Encouraging Canvaswork classes

It’s been a while since I last wrote about my Canvaswork module for the RSN Certificate, and that is partly because I am still finding my feet, even after three classes. I think my comfort zone isn’t even dimly visible on the horizon most of the time. Still, the two classes I’ve had since that previous FoF have been encouraging – the tutors appear to have faith in me even if I haven’t smiley, which makes me feel a bit better about the whole enterprise. So what have I been doing since that frantic surge of last-minute sampling back in January?

Well, I attended a class with Helen McCook, who okayed several of my samples, such as the roof and the Turkey rug bushes; she advised stitching several samples and cutting them to different pile heights to see what would work best, keeping the bushy look but without them becoming too prominent, as they are meant to be on the horizon and should therefore recede into the background rather than push themselves forward. I did this at the next class, where Angela Bishop suggested I blend other threads besides wool into the mix, so I tried it with an anonymous green thread from my stash (blue arrow) as well as some of the vintage green silk I inherited from my mother-in-law (orange arrow); neither were very noticeable once trimmed, so I will have to try another one with more strands of silk mixed in.

Sampling various bushes The trimmed bushes don't show the non-wool threads well

The main point I took away from Helen McCook’s class was the fact that my idea of what the sky’s stitch blending should be was incorrect. I’d started sampling the stitch blending going horizontally, but it should in fact be done vertically. As Helen pointed out, in canvaswork the smallest stitches should be at the vanishing point – in my case the horizon. From there they go larger the closer you get but (and it took me a while to get my head around this) this goes not just for the foreground, it also applies to the sky! So the stitch blending in the sky should be larger stitches at the top merging into smaller at the bottom. She also felt that although the Parisian (small) and Hungarian Grounding (medium) in my sample worked well together, Victorian Step (large) had too much of a diagonal component to blend in. So I started looking for suitable large stitches with a more horizontal look, and found one called Water, which is basically random-length satin stitches. Because I wasn’t sure that that would provide enough contrast (they have to blend, but I don’t want them to be too similar) I found some others and stitched a sample of each, which I could later use to blend into the other two stitches.

Possible large sky stitches

The stitches were, from left to right, horizontal Milanese, Willow, Pavilion and Water. Willow immediately revealed itself as a non-starter – too blocky – but I continued with the other three to see how they would blend into Hungarian Grounding (and that in turn into Parisian), trying to make the transitions gradual so that there wasn’t a clear horizontal break between one and the other. I showed these to Angela at my third class and we agreed that the pattern in Pavilion was too strong. Water blends in beautifully and looks least stylised, but I thought it was all a bit samey, and Angela worried that it would also echo the stitches in the paving too much (of which more later), so I will most likely go with the Milanese version. It has some patterning but the lines move forward (albeit in a zigzag) rather than turning back on themselves (like the diamonds in Pavilion).

Sampling the sky transitions

One thing I found in my two classes was that different tutors have different approaches and ideas. Helen advocated creating different “stitch languages” (so that, say, flowers are done in one set of thread types and stitches while leaves are done in other threads and stitches, which don’t overlap) while Angela at one point suggested that when using vertical Parisian for the smaller tulips near the paved area I could keep the same stitch but change to green to just make a colourful jumble of tulips and leaves. It’s a bit confusing when they do that…

Talking of tulips, I got on with those as well, the two big ones that stand out in the foreground. The pink tulip has so far been sampled in soft cotton but in the near future I must try blending the various shades of pink Carrie’s Creation overdyed stranded cotton which I’ve picked for that (so far I’ve only worked out how many strands are needed for full coverage – six, separated and recombined and used with a laying tool). I started out on the red tulip using soft cotton as well, but have since tried out threads that would work on the actual piece, and the main colour is going to be a lovely orangy red Caron Watercolours called Bittersweet. The pink tulip is going to be done in modified Florentine/Bargello, so I had to sample some standard Florentine as the assessors need to see you can do that too; and Helen suggested doing some of the petals in angled Florentine to make them more distinct from each other. When doing similar length stitches at an angle you can use whatever slant you like, but I found that with stitches of different lengths it’s easier to stick with 45 degrees, so that’s what will happen in the final version.

Sampling the red tulip Standard Florentine stitch Sampling the pink tulip Six strands of cotton give good coverage

Another prominent shape in the foreground is a large bud on the left, which is mostly green but with a hint of red in it. I sampled this before my third class in Cashmere stitch using blended perle #8, and I really liked the look of it. The shading isn’t in the right place yet, as I had only two shades of the yellowy green (I have since bought two more…), and the single strand of red applied over the top (which Angela suggested I try, to see if it would work) looks a bit messy so I will try and blend that in while stitching the greens, but on the whole it’s probably the part I’m happiest with so far!

Sampling a Cashmere bud Adding a hint of red

Then there was the paved area. One of the things the brief requires is at least two each (and Helen suggested picking three to be on the safe side) of four stitch types: horizontal, vertical, diagonal and crossed/textured. I’d intended the paving to be horizontal, but as I didn’t have many diagonal stitches yet Helen suggested using one there – she pointed out that “diagonal” includes anything slanted, so it could be as near horizontal as possible and still count as diagonal. I sampled various slants in Oblique Slav, settled on a 1 in 5 incline, then tried it in a linen thread I had lying around and didn’t like it. I then sampled it in flower thread or blomstergarn, coton à broder, and floche. Flower thread, with its unmercerised matte appearance, was the clear winner, and led to my acquiring the Danish Handcraft Guild‘s complete set. I have picked six shades that should work together well, covering the paved area’s brick-like colour as well as the rather surprising grey and almost-white in some areas.

Sampling oblique Slave Oblique Slav in linen Oblique Slav in different threads

Which bring me to one of my main stumbling blocks in this module – colour blending. In other people’s projects I noticed that in some areas the blend of, say, six threads might change composition (e.g. from 3 dark, 2 medium, 1 light to 2 dark, 2 medium, 2 light) every three or four stitches. I asked Angela how on earth you got into a stitch rhythm changing blends so often. “You don’t,” came the reply. Her tip was to load up ten or twelve needles with the various blends so at least you didn’t have to keep stopping and re-threading all the time, as there is the additional snag of incredibly high thread usage in this technique. Oh joy. So far I haven’t dared to think in any detail of the blended sky yet, but I have sampled some fuzzy threads (Madeiral Lana and Rainbow Gallery Wisper) to represent the white haze near the horizon, in one case blended with the silk used for the sky. I need to play a bit more with the proportions to get it looking quite right.

Madeira Lana and Wisper for the sky Blended haze

Otherwise, I have been sampling, sampling, and then sampling some more. Angela has told me to bite the bullet and actually put some stitches onto my proper canvas, but so far it hasn’t happened. The sampling has given me lots of ideas for things to use though (and some for things definitely not to use). Here are a few that made the Useful list: woven plait, fern stitch, slanted gobelin (encroaching, plain and split), brick stitch (and the decorative but less useful herringbone snowflake), upright cross & alternating continental (with a rather messy vault stitch), raised spot (three ways) & vertical Parisian, upright double cross & spot stitch (with some other odds and ends), kalem & lazy kalem (with another fern stitch).

Woven plait Fern stitch Slanted gobelin (encroaching, plain and split) Brick stitch
Upright cross and alternating continental Raised spot and vertical Parisian Upright double cross and spot stitch Kalem and lazy kalem

One idea of Angela’s that I haven’t sampled yet is to use Upright Double Cross for the complex areas where the blue sky shines through the leaves on the tree – work the upright cross underneath in blue silk like the sky with the diagonal cross in green wool (or whatever I will be using for the tree) over the top. I really like that idea, and it is yet another example of the way colours can change even within stitches. One day that idea may become a natural one, for now it definitely has to be suggested to me before I see it.

Where tree and sky mingle

Since the third class I’ve done only two more bits of sampling, a fringed pale lilac and yellow tulip for which I want to use some ribbon inherited from my mother-in-law (I haven’t quite got the look I want yet), and the wooden parts of the windmill’s sails, worked in ribbon over tent stitch. I like the way that’s come out, especially in the loose-lying version (on the right), but the one held on with stab stitches would be more secure. Still, no rush for that decision. I’m taking this module at a very sedate pace.

Using a variegated ribbon for a fringed tulip Two possible tulips Tent stitch base for the sail Ribbon windmill sails

And finally a little update on Bruce – in response to my questions Anne Butcher, Head of Teaching, wrote a detailed reply with helpful comments, and although I specifically did not ask for a reassessment she said I should not have lost points for using S-ing, so my final score is 89% (tantalisingly and ever so slightly annoyingly 1% short of a Distinction). Upgraded Bruce & Haasje have since been framed and are waiting to be put on the wall, so I can be proud of them every day smiley.

Bruce and Haasje framed

Getting into canvaswork – slowly

After my first Canvaswork class back in November I was feeling more optimistic about the endeavour, but I’m afraid my doubts returned as I was getting closer to my second class. To say that I am outside my comfort zone with this module is putting it mildly, and whether or not that was the reason I didn’t do much homework, the fact is that with less than one week to go I was feeling woefully underprepared. I managed to change my booking from 22nd to 29th January and solemnly undertook to do some serious sampling in that extra week.

By the way, just so that you don’t think I’d been completely useless over Christmas – I did do some work. I sampled some herringbone stitch in two different ways to see if I could work out a method which didn’t involve coming up underneath previous stitches, but which would still look the same (I could, the only visible difference being at the back of the work; the blue arrow points to the “by the book” version, the green arrow to my alternative). I also made a start on my colour plan, but as I wasn’t absolutely sure what the colour plan was meant to look like and how shaded the shading needed to be, I abandoned that halfway through. And I also… er, no. That was it.

Two versions of herringbone Different stitch patterns on the back An abandoned colour plan

In order not to feel immediately overwhelmed I started my new regime by sampling a very small roof. From the start I had envisaged it in slanted buttonhole stitch, and as it was orange and I have oodles of orange wool left from the Jacobean module, I used that. First I transferred the exact outline to my sample canvas, so I could see if the wool gave enough coverage. It did, and I really like the effect of the stitch. I later sampled it again in a slightly darker orange – I’ll decide further down the line which one to go with. (The picture also shows some outlines of bushes for later samplings.)

A roof outline Sampled in one orange Two orange roofs

Next was the sky. There were two things to decide there: threads and stitch transitions. Because the sky is what we call in Dutch “strakblauw” (literally “taut” or “stretched” blue) I wanted a thread with a smooth texture, not rough or matt but not overly shiny either. And it so happens that in my stash there are two rather lovely series of blue silks, one in Caron Soie Cristale and one in Soie Alger. But there was possibly a snag – when the surface has to be fully covered, canvaswork takes a lot of thread (demonstration to follow), and I wasn’t sure whether my stash of silk would be enough. Buying more, especially of the Soie Cristale, would be difficult and would in any case mean different dye lots. But I do have a lot of DMC perle #8, and some of the blues looked quite suitable. They are more textured than silk, and a bit shinier (the silks I picked are both spun silks), but it was worth a sample.

Possible perles for the sky

Did I mention canvaswork takes a lot of thread? In any other embroidery that I do, perle #8 would count as a relatively chunky thread. Here it took a triple thread to get sufficient coverage. As for the texture, I wasn’t convinced, but I was going to wait and see how the silk behaved before deciding.

Perle sampling

Soie Cristale is a 12-stranded silk, and I started out sampling with six strands. I got to use my beautiful laying tool, as getting the strands to lie parallel not only looks nicer, but also spreads the thread more. Even then six strands didn’t quite cover the canvas, so I tried again with nine, and worked a larger area to see the effect. Then I did a similar area in blended perle #8 to compare. I can tell you now that there is simply no comparison – silk it has to be! So smooth, so pretty, so lovely to work with *swoon* … Soie Cristale is on the Definite list.

Smoothly stitching silk with a laying tool Silk versus perle Silk versus perle close-up

One of the required items in the brief is a stitch transition, where at least two different stitches (and bear in mind that in canvaswork a “stitch” is often a particular arrangement of several stitches) have to gradually blend into each other. This means they must have some similarities to begin with, otherwise the join will be far too visible. I decided on Parisian stitch (underlined in blue), Hungarian Grounding (green) and Victorian Step (red). In spite of the interesting names, these are all arrangements of parallel straight stitches. The first two are usually worked with the stitches running vertically, so I turned them 90 degrees as I think the sky will look better with a horizontal sweep. It took a bit of pencil-and-squared-paper work to get the second transition to blend, but in the end I had something that I could present to the tutor as a feasible option.

Sampling transitions

By the way, a lot of sampling is done in just any old thread, unless you are actually trying out coverage or the way colours work together; the above transition was done using a spare ball of perle #5, and gave me the additional information that even that thickness does not cover the canvas when used horizontally (it may work diagonally, where the lines are closer together).

What else did I do before class? Ah yes, Turkey rug stitch. I want to use it for some bushes, with blended threads. I first tried it out with two (not very blendy) shades of Appleton’s, only to find that I’d forgotten how to do Turkey rug (which I first used on the bodies of two stumpwork butterflies) and was making the securing stitch too long (green arrow; the red arrow shows the correct length). Having refreshed my memory on this count, I decided to try it out in the proper threads, a blend of three shades of green Heathway Milano wool (the blend to change as a darker or lighter look is needed). As it turns out, a triple thread may be a bit too thick – after a few rows it gets difficult to see the canvas for the next one – so I’ll try one of the bushes with a double thread. I haven’t cut the loops yet, that’s another thing on the To Do list.

Turkey rug with securing stitches that are too long Turkey rug in the right thread and colours

Finally, I transferred the two big tulips to the sample canvas and made a start on the red one. One possible stitch for this tulip is web stitch, which has diagonal stitches couched down to create a woven look. My thought is to leave out some of the rather dense couching and use the remaining couching stitches to create shading, partly by working them in a different colour from the diagonals, and partly by spacing them further apart when the shading needs to be lighter. Because I’d been working on Bartram the Bayeux ram, almost without thinking I stitched the diagonals as laid work (bringing the needle up right next to where you’ve taken it down so that there is hardly any thread at the back of the work) rather than satin stitch – but the Canvaswork brief, in its Tips section, advises stitchers to always take the thread the longest way round to help with tension. However, that really takes an awful lot of thread, and with all the couching stitches would make the back quite bulky. I put it down as something to ask the tutor.

Starting on web stitch

And that’s the point I’d got to when it was time to gather all my frames and hoops and bits and bobs to go to class! But more about that in another FoF.

Can we canvas? Yes we can!

Until recently I didn’t really “feel” Canvaswork, so I approached my first proper class (which initially had been planned for last July, but got postponed several times for various reasons) with some trepidation. I came armed with two outlines which I knew to be far too detailed, a framed-up canvas which I knew wasn’t tight enough (but which by this time did at least have the required rectangular running stitch outline in sewing thread), a few samplings in the wrong sort of thread, and about one idea. I did not feel confident.

Two detailed tracings Framed up, but not quite tightly enough (and as yet minus outline) Possible stitches Some sampling

The tutor assigned to this class was Angela, and I’d been looking forward to seeing her and perhaps having a little Bruce chat with her, but unfortunately she had gone down with Covid (apparently feeling rather rough with it, poor her) and so the class was taken by Helen Jones. With only four students we each had plenty of time to discuss things with her, and for me the first thing was indeed to get that canvas tightened. I unlaced part of it, turned the bottom roller once and re-laced. It is now most definitely taut as a drum, but as that is difficult to photograph you’ll have to take my word for it!

The next thing was to simplify the outline. I was surprised at how far you take this process in canvaswork, and I fear mine probably still has too much detail (especially in the windmill) but this was as simple as I felt comfortable with, and Helen OK-ed it. To make it easier to transfer she suggested tracing the pencil lines in marker pen; this was also a good opportunity to get the horizon level. In the photograph the furthest edge of the paved area which forms the strongest horizontal line in the piece is actually slightly curved, but making it perfectly level would help to “anchor” the design when transferring it – if the horizon didn’t follow a straight line of holes on the canvas, I’d know I had to reposition it.

Simplifying the outline Tracing the outline and levelling the horizon

Having got used to prick & pounce and paint for transferring the design at RSN classes, canvaswork is a bit of a wayward module. There is no way the canvas would take the pounce in any meaningful way, and as you have to transfer the design when the canvas is on the frame you can’t just bung it onto a light box either. Instead, you build a squat tower of books with the design on top of it, place the frame over it so that the canvas rests on the design, and then trace what you can see of it through the holes with a permanent marker. It then becomes abundantly clear why the outline has to be simplified so much: the canvas simply will not take any great level of detail. It is also surprisingly difficult to manipulate the traced design if its position is slightly off, sandwiched as it is between the books and the frame. But eventually I got that nice straight horizon to line up with a row of holes, and drew it on.

Propping up the frame The horizon is in!

I can’t guarantee that what eventually ended up on the canvas is exactly like the design outline – some of the squigglier lines were difficult to trace precisely – but again it got the OK so perhaps I was being a bit too fussy. What definitely did need addressing was the fact that I managed to leave off an entire hedge, which I didn’t notice until I got home and showed the canvas to Mr Figworthy! It has since been added in.

Outline minus hedge Outline with hedge

Because it felt silly to do absolutely no stitching at all in class, I did do a tiny bit of sampling: it’s a herringbone variation which takes shading rather well, and which I hope to use to bring texture to the green bits that aren’t worked individually. It is rather fiddly, as you have to bring the needle up underneath previous stitches half the time, but I think it will be worth the effort.

Herringbone variations sampled

My next class is in January; until then I’ll be colouring in a print of the outline (officially “making a colour and shading plan”), choosing stitches and doing a lot of sampling. I’ve got some ideas for the two large tulips in the foreground and various other bits and have sketched and scribbled a few ideas (yes, my handwriting is atrocious) to be translated into sampling at some point.

A few sketches

Due to canvaswork being the Mary Mary Quite Contrary of embroidery, those two big tulips will be worked first. In all other techniques you work the background first, and then the things that are a bit nearer to the viewer, and so on, until you reach the things in the foreground. If parts of the design overlap you stitch the overlapping bit last, which looks more natural and convincing. But in canvaswork you stitch the foreground first, and end with what is furthest away in the picture. As far as I understand, this is because the further back in the design you go, the smaller the stitches get – and it is much easier to work small stitches around large ones than fit large ones into a background made up of small stitches!

Having to end with lots of green and a big expanse of sky after doing all the interesting foreground bits may sound like starting with the fireworks and going downhill from there, but I rather like it – I think those tulips will entice me into a technique which is entirely new to me and feels unfamiliar and challenging. Let’s hear it for the Tempting Tulips!

Planning Canvaswork

After Bruce (yes, there will be an “after Bruce”, and not too far in the future with any luck) the next module for my RSN Certificate is Canvaswork. Surprisingly, even though I haven’t officially started, it’s already been through quite a few ups and downs. First of all I didn’t think I’d do it at all; I just wanted to do Jacobean and Goldwork (and even the Jacobean was mainly because they wouldn’t let me do Goldwork without it). Then I got the various RSN stitch guides and thought it might be rather fun to do Canvaswork after all, especially as some ancient seaside scribbles led to ideas for a possible design.

Early scribbles for a sea shore idea... ...and how it might look in a Canvaswork project

I liked it. I still do. It may one day make it to canvas. But unfortunately it doesn’t fit the Canvaswork brief, which specifies “depth and perspective”. A pity, because not only did I have some photographs from a visit to an aquarium in Brittany some years ago with weird and wonderful and usable creatures, but I’d also had a bit of a splurge on seaside-y textured threads. Fortunately some of these have since come in handy in other projects, for example in Septimus the Septopus.

Rainbow Gallery threads from eBay Rainbow Gallery threads plus one other from West End Embroidery Finished and lit from the side

Keeping the need for perspective in mind, I was tempted by a picture I took at Buckler’s Hard in the New Forest of some oystercatchers foraging. But in order to make them big enough to be both recognisable and stitchable I’d have to zoom in so much that most of the background, and with it any perspective, was lost. Exit the oystercatchers.

The oystercatchers at Buckler's Hard don't quite make the grade

Now from the first two modules you may have gathered that I like having some personal touches in my Certificate projects; beyond the fact that I’ve designed them, I mean. Our very own pussycat made it into the Jacobean project together with references to a favourite poem, Dutchness, and my mother. Bruce and Haasje carry memories of my favourite cuddly toy and a favourite aunt’s bedtime stories. What to put into Canvaswork? Well, there is a place which is very special to me and which, before moving to England, I would visit every year; where I got engaged; and which I remember going to with many special people like my mother, several aunts, and my in-laws: the Keukenhof, that famous Dutch bulb garden. And with its swathes of flowers, trees, ponds, fountains, sculptures and buildings surely there must be suitable scenes with plenty of depth and perspective. This one, for example, which I took myself some years ago:

Lots of perspective but a bit too much detail

Very pretty, I’m sure you’ll agree, and it’s got the different textures of flowers, path and trees, but there may just be a bit too much going on in it. Canvaswork is by its very nature rather more chunky than the other techniques, and it might be challenging to capture all the detail. Moreover, I decided that if I was going to celebrate my Dutch heritage in canvaswork, I was going to go all out. Not just tulip overload, but also a windmill. And the Keukenhof just happens to have a pretty good one smiley.

A floral sweep with a washed-out sky No clutter, but not a lot of perspective Good image, but it lacks sweep Sky, sweep of flowers, path, mill - but landscape orientation

Three of these pictures come from the Facebook page of a travel organisation that specialised in Keukenhof trips, so the first thing to do was ask for their permission to use one of them. They were extremely kind about it and said yes, take your pick, just send us a picture of the embroidery when it’s finished. I warned them that might be a while…

As for which to choose, well, all four have a lot going for them: perspective, different textures and lines, and some large areas in which to show off stitch transitions. They all have their drawbacks too – the first has lots of people and other “cluttery” things like bins and benches in it and the sky is overexposed; the middle two haven’t got the curvy sweep of flowers which draws you into the picture; and the last one again has rather a lot of clutter, and it’s in landscape orientation, which doesn’t fit my small frame quite so well. I may try cropping that one to portrait orientation, getting rid of some of the undesirable elements at the same time (although I’d still have to ask the tutor whether I’d be allowed to ignore that one rather prominent rubbish bin). And in spite of the lack of sweep, I like the third picture because it is nice and bold and hasn’t got too much fiddly detail. Anyway, I’m taking all these pics to class, so in between mounting Goldwork we’ll discuss which one is most likely to work. I’ll let you know!