JMonkey's Desem
This post illustrates starter maintenance and an end-to-end bake of a whole wheat Desem loaf using home milled hard white spring wheat berries (Prairie Gold) using JMonkey's Desem post as a blueprint. It is relevant to some of the starter yeast-vs-LAB ratio discussions on TFL this past week. It is fun to find such useful posts and discussions from the TFL "archives" back in 2009. He summarizes the style succinctly, and helps dispel a lot of mystique around this style of bread:
I enjoy making Desem, which is really just a traditional Pain au Levain (French sourdough) made with whole wheat flour. From what I've read (and if others have read differently, please, chime in!) the French, historically, have disliked a strong sour flavor in their breads, and so bakers had to work very creatively to eliminate as much sour as they could, especially when sourdough was the only levean they had!
Here's what they did:
1. They kept their starter firm.
2. They kept their starter cool (50 to 60 degrees F)
3. They used quite a bit of starter so that the bread would rise quickly and the bacteria would not have much time to produce a lot of acid.
I've been interested in heading in this direction for my regular home milled whole wheat bakes, and a number of Desem posts from notable TFL bakers caught my attention, which I summarized in this thread.
Starter
I've been maintaining a cool 50 % hydration starter fed at a 1:1:2 ratio for a couple of weeks and have had some luck achieving the desired temperature range of 50-60 F by storing the starter in a mason jar inside a wide mouth thermos with large ice cubes refreshed at the 24 hour feeding cycle. A lid and silicone jar top provide some insulation to prevent direct contact between the starter and the ice cubes. From occasional measurements with an IR gun this seems to reach an initial temperature of a little below 50 F and it slowly climbs to a little below 60 F by the next feeding, which can be adjusted by the number of ice cubes. The stiff low hydration starters don't readily provide the same peak volume feedback associated with higher hydration starters, so this exercise has required a little bit of blind faith. In this temperature range, feeding the starter the night before and using it the following morning seems to work out fine in practice. One recipe I read described a ripe Desem starter as resembling the texture of a kitchen sponge. Using a pH meter or Doc's weight loss approach may provide a better mechanism for optimizing the maintenance schedule, and I'm interested in any additional thoughts in this direction.
Dough
The dough was mixed for an initial conservative 75% hydration per the above post, and I used a refrigerated overnight saltolyse, so very little gluten development was required on the bench. I used the water allowance from the low hydration starter for the laminate-roll-spray-and-cut-up style mixing and continued to add water by feel with the sprayer. I bulk fermented at a warm active kitchen room temperature (77-80 F) to roughly 1.3x using an aliquot jar, at which point I pre-shaped, rested, folded, stretched and rolled it into a tight burrito. I dusted it and placed it into a banneton for a very short final proof (probably 1.5x via the aliquot meter). The dough was very strong. I have found the HWSW Prarie Gold exceptionally easy to work with for a whole grain flour and will be interested to try a similar bake with some heritage grains soon. I slashed it then placed it in a batard clay baker in the oven at about 475F for 20 minutes and another 18 minutes uncovered, after lowering the temperature slightly.
Comments
How does it taste?
Paul
This is a very mellow loaf of bread. It is light in texture (for whole wheat) and taste. It may be more versatile and better in a supporting role than the average sourdough loaf for that reason. I expect it would please many, but in so doing might put off some others. The flavors are subtle and the wheat itself is stronger in the mix, so the flavor is likely to change more significantly with the variety of wheat used. It makes me want to try repeating this with some heritage red wheat varieties, like the turkey red or red fife berries I had started with. I loved the smell of those while milling and mixing but was slightly disappointed that it didn't quite translate to the flavor of the final bread. I'm curious if this formulation might allow more of that. I have recently picked up some Yecora Rojo, and will try this again with that grain. I think some of the loft is due in part to the Prairie Gold wheat itself, which is extremely easy to bake with for whole wheat.
Really good looking loaf. I may have missed it, but what is the percent whole grain, is this 100%?
Benny
Thanks. Yes, it is 100% whole grain. Desem is traditionally a100% whole wheat loaf. I'm using a hard spring white wheat here, so it has a neutral color.
Thanks for your hard work and for sharing your learnings in this thoughtful, well-presented post. Your bread looks like a real achievement and well-deserved reward for your efforts.
A few months ago, I encountered a very long blog post about Desem elsewhere on the interwebs. It really captured my imagination, but I didn't have the wherewithal to pursue it at the time. You've probably reignited my curiosity. I'll be following your Desem journey with interest and may, at some point, dip my toe in the water.
–AG
Thanks, I'm glad to pick up the trail. I'm still trying to understand this approach and have some doubts about my implementation (as with other breads where I haven't had the the chance to try the real thing). That said, the ingredient list really does make this "yet another sourdough". I still haven't done the accelerated steam bath final proof, and I probably should to do the post justice. The process I am following does produce a mild flavor and the loaves do feel light for whole wheat. This experiment has had me lowering the temperature of my bakes to <= 475 F, away from the usual >= 500 F temperature I had internalized being necessary for oven spring in hearth loaves. The crumb is softer and the bread doesn't seem to stale as quickly. I think I had been slightly overbaking my whole grain loaves before this. I am curious how the low temperature in starter maintenance fits into this, as I typically associate that with promoting LAB (or their byproducts), which doesn't seem to be the goal here. The stiff + cool starter discards from this approach do behave differently than the typical starter, and the dough still has plenty of integrity when it is done. I think I could make a passable loaf from just the accumulated discards.
100% home milled whole wheat: starter fed w/ prairie gold wheat, main dough mixed with yecora rojo (estimated 75-78 % hydration)
100% home milled whole wheat: starter fed w/ prairie gold wheat; main dough mixed with yecora rojo; estimated >= 85 % hydration; lamination mixing
Yecora Rojo Desem @ 88.5% hydration
ingredients: https://fgbc.dk/15nl (scalable bakers percentage link)
The starter has been fed with Yecora Rojo, so there is relatively little Prairie Gold in the mix.
As a practical matter, I've migrated more or less to the 60% starter hydration level JMonkey mentioned. I started at 50% but have been using the sprayer to bring things together more easily by feel. I have weighed the last two starter feeds after this and both came out to 60% hydration. At this hydration in the 50-60F range you do get fairly good starter rise feedback in <= 24 hours if using something close to 1:1:2, so no magic is required. I have also found the gold standard Weck starter jars fit perfectly inside the 47 oz Wide Mouth Thermos. The jar lip catches perfect around a ring at the top of the thermos such that the jar hangs securely from the top while allowing the themos cap to screw completely closed -- as if by design. That leaves enough space at the bottom for a handful of regular ice cubes. This is good news as it should allow me to increase the starter volume for 2 loaves at the high volume PPF. I'm also considering purchasing a small thermoelectric cooler to further streamline things.
I still prefer the look and feel of the initial Prairie Gold loaf slightly. That was mixed by feel, and there is probably a fair amount of process variance in these bakes, so I can't yet attribute differences to the wheat. I'll try increasing the hydration a little more to see what comes of it. I believe an extremely open crumb is not the goal with Desem. A slightly more open crumb, however, may make this loaf more inviting for those accustomed to refined white flour.
This bake used an 85.7% hydration saltolyse, allowing a smaller portion of water (52.5 g) in the final mix to achieve the target 88.5% final hydration.
This feels slightly easier than previous 75% hydration saltolyse, with a larger portion of water (90 g) in the final mix for the same 88.5% final hydration.
Ever since our last interaction. But I'm having trouble understanding what the advantage is over the 100% whole grain loaves I'm accustomed to making with a more conventional approach. There's not much information out there that I've come across, and some is pretty conflicting. I've started converting my starter to something more stiff and it now lives in our 59°F cellar. Yesterday I fed it 1:2:3 and today it had doubled in 24hrs so I fed it 1:4:6 (the culture was originally fed 1:5:5 regularly at about 69°F). My only working theory at this point is that the culture will adapt to thrive under these conditions then excel when mixed for dough at room temperature. This may or not be the case (we'll see), but still results in something distinct from a true desem. Are you building up your culture to be 50% of the total dough weight before using, as I've seen some methods call for? Is it supposed to taste more sour, or less? Is the final dough meant to achieve an equivalent fermentation state in half the time?
I'm interested in learning about whole wheat milling and baking, and my searches kept turning up older desem posts from notable bakers on this site who all raved about the whole wheat friendly flavor. It seemed like a good place to start. I quoted four of them a while ago in THIS POST. I like dmsnyder's description (I think some can overstate the mild sweetness):
That was what piqued my interest initially, and I think one has to keep this objective (the flavor profile) in mind in order to evaluate any advantages with respect to other approaches. In addition, I think the principles are well aligned with the more active nature of whole wheat during fermentation. My sense is that most sourdough approaches popular with home bakers today are optimized towards producing a rich sour flavor using the neutral palette that refined flour provides (San Francisco Sourdough being the de facto standard in the US). I have somewhat of a "sour tooth" in general, but I'm in the crowd that feels the inherent flavors whole wheat grains bring with them are more complementary to a fermentation process that reduces (but doesn't eliminate) this acidity. There are various accounts of desem, but I believe the three English language references currently available are: Laurel's Kitchen Bread Book, The Bread Book, and The Bread Builders. There is some discussion of these books by PiPs in THIS THREAD. All of them trace back to the Lima Bakery in Belgium by way of the Baldwin Hill Bakery in Massachusetts. There are changes here and there, but they all share a common theme, which he summarizes well:
I recently picked up The Bread Book on Debra's recommendation, and I think it provides the most succinct explanation of the three. I started with JMonkey's post because of its simplicity and use of bakers percentages, but I have been reading all of these books and adapting the process based on my observations and the strength of the starter. They all share the common theme of maintaining a cool, stiff starter (favoring yeast:LAB) with frequent feedings (minimizing acid load that would be passed on to the dough) and a high enough inoculation during bulk fermentation to get the job done in a relatively short period of time (longer fermentations will eventually favor LAB). For example, Laurel's Kitchen Bread Book mentions a 7-hour process (4 hour bulk fermentation + 2 hour (or shorter) hot final proof + bake) from beginning to end. I believe this relates to the phenomena of LAB vs yeast lag, where yeast responds more quickly to a new food environment, but LAB eventually catches up and dominates in the end with enough time, which can lead to structural problems.
For a young desem culture, this may require a very high inoculation of 50% or so. As the culture matures, it can (and should) be reduced. In The Bread Book, Thom Leonard suggests building a leaven that is roughly 1/3 the size of your final dough, which is similar to the prescription in JMonkey's post. (I'm currently baking regularly and continuously build a leaven with just enough left over to feed the next one, so there is no distinction between the starter and leaven.) Tradeoffs associated with the use of a relatively high percentage of pre-fermented flour are discussed by Debra in THIS COMMENT.
Thom Leonard addresses this "balancing act" in The Bread Book:
He further explains:
My fermentation times have been decreasing and I have recently dropped from 30% to 20% PFF with noticeable improvements in dough strength later in the fermentation process, while still producing a pleasant mellow flavor with a subtle tang and mild sweetness. I did have some early versions that took much longer (> 9 hours) and they were noticeably much more acidic. It is desirable to maintain an inverse relationship between dough hydration and the percentage of PFF, and I have gotten into trouble with high hydration high percentage PFF doughs.
The underlying principles are covered in more detail in THE POST I mentioned previously. This Maximizing Yeast in SD Starter post is another one worth reading. It is worth noting that the same principles show up in many other bread baking traditions with various modifications. As JMonkey mentions in his post, this is very similar to a French Pain Au Levain made with whole wheat.
I've recently purchased a pH meter and will attempt to quantify pH of the starter and dough. According to THIS POST, it may also serve as a fairly reasonable quantitative measure of the final bread flavor to complement the usual photographs. I'm still learning and experiment with this myself, so please don't take me as an authority on the matter.
I really appreciate the thorough response. I'm captivated by the process and intend to pursue this further. Since I had already been working with 100% whole grain it will be interesting to see how the results compare over time. It seems that most who come to this method have not been satisfied with their whole grain loaves previously, which makes it hard for me to gauge their results as they compare to mine. As mentioned earlier I've started maintaining my "stiff" starter in the cellar in an attempt to strengthen the culture (not desem, but in that spirit). I've just mixed my first dough using a hard red I'm familiar with following my standard formula. It's early days obviously but I'm curious to see if there's any difference so far (bulk time in particular). Thanks again for the detailed response, it looks like The Bread Book will be my next purchase.
Yes, you seem to have a process dialed in that works well and produces bread you like, which was the main reason for my cautiously qualified response concerning "advantages". The desem process is one that seems to align with my preferences and has produced whole-grain loaves that I'm happy with (although they could always be better). I've largely leaned on testimonials from others to settle on the approach. I don't have a lot to compare it to, but did tinker with a more standard approach early on, so I'm excited that you are deciding to experiment with it.
I picked up the mill and started milling and baking early in the summer with the usual early stage challenges. Shortly after that, I moved to a cabin where I ended up baking from a large sack of Central Milling bread flour using the fairly standard approach of a warm (often hot and humid) bulk fermentation and retarded final proof. It worked like clockwork (once I got a cloche for the countertop oven) and I rarely changed anything. By late Fall I was back at home and eventually started trying to replicating the process with whole grain flour. I prefer what I'm making now, but I've learned a lot since then and can't attribute it entirely to process. It isn't clear to me that I couldn't produce a similar flavor to what I'm making now with a different approach.
Anecdotally, it seems the few whole-grain sourdough approaches with nice results often employ a cold final proof and emphasize extremely gentle dough handling and passive gluten development, more so than intensive active gluten development. In contrast, in the case of desem, a short, hot, and humid final proof is employed.
In Laurel's Kitchen Bread Book, they state:
It isn't exactly clear to me what she means by "light" here. Crumb terminology tends to be fairly loose, and I think I often conflate open and light crumbs. The few photos I've seen of desem loaves (including The Break Book cover) seem to be tall and springy with closed but light crumbs. I would like to better understand any inherent differences between warm and cold final proofing. I would expect a warm (95F) final proof to be highly favorable to LAB (especially at the later stage where LAB vs yeast lag is not relevant), which would increase acidity. Perhaps the acidity is low enough that the process is used primarily for texture and the added acidity at this stage is tolerable or even desirable. What would be the impact on the texture and flavor of a cold final proof here if everything else is held constant? Or maybe a warm final proof and quick chill to firm up the dough and improve scoring?
I think there are also other ways of achieving a yeast favoring starter that is friendly to whole-grain baking and their "active" nature, such as this 100% SPELT LOAF, which I believe exploits LAB vs yeast lag with a multi-stage young levain build. That is also something I'd like to try.
I do think approaches that use a retarded bulk fermentation and/or final proof are very schedule friendly. The desem approach works very quickly, so the timing windows are a bit more demanding, although using the two-day (evening pre-dough, morning mix) "epoxy method" helps.
The Break Book is 101 pages with a short 16-page discussion on desem bread baking and starter maintenance. It sells for $30 or so online, so at roughly $2 per page it is a very special interest purchase, but it is worthwhile for me given my current focus.
My theory is that in Laurel (and Allan Scott's) chilly northern California environment, they've need that super hot proof at the end to have the bread ready for selling at the market. In my Arizona environment I use a regular proof in the 70-80F and my desem turns out fine. The key is to get max rise with no acidity.
I agree with JMonkey that desem is just pain au levain made with 100% whole wheat, natural yeast, and kept at all costs from being sour. This mostly takes place through stiff starters, moderate temperatures, but can also happen with salt. I took a quick peak at my somewhat outdated Desem primer from 2001 (when I was still baking by volume); and I agree with what's been stated here and in the comments. I prefer my crumb a little tighter than the one photographed above, but that's just tastes. I have a photo of mine at the top of this Desem Primer
Though I bake lots of types of breads these days, I've kept the desem process for creating new starters. Soaking whole grains of wheat, mashing them into a pulp, and leaving them in cool temperature in a tight low hydration ball is my favorite method of creating a new starter. I find it easier, more forgiving, and more reliable than methods using white flour with a liquid consistency. The main downside is I have to wait till Arizona winters to get the 50-70 temperatures required, usually outside overnight.