward turns are blue and backwards turns are red.
The tablets are warped so that each tablet has, in clockwise direction, colours 1 – 3 – 2 – 3 (although often when I wove it, the background colour and the second colour where the same colour, making the warping 1 – 2 – 2 – 2). This allows for quick speed warping using the continuous warp method Linda Hendrickson describes in her description of a Continuous Warp (Linda Hendrickson). After the tablets are all warped, they are set up so that all the cards are threaded in the same direction, and then turned so that colour 1 is – moving from left to right – in the top front hole, top back, bottom back, bottom front, and repeat. Turning all the cards forward should result in straight diagonals moving to the left. If the diagonals are jagged, either flip all the cards vertically and reset, or turn every second card two quarter turns. This is all very similar to setting up for Egyptian Diagonals, except that the tablets are not all threaded two dark, two light.
In this version, which is a slightly different pattern where the handle is longer and the guards are much bigger and the background is made with solid coloured diamonds, I was weaving a pink belt; hence the background colour is pink, colour 1 is purple, and colour 2 is white. And I just noticed I’d left out the dashed lines inside the daggers, so will leave those as an exercise for the reader.
Patterns like this are easy to develop on grid paper; the basic rule for me is to draw out colour 1 first with the following constraints: if the diagonal changes direction, there can be no gap and for the two cards we must go from one diagonal straight to the other. If the diagonal does not change direction, then there must be a (temporarily) blank square between them. I can test the design by adding the dotted lines using the same rule. If there are two parallel diagonal lines there must be a dotted line between them. And the backgrounds I usually use either the dotted diamonds, or solid diamonds that are properly offset from the main pattern (so that there is room for the dashed lines in between).
Over time I’ve started developing diagrams that use the background (white or gray) to indicate the direction of the turn, and show every thread rather than just the dominant colours, but I find that when I’m just doodling patterns this method suits me just fine.