Advanced Carving: Venice Room (Part 12) Workshops 132-143

Shadow in a relief carving is carved, not lit. One edge of every leaf drops low, the far edge stays high, and the surface between them tilts. The neighbour leaf then tilts the other way. Light at four in the afternoon finds edges that noon light walked past.

Your leaves are deep. You went right down to the ground in places. Stand back, and the whole panel still reads like a rubbing, flat as a plate.

Have you carved a leaf that measured deep and looked shallow? Do the hollows go black on your bench and grey on the wall? Depth is what you measure. Shadow is what people see, and the two are not the same thing.

Workshops 132 through 143 of the Rococo Venice Room panel are where this job stops being about removing mass and starts being about aiming light. Three flowers, a run of beads, one double leaf, and a mistake I made with the camera running.

What the shadow is actually made of

• A tilt. One edge low, the far edge high, never both at one level.
• An alternating tilt, so the leaf next door leans the opposite way and the light never repeats.
• Roughly 3 mm (1/8 in) of air held back for the undercut you have not made yet.
• Waviness.
• A hard stop drawn on the wood before any waste comes off, so you know where the pass must not go.

Why does a deep leaf still look flat?

A leaf reads flat when both of its sides sit at the same height. Take it down 12 mm, half an inch, and you have a hole rather than a form. What the eye reads as shape is the step from light into dark across a surface, and a level surface has no step in it anywhere.

On this panel I stopped asking how deep and started asking which way. Every leaf in these twelve workshops got the same question put to it out loud. Which one goes down? I would put a finger on the low end, look across the work from where the client will stand, and answer. Then I dropped that side almost onto the surrounding surface and left the far side standing.

Those concaves make a big difference as far as the shadow and light. Huge difference. The whole trick is one decision, and a decision costs nothing.

Which edge of a leaf drops first?

Point of view decides it every time. Before I drop anything I walk to where the piece will really be seen from. This panel goes into a room, so a person stands to one side and looks across it, never square on from above the way my camera does. From that chair the low side of a leaf shows its whole hollow and the high side shows nothing but an edge.

Then I redraw. The pencil goes back on long before the shape is final, because the waste I just took off carried half my line work away with it. A drawing you can still read is the difference between deciding and hoping. I redrew that first leaf more than once, and I would do it again.

One rule I carry over from the old work: a small leaf sits no higher than about 3 mm, an eighth of an inch, above the line of the long leaf running underneath it. Any higher and that leaf stops belonging to the bunch. It starts to look glued on, and a person catches that in half a second without knowing why.

Why should the next leaf tilt the other way?

Light moves through a room all day. A panel where every leaf leans the same way looks alive for about two hours and dead for the rest, because the sun that lit those hollows in the morning is coming from somewhere else by dinner. I sloped two of these leaves one way and the third one the other way, purely for the shadow's sake.

Try it on two leaves side by side. Drop the left edge of one and the right edge of its neighbour, then carry the board to a window and turn it slowly in your hands. One of them is always dark. A carving that keeps changing through the day holds a person the way a flat one never will, and the eleven parts before this one were building toward exactly that.

How much air do you leave for an undercut?

Three millimetres, 1/8 in, held back before I need it. An undercut is what turns a step into a shadow, and you cannot make one if the wood behind the edge has already gone. While the shape is still rough I leave that sliver everywhere an edge will eventually float.

Undercutting runs down the sizes. The first pass in is a number 11 at 3 mm, right up to the pencil. Then 2 mm, then 1 mm in the tight spots, and half a millimetre where the shadow has to read black. Some of those places are fragile by then. Fragile and awkward. A weak corner without its shadow is a corner nobody looks at, so I go in anyway, mallet in hand for control.

An opening is a different animal. In one spot I went right through the wood on purpose, because that section was reading heavy and real light through a real gap makes it lighter. An undercut fakes air. An opening is air. Both are decisions, and neither one is an accident.

Which gouges do the shadow work?

A number 11 veiner does most of it, in every width I own. Everything on this panel came off a small set used over and over rather than a wall of tools, and the width in your hand matters more than the name on the handle. Here is what the twelve workshops actually used.

The tool	What it does on this panel	Widths I used	Why that one and not some other
The 11 veiner	outlining, undercutting and dropping one edge low	0.5 to 30 mm	one sweep carries nearly the entire job here
Number 7 gouge	slopes, concaves, rolling a petal back over	14 to 76 mm	waviness, and it leaves no flat spot behind
Number 9 gouge	the plug pass that sinks a centre	20 mm, 3/4 in	round bottom, deep bite, quick way down there
Number 3 chisel	stops straight down, blending a ridge away	any width that fits	nearly flat, so it cleans up very well
The back bend	inside a curve a straight tool misses	6 to 13 mm	reaches right under the roll without walking around
Spoon and skew	the dome hidden inside a flower centre	2 to 13 mm	gets where straight steel simply will not go
A steel punch	texture laid straight over a carved grid	more than one size	blends two fighting directions into one calm surface

Sharpening is not a stage that happens before the job. Halfway down one hollow in workshop 143 the tool needed sharpening, so I stopped, took care of it, and came back. Much better now. If you want that part on its own, it lives in the sharpening course, and the full set is laid out in all about tools.

What is an eye in a carved leaf?

An eye is the small pocket of shadow where two movements meet. On a real acanthus leaf it sits at the root of every lobe, and it is half the reason such a leaf looks alive at all. These leaves are not acanthus. The design comes off 15th and 16th century buildings in Venice, and the leafage does as it pleases. Even so, wherever I see a spot that wants an eye, I give it one.

The pocket takes a single pass. A small number 8 at 3 mm goes straight in like a plug, and I blend outward from there. Beside it a pipe has to pop up, a convex, or the eye is only a dent in the wood. Eye down, pipe up. That pair is what the Corinthian acanthus is built out of, and it carries over to leafage that has no name.

Divisions behave the same way. Around the middle of the first flower I marked five, then made sure no two of them sat at one level. Each slips a little under the next, and the pattern never repeats itself around the ring. The unevenness is the entire reason it does not look stamped out by a machine.

Why carve a flower wavy instead of round?

Nature holds no clean circles. A petal ring turned smooth as a bowl announces itself as a machine part the second light crosses it, and you will never find a perfect shape out there in creation. I like waviness when I carve flowers, and waviness is only a plus.

Wavy does not mean careless. The wave still lands where a Fibonacci sequence puts it, and by that measure it is perfect. What it is not is symmetrical. Two different words.

The order of work surprises students. I carve the centre first, before the ring of petals around it, because the middle is the strongest wood on that flower at that stage and there is less chance of snapping something off. Petals next, then the roll, then the undercut. Backwards from how the finished thing reads.

Sinking that middle takes a number 9 at 20 mm, three quarters of an inch. Five millimetres down, sometimes six, roughly a quarter inch, and I stop. Then the wavy edge gets chopped toward the centre so every pass after it runs downhill instead of uphill.

What makes a petal look like it was applied separately?

Carve the edge nearly paper thin and hollow the wood out from underneath. Look at the finished panel and those flowers seem to have been made separately and glued on. They were not. Same board, one piece of wood, and the whole illusion is a thin edge with deep shadow behind it.

Rolling does the other half of the work. A petal that lifts from the middle, goes over, and then rolls back underneath is doing three things at once, and I get there with a number 7 held upside down, a little at a time. Roll it hard where the flower crowds the leaf next door. Rolling stretches the shape, and a stretched petal fits a tight gap that a fat one would fight.

Then round everything, slice by slice, in all directions. I do not want one flat spot on a flower. Not even one flat surface, or it stops reading as three dimensional and starts reading as a shape that owns some petals. Slow work. It is where the Hashemi flower project earned its hours, and I spent fewer here because of it.

How do you carve the centre of a flower?

Draw a grid on it first, in two colours. A flower's middle is the one place where carved texture can carry real information, so I lay out two sets of lines that swirl against each other. One set runs in toward the centre. The other arcs the opposite way, closer to a golden ratio curve than to a plain radius, the way you would see it in nature.

Finding that middle is not guesswork. I take the two edges, divide the span into ten, and move six parts across. The centre lands off true and lines up with the movement running through the rest of the panel, which a dead centre never manages. That one move is the whole measurement. Then orange for one set of lines and pencil for the other, so I can tell which ones I have already carved.

Each line gets one pass straight down, angled a little, then a second pass about 1 mm away leaning in to meet it. A V. Not a V tool, though, which would be quicker and is not authentic to the period. I use a skew chisel and pay the extra hour. Watch the grain in those little arcs, because it turns direction inside the curve and it will chip if you hurry it.

Last comes the punch, straight over the V work I just finished. Feels like throwing away an hour, and it is the opposite. You can still read every line through the texture, and the punch blends two competing directions into one surface. Three flower centres on this panel got the same treatment, which is also how the Venetian rosette holds together.

Where do beads belong on a moulding?

Split the width six parts out of ten, and the bead line goes there. A row of beads is the oldest way there is of putting rhythm beside a leaf, and the bead and reel has been doing that job since antiquity. What changes from century to century is the shape lying next to the row.

Two of them were common at the end of the 15th century and the start of the 16th. Concave against convex, which exaggerates the beads and throws hard shadow. Or double convex, two rounds side by side, softer, and that is the one I chose here. Six millimetres each, a quarter inch, worked down the number 11 widths from 7 mm to 1 mm before any rounding starts.

And then I broke my own rule on the far side of the panel. I moved the division lines off that mark, dropped the overall depth on that side, and popped the beads a little higher instead. It looks much better that way. A moulding is a proportion you check by eye with the wood in front of you, not a formula you obey, and the same goes for the volute beside it. Acanthus over mouldings works the same argument from the other end.

Do you need a back bend gouge?

A back bend reaches inside a curve where a straight tool cannot follow. You can get the job done without one by walking around the bench and coming at the wood from the other side, and I show that in the video, because most people do not own one yet. With a back bend you set the angle and go, no walking.

Read the numbers on the handle rather than the shape. A Swiss handle stamped 25/13 is a back bend with a number 5 sweep, 13 mm wide. The 20 says back bend and the 5 says sweep. Mine is reground into a skew on top of that, so one corner can go very deep. Henry Taylor made the back bend I picked up for the beads, and I like how their tools feel. The handles that say Swiss made are Pfeil.

Then step down through the widths. A 30 mm number 11 gets me to depth fast, and the moment it starts tearing I move to 25, then 20, then 18, then 15, then 10, then a 7 mm spoon to scoop out the last of it from two sides. One hollow, six tools, and only the last of them is about how it looks.

Why does a carving look worse on camera than on the bench?

A lens flattens the exact thing the tilt was for. When I look at this piece on a screen it does not look as nice as it looks in real life. On the bench two eyes see around a hollow and the shadow shifts as your head shifts. A camera has one eye and it holds still.

Worth knowing if you learn from video, mine included. A section that looks like a mess on your monitor is often finished work under a bad angle and a heap of chips. I said it looks like mess out loud more than once here, and the carving was fine every time.

I bought a set of rasps about a year ago and I still do not trust them. I used them on a few spots here, left the surface rough because that part gets painted anyway, and went back to gouges. They save time now and then. More often they make problems, and my normal tools do better.

In workshop 143 I rushed a number 7 and broke off a separation I wanted to keep. I could have glued it back. I did not. It opened the area up, my separation moved across, and the section came out lighter than my drawing, which is luck rather than method. I am not going to tell nobody except you that I made that mistake.

My granddaughter was visiting the day I carved the first flower, and you can hear her in the background of workshop 134. I left her in. These videos are exactly that. A real bench, a real client waiting, forty years of this work, and a man deciding out loud. Everything above is what the panel needed, and the pass by pass version sits in the workshop videos themselves, which is what tuition to the school opens. Not sure this is your project yet? Then start somewhere smaller, or bring a question to the techniques and questions course.

Shall we carve?