Alexander Grabovetskiy, woodcarver, School of Woodcarving · Published 2022
How do you get the background of a relief carving flat?
Quick answer: A relief background comes flat by working crosswise to the fibres in shrinking bites, riding each pass on the ridge your last one left standing, then settling the final depth with a router plane. Its sole rides uncarved wood, so the blade reaches one height everywhere instead of wherever your wrist happened to stop.
Ground that looks flat and ground that is flat are two different things, and the wall in a finished room tells them apart at a glance. Low light rakes across a carved panel and every dip you left behind throws its own little shadow, right next to the shadows you meant to put there.
This workshop drops the whole background around the flowers to come, on the second of two mirrored panels. Nothing is shaped yet. What happens here decides whether the flowers sit on a clean field or on a run of small accidents nobody can fix later.
Three habits do most of the work here. Run the gouge across the grain, ride each pass on the ridge your last one left standing, and then let a tool with a sole in it settle the final depth. The last of those is a furniture maker's plane, and I think carvers should own one.
Before you drop the background
- Work across the fibres. Safer, and you feel the difference immediately.
- Bites get smaller as the line gets closer.
- A router plane rides the uncarved surface, so it cuts one depth everywhere.
- Flat means the same height front and back, checked, not assumed.
- A window you forgot to pierce on the saw can still be drilled through.
- Sweep is what makes a gouge, and without any sweep at all you are holding a chisel.
Why does carving across the grain feel safer?
The fibres have nowhere to run, and that is the short version of it. Run along the grain and a pass can dive, following the fibre down past where you wanted to stop, and on a background about to be five millimetres thick, diving is expensive. Cross those same fibres at an angle and each one is severed instead of split.
Always better to work crosswise to the fibres. Always safer. My opinion, and forty years have not talked me out of it. Watch the shavings and you can see why, because they come off short and curled rather than long and tearing.
A second reason gets mentioned by nobody. Crossing the fibres, you feel resistance change under the edge before anything goes wrong, so the tool warns you. Running along them, a dive happens faster than your hand can answer.
What is the ridge trick that saves your shoulder?
Make one pass and then stop and look at what it left behind, which is a shallow trough with a small ridge standing beside it. Now place the following one so that ridge sits roughly in the middle of your tool rather than under one corner.
Do that and the gouge takes waste from both sides of the ridge at once, balanced, with the wood supporting the edge instead of levering against it. A lot less effort. I am not using much force at all, and the wood is coming away faster than when I fight it.
Beginners put every pass on fresh ground and wonder why their forearm gives out. The ridge is free help, and it is sitting right there after every single pass you make.
Roughness at this stage is fine, which surprises people. It does not have to be super clean at the beginning, because a lot of that material is going to be gone anyway. Cleanliness is a later problem. Right now the only questions are how low and how even.
How deep should each bite be as you near the line?
Deep at first, then progressively shy. Out in open ground I feel safe grabbing about an eighth of an inch, three millimetres at a time. In some spots the pass is no more than 3/32 of an inch, which does the job just as well.
Closer to the line I get, a lot less material I am taking, maybe about a 16th, one and a half millimetres or so. The last few passes barely remove anything, which is the point. A small bite that goes wrong is a small problem.
And I work from outside in, toward the flower, never the other way. Safer, always. Where a shape blocks the approach I turn the whole panel around and come at it from the far side, and often that second angle shows me the true depth I could not judge before.
No need to reach final depth on the first side either. Take it partway, rotate the piece, come in from behind, and suddenly you can see exactly how deep the ground between two shapes has to go. Two half measurements from two directions beat one guess from a bad angle.
| Tool | Reach | What you get back |
|---|---|---|
| Number nine at 10 mm | Open ground | Waste gone fast, plus a ridge for the next pass to ride |
| Spoon 11 | Pockets a straight tool cannot enter | Access into places nothing straight will reach |
| Number two, tapped | The drawn line itself | A clean wall, provided the mallet stays gentle |
| Router plane | Anywhere with uncarved surface beside it | One depth everywhere, and you can check it |
One warning about row three, because it nearly cost me the panel. When the background is down to five millimetres, a firm tap with a mallet can send the edge straight through to the other side. You can easily, easily just go through, so do not bang with that mallet. A little, little force, and then check.
Should a wood carver own a router plane?
Mine says yes, and this is the one opinion in this workshop I will argue about. A router plane is a blade hanging below a flat sole. The sole rides on wood you have not carved, so wherever the blade reaches, it reaches at exactly the same depth. Your hand cannot do that.
If you are a wood carver, it is a good tool to have, and not just for the furniture makers or for custom woodworkers. Even doing only relief, it is a really good tool to have.
Watch the timing, though. Use it while you still have flat surface for the sole to sit on. Leave it too late, after everything around the ground is shaped and standing proud, and there is nothing left to reference from. On this panel I brought it out early, and the surface underneath came up beautiful and clean.
The reason it matters here is not beauty. Proof is the word. Putting the plane down, I know for certain the height of that detail is exactly the same on the back side and on the front side, and on a mirrored pair that certainty is the whole job.
Which blade goes in the plane, and when do you swap it?
Whichever one fits the space. I started at five millimetres, went down to about a 16th of an inch, a millimetre and a half, and finished with 3/32, a little less than an eighth.
Swapping matters more than the sizes. Mine is a Veritas, and its blades change in a couple of seconds without losing the setting, so the depth after the swap is the depth before it. I like Lie-Nielsen too. Veritas wins here on that one detail.
When the plane goes away, the blade comes out and lives in its own storage, separate from the body. Blades left in a plane get knocked, and a knocked blade is a setting you no longer trust.
Is that tool a chisel or a gouge?
Hold it up and look along the edge. A chisel is always flat. The absence of radius, the absence of sweep. If it is flat, it is a chisel, and if it has a sweep, then it is a gouge.
Which makes the number one in my hand a spoon chisel and not a spoon gouge, even though everybody at the bench calls the whole drawer gouges. Mine is an antique dovetail number one by Marples and Sons. The old Marples, when they used to be really good.
Why bother with the distinction? The number tells you the sweep, and the sweep decides which shape a tool can enter. Reach for a number two where a number nine belongs and you will feel the edge refuse the curve, and then you will start using force, and force is how panels get ruined. We have a whole workshop on choosing between them.
Can you fix a piercing you forgot to saw?
Easily, and it happened to me on camera in this very workshop. Two sections had to be pierced right through, and I had not taken them to the scroll saw. Not a big mistake, just a mistake, and the fix took under a minute.
A drill solves it in seconds: a few holes through the waste and the window is open. The slower answer works too, if the drill is across the shop. Take a number nine at an eighth of an inch and plug cut your way through, straight down, the same operation the drill does but by hand.
Either way the pencil lines around the hole disappear with the waste, and that does not matter at all. Nobody sees the route a window took to become a window. Lines get redrawn. I reestablish them afterwards from the first panel, and nobody looking at the finished wall will ever know which windows were sawn and which were drilled.
Must you draw every detail before carving?
Over-drawing wastes an evening, so the answer is less than you think. Doing the original drawing on paper, I work it out properly, every curve, every division. Putting it on wood is a different job entirely.
What goes on the blank is roughly where the flower will be and nothing more. Following all of those curves exactly does not matter at this stage, because I am still going to change them. An idea of the position is all I need, so no details go inside when I draw it on the wood. No reason to spend more time than the work needs.
Lines disappear as you carve anyway. When I drilled through those two forgotten windows, the pencil around them went with the waste, and I was not worried. I am not afraid of losing this line. I can reestablish it, and I reestablish it from the panel that is already finished.
That is the quiet argument for keeping a molding profile and a finished reference within arm's reach. Drawing is cheap when the answer is lying on the bench beside you.
Where does the molding line come from?
Off the first panel, three ways, each checking the one before it. First the width: calipers onto the old piece, line to line, then straight across to the new one. Calipers are probably my favourite measuring tool and probably the most accurate.
Then the drawing, and my trusty tool for that is my finger. Butt it against the edge as a fence, run it along, and the pencil follows at a fixed distance. Sharpen the pencil first. A fat line is a fat error.
The molding itself has a shape worth naming while we are here. A bullnose across the top, then fillets running around it, and that whole profile has to travel downhill and mitre into its neighbours. Furniture makers will recognise the trouble, because sometimes a molding plane simply cannot reach a run like this, so the answer is to carve it by hand, and a panel this size is excellent practice for exactly that.
Last comes depth, from the measuring arm, set against the first panel and left sitting about half a millimetre above the final surface. That half millimetre is deliberate. That margin is the one I still get to remove, and once you take too much you cannot put it back, because this craft is an art of deduction, an art of subtraction.
Which is also why I refuse to hurry. I do not want to rush it, and no reason exists to. Even when a deadline is sitting there, rushing is not really a good thing. This panel belongs to the Venice Room, and a room that size is not finished by anybody in a hurry.
The rest of it sits nearby in the school. Workshop one covers slicing and the first outlining. The course itself runs fourteen workshops from bare board to finished panel. Sharpening decides whether any of this is pleasant, acanthus over moldings takes the same combination further, and the furniture panel does it on a bigger board. Everything on my bench is listed in the tools course.
Your background wants to be one depth, and your hand alone will not get it there. Access to the school shows you the rest.
How it's done
- Open the ground with a number nine, working cross grain. A 10 millimetre number nine moves waste quickly, and every pass runs across the fibres rather than along them. Bites of about three millimetres out in the open are comfortable.
- Ride the last ridge with the tool you already hold. Place the ridge your previous pass left roughly under the middle of the tool. Waste comes off both sides at once and the wood supports the edge, so the work needs far less force.
- Shrink the bite as the line gets close. Within a few millimetres of the drawing, drop to a 16th of an inch, a millimetre and a half, sometimes less. A small pass that goes wrong is a small problem.
- Rotate the panel and come in from behind. Stop short of final depth, turn the work around, and carve the same ground from the far side. That second angle shows the true depth between two shapes.
- Stand a flat tool on the line. A number two held upright, tapped gently, is what makes that wall. Gently is the whole instruction: with five millimetres of background left, a firm blow goes straight through.
- Settle the depth with the router plane. Bring it out while flat surface still exists for the sole to ride on, take light passes from both directions, and check. Now the height is the same front and back, proven rather than hoped.
Frequently asked questions
Do you need a router plane to carve relief?
Carvers managed without one for centuries, so no. What it buys is certainty rather than possibility, which is a different thing. On a single panel your eye will get close enough. On a matched pair, where two grounds have to agree at the same height, the plane turns a judgement into a measurement you can repeat tomorrow.
Would a bench plane do the same job?
Not in a carved ground. A bench plane needs a long open surface to ride on and it cannot reach into a pocket surrounded by shapes. The router plane hangs its blade below a small sole, so it works down inside a recess as long as some uncarved wood remains beside it for reference.
Are spoon gouges worth buying this early?
One or two, yes. A spoon number 11 at two millimetres reaches into tight pockets where a straight tool simply will not fit, because the shapes around it are in the way. That single tool opened several spots on this panel that nothing else on my bench could enter.