#sancai

Src: Fig 1 of geographical distribution of three main Tang Sancai kilns, J.F. Cui et al., Lead Isotope Analysis of Tang Sancai Pottery Glazes from Gongyi Kiln, Henan Province and Huangbao Kiln, Shaanxi province, pg 598

From chemical analysis, we know that Sancai glazes do tend to share some similarities. Common characteristics of Sancai glaze are that they are lead-based, twice-fired at relatively low temperatures (700-1000C), and very viscous. A number of elements are used as colorants, the most common of which are iron (Fe), copper (Cu), and cobalt (Co), which could be obtained from minerals. When oxidized they turn amber, green, and blue respectively (think of rust and old copper like the statue of liberty). Other elements like antimony and manganese could also be used to make brighter yellow or even purple shade.

Different people used different methods of producing similar colors, though, so there was not necessarily a 'standard' recipe for each color or anything like that. Still, I think the most commonly known mainstream shades are this bright grass green, this amber color, and an off-white base clay color.

Src: 唐三彩仕女陶俑 A Blue and Amber-Glazed Pottery Figure of a Court Lady Tang Dynasty, Lot 249, Sotheby's

Jiang Qiqi's thesis on Tang Sancai puts particular emphasis on the possibly foreign-introduced use of cobalt colorant as a vital step forward in ceramic technology, and an important factor in the development of Tang Sancai as a style. It allowed for the use of blue tones that were previously not possible, broadening the spectrum of colors that artists had access to in this medium, washing the Tang Dynasty in color. To me it's a splendid example of how the Tang Dynasty exchanged ideas with foreign cultures, then innovated new concepts and technologies from what they learned.

Since the main three colors are well-represented in the other two sets in the Sancai line, I wanted to have at least one set featuring a slightly less-common shade variety of one of the three main Sancai colors. I decided on green, which sometimes occurs in this pale, almost teal-leaning tint. This is described in Chen Siya's thesis as "opal green" (however a lot of the color names are listed in Japanese which I don't speak so it's possible I missed one that's closer to the teal-ish shade).

Src: Table 4.3, Glazed colors of utensils in the Ye City area, Chen Siya, Production, Circulation and Consumption of Lead-glazed Pottery, From Latter Northern Dynasties to Tang Dynasty, pg 198

I'm not sure if cobalt was necessarily used in this to make it bluer, because pigments + chemical reactions + firing process do not mix like paint does, but it does have a blue hue to it that kind of ties it back to the whole cobalt colorant thing.

Drip-Glaze Effect

When I first decided on the Tang Sancai theme, I started Pinterest-ing pictures of Sancai ceramics just to get a feel for the possibilities. I didn't want to just recreate the outfit of a sancai figurine—not that there's anything bad about that; lots of hanfu shops do reconstruction designs, I just wanted to maintain some personal creativity here, being focused on design and all, especially since the other two sets in the Sancai line already take fairly heavy inspiration from figurines.

Turning away from figurines, we can take a look at household wares, or what might be called "utensils," especially vessels like pots and vases. I primarily referred to Chen Siya's Kanazawa University doctoral thesis, "Production, Circulation and Consumption of Lead-glazed Pottery, From Latter Northern Dynasties to Tang Dynasty" for analysis of glazing techniques. Utensils and figurines share many glazing techniques, but there are some methods that may be more common in one category than the other.

Src: Pinterest link, previously linked to Christie's but unfortunately the listing page is no longer available (it was available when I first came across it but I didn't save it so I don't have any metadata on it anymore)

There are a lot of Sotheby’s/Christie's and similar antique auction photos on Pinterest. I was particularly drawn to this kind of glaze that appeared mostly on large round vessels like vases and pots, where it looks like the surface was colored white first, and then poured over with colorful glaze so that the colors ‘melted’ down over the white, forming this really beautiful abstract pattern. (I hypothesize that the glaze was poured because of Jiang Qiqi's thesis which states that "Pouring, dripping, filling and brushing were the methods of applying the glazes. Pouring is the main process for the glazing operation" (Jiang 212), but it seems like dripping could also apply here.) Anyway, the one I focused on for inspiration is this three-footed pot (三足瓮 / san1 zu2 weng1) from Christie's.

L: Figure 4.17, Examples of decorative positions // R: Table 4.5, Classification of decorative positions // both Chen Siya, Production, Circulation and Consumption of Lead-glazed Pottery, From Latter Northern Dynasties to Tang Dynasty, pg 201-202

Chen Siya's paper classifies the positioning/location of glazed parts of ceramic utensils into 7 types (pg 201-204). It seems like this placement would count as the exterior wall being fully glazed "except bottom," since in Figure 4.17 the example for type 4 also has a significant stretch of the bottom unglazed, which would fall into types 2, 4, or 7.

Src: A Blue, Amber And Straw-Glazed Pottery Tripod Censer, Tang Dynasty, Lot 255, Christie's

The inner wall of this pot is not visible in the photo, but other similarly formed Sancai pots from Christie's show at least some coloration on the inside of the lip (as in the one shown above), so I think it's likely that this pot is either type 2 or 4. From there, Table 4.5 lists "sanzufu" (I believe this is a mistranslation of 三足瓮, which would be sanzuweng, because there is a character pronounced fu that looks very similar to weng; there are a lot of typos in this paper so a mistranslation is likely) in types 4 and 5, making the only overlapping one type 4 in the 7th century. From this I deduced that it was possible that this vessel may have been made in the Early Tang Dynasty.

Design Breakdown

The Pegasus / 天馬行空 is a 5-piece set based on feminine silhouettes from the 7th century during the Tang Dynasty. The five pieces included are:

齊胸破裙 (chest-high skirt)

陌腹 (short overskirt)

披帛 (shawl

對襟短衫 (long-sleeved inner shirt)

唐褙子 (short-sleeved jacket)

1 齊胸十二破裙 / qi2 xiong1 shi2 er4 po4 qun2 / Chest-High Twelve-Piece Skirt

First, to create a good ‘layering’ effect, I needed to have enough surface area to work with, which is why I chose to make the skirt in this set a chest-high skirt. It's an iconic silhouette of the Tang Dynasty and gives me more visual space to build on top of.

The fabric I chose for this is an off-white 100% linen with very subtle organic striations. The white color in Tang Sancai is not actually white; it’s a colorless glaze that gets applied over the clay base, which is itself an off-white color. As a result, you can see that the natural texture and impurity in the clay base sometimes shows through, so I wanted the fabric to reflect that a little bit.

Poqun have the advantage of not having pleats, meaning that they can be made from natural fiber without needing to re-press the pleats every time you wash it (which is the main reason why many pleated hanfu skirts use polyester).

There are also shoulder straps, because ain’t nobody got time to be hiking up their skirts around their torsos every five seconds.

2 陌腹 / mo4 fu4 / Short Overskirt

Also called the 腰裙 / yao1 qun2 / "waist skirt," the mofu is a layering accessory that appeared to be most popular during the early Tang Dynasty. This is a slightly less common/well-established hanfu piece, so some research had to be done to stay defensibly authentic.

Research

The most popular depictions of the mofu come from frescoes on the walls of the 新城公主 / xin1 cheng2 gong1 zhu3 / Princess Xincheng's tomb and 燕德妃 / yan4 de2 fei1 / Concubine Yan's tomb.

L: 新城公主墓侍女壁畫, mural of ladies-in-waiting from the tomb of Princess Xincheng, discovered 1995 // R: 后墓室东壁南侧伎乐图, 唐高宗咸亨二年（671）, 燕德妃墓后墓室东壁南侧, 唐代壁畫珍品數據庫, mural of female entertainers from the southern face of the eastern wall of Concubine Yan's tomb, 671CE, Tang Dynasty Mural Database

Structurally, they appear to be basically the same as a shortened poqun (although we can’t be 100% certain of that because we haven’t found a surviving artifact of the actual thing yet). We know they're tied around the bust because there are ribbons visible holding them there in the paintings, and we know that some (but not all) of them have a contrasting border trim, and that many are striped similarly to the full poqun skirts.

Liu Kaixuan et al.'s paper "Archaeology and Restoration of Costumes in Tang Tomb Murals Based on Reverse Engineering and Human-Computer Interaction Technology" digitally models the outfit of the second lady from the left in the mural from Princess Xincheng's tomb. Their model of the the striped poqun is in line with what we know—alternating trapezoidal pieces of fabric sewn together—but the mofu they constructed was flat and rectangular.

Src: Figure 1, 3D interactive pattern-making for Tang Dynasty tomb costumes, Liu Kaixuan et al., Archaeology and Restoration of Costumes in Tang Tomb Murals Based on Reverse Engineering and Human-Computer Interaction Technology, Sustainability

This is definitely not impossible, but I have more reasons to believe that the mofu may have been similarly patterned tot he full-length poqun skirt. Although the mofu of the lady in question is a solid color, you can see that the other women around her are wearing mofu that are striped similarly to their skirts. This can be achieved by simply using a striped fabric, but if the assumption is that the striped poqun is color blocked, why not the mofu? We know that non-striped poqun are still blocked out in the same way even when all the trapezoids are the same color, and we see the woman on the far left wearing a solid-color poqun as well.

Also, taking a look at the mural from the tomb of Concubine Yan, the second woman from the right appears to also be wearing a single-color red mofu; however, there are clearly many defined lines running down its length, leading me to believe these may be seam lines. The mofu also all seem to be relatively loose and free-flowing at the hemline, which to me suggests that they may be wider at the hem than at the bust, which is one of the defining characteristics of the poqun construction.

Design

This was probably the most fun part of designing this set, as well as the most frustrating. I wanted the mofu to ‘drip’ down the skirt like the colored glaze. I really liked the uneven edges that were formed by the dripping glaze on these vases. I wanted to apply that to the hem of the mofu to get that three-dimensional layering effect.

The trouble with this is that organic curves, especially the concave kinds formed in between the glaze drips, are really really annoying to hem. We work with a number of hanfu workshops across China, and many of them are very small 1-3 person side-gig businesses that specialize in hanfu; they don’t have the power of a whole factory line behind them. A big part of my back-end job is also to communicate exactly what I need the design to look like to the people cutting & sewing it, and if you think about it, it would be an absolute nightmare to try to get the shape exactly right every time, plus ask them to hem it. 

My solution has to do with modern technology: laser cutting! I do a lot of laser cutting for rapid prototyping purposes, but laser cutters are a very versatile tool and can be applied in a variety of situations. The key advantage of laser cutting fabric, especially polyester as in this case, is that because the cutting is done by heat/light, the edges are automatically burned/melted shut, so the fabric won’t fray even if you don’t hem it.

Digital pattern pieces and sewn photo (sorry about the annoying watermark but people have stolen my art before and im not taking any chances x-x)

I’m sure there are craft people and cosplay people around here somewhere who’ve done similar things with polyester fabric or rope—you can seal the edges by holding a lighter up to the edge and letting the fibers melt, gluing themselves to each other. Using laser cutting also gives me more control over exactly how I want each piece to look, since the laser follows a specific modeled path over the print, which I can just CAD and then send over digitally. 

Even though there are creative modifications and modern technology at play, the skirt still retains the traditional cut of the broken skirt: right-trapezoid skirt pieces seamed together leg-to-leg.

Printed + laser cut pieces, construction doodles on graph paper

Typically, because of the difference in length between the two legs of a right trapezoid, there’s a little piece that ends up poking out at each seam, which you trim off before you hem the whole skirt. To avoid needing to cut the fabric with scissors, which would require hemming, I designed the bottom of each trapezoid to match up at the edges, so that the organic shape of the skirt takes care of that problem beforehand. There are two different variations of the shape, which match up with each other on either side, meant to be seamed together in an alternating ABABAB pattern. When put side by side, they look something like this:

The other important part of this design that I have yet to talk about is the fabric pattern. I knew I wanted it to have mixed-color sancai patterns, to mimic the glaze, so I started putting some colors down and blending them out. Unfortunately, my first attempt mostly referenced the colors in these figurines, and kind of looked like vomit. 

The original idea + how it actually turned out. This was. Not... a look.

So that got scrapped. 

That's when I found the sanzuweng I mentioned at the beginning of this post. I had scheduled three sets to be in the Tang Sancai collection, so I figured that I should depict a range of possible colors to demonstrate the possibilities. 

So I tried again, this time with teal and a burnt golden color. It’s really really hard to get a realistic ‘melting’ pattern when drawing digitally by hand—even if you have a reference, it’s hard for a digital program to mix colors in the way that melted pigment suspended in liquid mixes; the math just isn’t really the same, and again it ends up coming out looking kind of like vomit most of the time. 

Instead, I went for a more geometric design that I found on that Sancai pot. It still has the melty elements that I wanted, but it’s limited to the edges instead of dripping all over like the other examples. I put the teal-green down first, carved out the white diamond resist parts, and then used the airbrush tool to add the golden color to the top/bottom, resulting in a really interesting burned-edge look. I really like how it turned out—the darker hemline emphasizes the laser-cut curve against the white background.

3 披帛 / pi1 bo2 / Shawl

The pibo shawl matches the mofu; it’s made out of the same fabric (polyester chiffon) and also laser-cut so that the two ends have drippy shapes as well. It’s not a completely even gradient; I wanted the color blending to look organic, so I used the airbrush tool on the ends to keep it natural.

The laser cutter bed was only 200cm wide, and the pibo was supposed to be 300cm long, so the pibo had to be cut in two 150cm long pieces, then seamed together in the middle.

This is not out of the ordinary even for traditionally cut pibo; you can cut 1m of fabric in half and then recombine each half to turn it into a ~300cmx50cm pibo if you’re okay with having a seam in the middle, but if you don’t want a seam you have to use 3m of fabric and waste 2/3 of it. This is fine if you’re making a bunch of pibo at once, but less ideal if you’re just making one.

4 對襟短衫 / dui4 jin4 duan3 shan1 / Long-Sleeved Inner Shirt

I don't have a ton to say on this shirt design-wise, it's a fairly simple parallel-collar shirt with long straight sleeves. The sleeve cuffs have a 聯珠紋 / lian2 zhu1 wen2 / linked-bead pattern print I illustrated based on winged horse medallions from the Tang Dynasty, but it's really just for color matching purposes here; that fabric was primarily created for the 身騎白馬 / shen1 qi2 bai2 ma3 / Chrysaor set (which is the next design background post coming up) so I'll explain that one there.

The fabric used for the main body of the shirt is a very light and silky cotton mix (60% polyester 40% cotton). It's meant to mimic the look of 香雲紗 / xiang1 yun2 sha1 / Gambiered silk, also known as tea silk or Guangdong silk, a type of silk with an extremely complicated dyeing process that originated in Guangdong. The production of xiangyunsha silk involves a multi-week process using natural materials such as gambier juice, river mud and soil, and the sun. It's only viable during certain seasons and produces an extremely unique double-sided fabric, black or rich brown on one side and usually patterned and jewel-toned on the other.

As a result, real xiangyunsha fabric is—unsurprisingly—incredibly expensive. Having used it in other designs and custom orders, however, I can tell you that this fabric gets remarkably close in touch. Of course there are qualities that you can only get from real silk, and the color is just a shade shy of the richness of real xiangyunsha, but it's definitely a recognizable imitation with a luxurious feel.

5 唐褙子 / tang2 bei4 zi0 / Short-Sleeved Jacket

This type of Tang beizi is often called the 武周唐褙子 / wu2 zhou1 tang2 bei4 zi0 / Wuzhou Period Tang beizi. 武周 is the name of the period of Wu Zetian's reign, the only female emperor of ancient China, who ruled from 690–705CE. This is usually considered the tail end of what's considered the early Tang Dynasty.

The main reference (left below) for this type of Tang beizi is usually cited as this tomb painting of a woman from Astana cemetery in Xinjiang from the tomb of Zhang Licheng, who died in the year 702CE, which is within the Wuzhou time period, thus the name. The collar starts out in a V-line shape before narrowing and meeting into parallel lapels further down the body.

L: Astana Graves Dancer, unearthed 1972, Xinjiang Uighur Autonomous Region Museum, Wikimedia Commons (cropped) // R: 唐三彩釉陶梳妆女坐俑, 1955年陕西省西安市西郊王家坟唐墓出土, 陕西历史博物馆藏, Wikimedia Commons

Some Sancai figurines have more solid fabric designs on their skirts, such as this sitting figure out a woman above on the right from the Shanxi History Museum, with dozens of tiny 3D flowers or designs that appear to have been added on top or other times impressed into the main clay body. 

I thought that was aptly represented by using a two-toned jacquard where the pattern is woven in, so that there's a texture change as well as a shade change between the design and background. The color of the fabric was also very similar to the shade of green most commonly seen in Sancai. This fabric is 40% viscose and 60% polyester, medium weight, giving it a certain durability along with a subtle sheen across its surface. Because of its weave, the inside is a mirror of the outside pattern.

-

Okay I'm out of images so I guess I'm done lol I even cheated by combining instances of two images next to each other into one screenshot

Working on the 身騎白馬 Chrysaor post and adding all the sources into the Sancai works cited sheet (there's a couple in there right now but must of them are still open on my browser im working on it) (there are. so many tabs open on my computer rn), happy new year everyone!!!