p4: A Study in White – Vanilla Ice Cream (Roasted Banana Style), Milk Caramel, Raw Apple Garnish

What originally brought me back to the interest of the Fat Duck cookbook was in fact a search for an ice cream custard recipe; in particular, one that didn’t use an cream, and ideally had some skim milk powder in sense I’ve got a leftover bag after my Sweet Bread creation. This was mainly prompted by the fact that I finally figured out how to get my ice cream machine cold enough at home (used to use it all the time in the dorm, the freezer there was PERFECTLY cold), so I can now put together some tasty custards and sorbets once again. Before now, I’d always have to use Heston’s trick of pouring crushed-up Dry Ice into a stand mixer with the cream base if I wanted any hope at a properly frozen and churned dessert.

Since I forgot my all-purpose, no-cream recipe that I used before (btw, I don’t use cream because that just means I have to go out and buy this pricier dairy item every time I get a whim; so much easier just to use milk and eggs that are at home); but I remembered attempting the Heston ice cream before, sans milk powder and when the freezer sucked. His Vanilla Bean Ice Cream perfectly fits all my requirements, and I can’t think of any better fun item to start off this project.

20150807_102920I did find two simple toppings to try out as well, both apple themed since 1: I’ve got a few in the fridge, it’s what works best at the moment and 2: the dish this recipe is originally used in is the ‘Cox’s Apple,’ and highly modernized and ultimate version of Heston’s experiments and interest in Tarte Tatin apparently. I actually wanted to do 3, but one of the ones I was dead-set on doing required 14 hours of sous-vide cooking, which I did NOT have the time to reproduce this last-minute; and another sorta-similar-resulting recipe used Pectin, a no-go for me right now too. But I do think I can get some somewhere…

20150807_102940The first is an ‘Apple Milk Caramel,’ a seemingly uncolored ‘caramel’ sauce made mostly from milk that has steeped from leftover apple cores and skins. The other, and yes I feel cheap even mentioning it, is ‘raw apple dice.’ But it’s not just cutting them up and garnishing, you’re supposed to preserve them in an acidified syrup of sorts before plating, but with the limit that it can’t be sit any longer than 2 hours. I’m curious to see if this actually has an effect on the apples and their flavor, or is just a high quality version of letting them sit in lemon-water.

What I’m Doing Wrong

First off, I’m going to go ahead and say it: I’m not using Vanilla Beans in this. As good and delicious as it would be, the fact is these guys are pricey, even if you find some ‘good deals;’ I only use them when I really want to make something special for a good occasion. This would certainly be the kind of recipe I’d do it with though; pure, unadulterated custard that shines on its own. The only effect of leaving this out should just be the lack of the vanilla’s flavor in any case; though it will also mean that I don’t need to perform the long steeping time to infuse the beans into scalded milk.

Counter to this, I’m also gonna use the opportunity to get rid of one of many VERY ripe bananas that I have. I’ve read a recipe on ‘Roasted Banana Ice Cream,’ the base of which is the same as a typical cream, only with the addition of a ripe banana that’s been roasted to develop its flavor, so I figure it shouldn’t negatively impact the final texture; or if it did, in a minor way.

20150807_213454Not like I expect to get a perfect texture either way. At the end of the day, I’m making this in a countertop ice cream machine; the book itself does state ‘churn in your machine until it reaches 23F’ (though I’m just going until it looks complete; I know enough about my machine to avoid ANY chances of interrupting the process and introducing warmth that I don’t ever want to stop it until it’s done), but the only way it can get as super smooth and perfect as his usually does is by either using liquid nitrogen, dry ice, or one of those awesome industrial machines. You know, the ones you see on Chopped and Iron Chef that they can put STEAMING HOT custard from the stove in directly and it freezes it within like, what, ten minutes? Those are just unfair. I HAVE developed at least one technique to ensure that my ice cream has the best chance to turn out without having to worry about the frozen container warming up too much on the counter before it’s frozen my mixture. And that is, of course, actually putting the machine in the freezer AS it’s churning the ice cream; one of the few advantages of having a small, dinky countertop ice cream churner!

Final change in the ice cream comes through the requirement of using ‘Fructose’ as the sugar source. I’m sure Heston has a tub of pure, neutral fructose to use for this; I do not. Instead I go for Honey, which is completely or almost completely fructose anyway, just with some added flavor. It works out well, since along with a banana ice cream I’ve been wanting to do a honey one too since I’ve got quite a bit leftover from my Nougat adventure. (Last minute note, I am now realizing that the fructose used in the book might be an actual powder; in which case, I may have fudged up t he solids-water ratio a tad)

20150807_215330Moving on, I’m also using honey in the ‘syrup’ for the raw apples; also, it calls for Lemon Juice and a bit of Vitamin C. I obviously don’t have the latter, and not going out to get some for just a measly 1 gram, and I can’t believe there’s no lemons in the fridge! No worries, an easy substitute for lemon juice is using ½ as much vinegar, and I just happen to have one from my work made from Calamansi! It’s a citrus fruit from the Philippines, tastes just like lime and tangerine juice, it’s perfect! It’ll help show any distinctive effects that might be ideally present.

By the way, as you’ll see in later pictures, I tried dicing the apples to a couple different sizes, mainly to see if it made any difference (though really because I wasn’t sure exactly how big 5mm was).

I also didn’t have any Malic Acid to use in the caramel, again a measly 1 gram (and note, I’m also reducing the volume on each of these by a lot, so less than that), but I ended up having more apple skins and core than required, so I left it like that in hopes they’d give a bit more flavor and potentially some natural malic acid to fulfill what may be needed by it.

Cool Science-y Stuff

If there’s anything important to ice cream creation, it’s two fiels: Proportion and Freezing. The latter is easy and understandable; the ideal temperature, and apparently the lowest one can get it to anyways, to freeze ice cream is -5C/23F for getting the fats and liquids frozen, keeping it moving constantly so that 1: the ice crystals don’t sit and grow into large structures that often create those ‘icy’ textures, and 2: incorporate some air for texture, the amount of which one gets in and thus increase the volume of the final mix is known as ‘overrun.’ Often the ideal result with most quality frozen custard is ‘100%’ overrun, or about doubling the size, whereas many mass-production frozen products increase this to almost insane degrees, giving you ‘more ice cream’ for the same, now cheap price. But generally, the colder one can get the freezing mechanism, and thus the faster this freezing can get done, while churning as quick as possible, creates the super tiny evenly-distributed crystals while getting SOME of that little moussey and creamy texture. Thus why the best ice creams, and all those in the Fat Duck, are usually frozen with Liquid Nitrogen while being whipped fast in a stand mixer.

20150813_003036But in order to get that ideal final texture, one requires an ice cream base that actual has the right molecular ratios to yield the result; it won’t even freeze if it’s a solution that just plain CAN’T, and when it finally does after certain forcing believe me it won’t be good. When looking into this, the one thing I actually learned, to my surprise, is that the percentages of Milk Solids, Fat, Sugar, and remaining Water/Liquid isn’t quite as strict as I imagined. There’s certainly an acceptable range of course, but different frozen-churned desserts have different proportions that affect the final flavor and consistency, allowing one to actually play around a bit. Milk Solids, which helps with air incorporation an fat droplet stabilization, range between 8-13%; overall solid matter averaging 30-45. Sugars actually help add to the total of solids, other than that the only real effect is flavor; having too much mainly affects the solid balance, which is why it’s harder to freeze in a proper consistency. The ideal range is usually quite narrow, 14-17% of the total but Gelato can go up to 24. Finally, Fat stabilizes, since it melts so slow; that and affecting how rich it gets, though interestingly enough also having an inverse effect on flavor release. It’s slow-starting, but lingers longer, while-as low-fat creams burst their flavor faster and for shorter periods, but consequently melt faster. One’s choice of percentage can range widely for these, from as low as 3 to as high as 20. Note I’ve made no inclusion of figures for Sorbet, Sherbet, or Milkshake ranges in any of these

To give one example, traditional Gelato keeps a Fat% of 3-8, Milk Solids 8-11, and Sugar 14-24, combining (give or take) into a final Sold% of 32-42. The book has an awesome table listing these for different styles of ice cream, along with typical % for commonly used, and Heston’s favorite chemical, ingredients. I took a picture and listed them for reference on the side, hopefully you can see all right. I know in the future I’ll be using this to calculate all my experimental ice cream recipes.

Also learned that different sugars, though using the exact same weight, have different affects on level; pure glucose is actually sweeter than sucrose, table sugar, a mix of glucose and fructose. Of course they also affect flavor.

Heston’s recipe, it should be noted, was crafted after much trial and error to be create a very clean-flavored, low-custardy yet ‘creamy’ result that allows additional flavors to shine, mainly so that it can be served with other food and not overpower it. Thus why it focuses on all milk, improving the texture with milk solids that will increase the dairy flavor and consistency without the strong fatty flavor/feeling.

‘Vanilla’ Ice Cream, Roasted Banana version
20150807_1031521 Banana
(1 Vanilla Pod)
333ml Whole Milk
60g Egg Yolks (about 4)
32g Honey (Fructose)
14g Skimmed Milk Powder
(3 coffee beans)


  1. (Cut Vanilla pods in half, scrape out seeds. Put pod in pan w/ Milk, simmer gently for 10 minutes, remove and cool to 140F)20150807_111012
  2. Poke Banana, unskinned, with toothpick, and place in 350F oven for 40-60 minutes, depending on ripeness. Peel and reserve20150807_105727
  3. Whisk Yolk and Honey for at least 5 minutes to a pale yellow, foamy consistency20150807_110039
  4. Add in 140F scalded Milk, whisking to combine, followed by the Milk Powder (and Coffee Beans)20150807_110739
  5. Heat to 160F, stirring often, and hold for 10 minutes to pasteurize
  6. Add in banana, mashing and whisking until fully incorporated20150807_111842
  7. Strain custard into a separate bowl, ideally over ice bath, and leave to cool20150807_112646
  8. Transfer to covered container and let mature in fridge 8-24 hours20150807_213359
  9. Churn in desired machine and fashion until 23F, or set and airy. Transfer back to container and move to freezer to store20150807_215759

Apple Milk Caramel
166g Whole Milk
50g Sugar
50+g reserved Apple Cores and Peelings
(1/3g Malic Acid)


  1. Bring all the ingredients to a simmer, then let sit off heat for 30 minutes to infuse20150809_180850
  2. Strain through a fine sieve into pan, return to hit and bring back to a simmer20150809_184025
  3. Let cook, knowing that it WILL curdle, ‘until refractometer shows that is has reached 74 Brix,’ or until the liquid get to a desired thick consistency when cooled but does NOT brown or color20150809_190227
  4. Strain through fine sieve again into storage container, move to fridge until use20150809_191825

Raw Diced Apple
40g Honey (Fructose)
75g Water
0.5g Salt
3g Calamansi or other Citrus Vinegar (5g Lemon juice+0.5g Vitamin C)
1 Green Apple (Granny Smith ideal actually)


  1. Combine Honey, Water, and Salt in pan, heating until dissolved20150807_215601
  2. Cool over ice or in fridge, adding in Calamansi Vinegar (or actual Lemon Juice and C)20150809_175737
  3. When ready, peel and dice Apple to 5mm brunoise, adding to sirup for, at most, 2 hours until needed20150809_180026


20150809_193307Really, this sort of ice cream truly does need a fast-freezing method; with its higher milk focus vs eggs, a trait that another blogger prized since they were looking for a less-custardy option (they even thought THIS had too much yolk, crazy person), it still ended up icy with my machine. Still better than my previous ice cream attempts. In my situation, I think next time I’ll use more yolks to improve texture; besides, I much prefer and look for a more custardy ice cream for my all-purpose base. That said, I see the appeal to this mostly-milk formulae, the lack of custard really helps it pick up and express other flavors. That roasted banana came out pure and simple, and I bet that vanilla would have been glorious if used. Can’t wait to use it again, also to actually better see the effect of the milk powder in comparison to something I’m more used to.

20150809_193423Now, onto other things. I will say the flavor of that syrup DID come through a bit with those apple, at least when only briefly drained, and as expected moreso with masses of smaller pieces vs larger, so not a surprise. However, I WAS using some more distinctly flavored ingredients in it, and I expect that, if following the original recipe, those would actually be neutral, the use being to provide a solution for the apples to sit in that is of equal water-sugar-acid ratios to the fruit itself, not diluting or changing its flavor. So at least I know that; oh, and I now have some leftover syrup that I’m sure I can do something tasty with.

I’m pretty sure I didn’t cook the ‘caramel’ as long as I needed; I would have returned it to the pan to keep going too, but I barely had any and there wasn’t much time until my 2 hours on the raw apples were up. Besides, it was saucy enough and the flavor shouldn’t have actually changed anymore, so I still got a solid impression. And overall what I tasted I liked; they were right, very lactic and sweet and nice, like a dulce de leche that hasn’t caramelized… oh wait…

The white chocolate of the milk caramel world, the back of my mind is quite tingled with the idea of introducing this to some other desserts in the future. Problem comes through the fact that it can’t be paired with any strong flavors; after over an hour of cooking, long waiting to end up with only a quarter-ish cup of sauce, I could barely taste it alongside my banana ice cream; then again, that might have been the more icy and distinct results of my frozen attempts. Not to mention that, unless REALLY looking while eating as-is, I couldn’t really tell there was any apple in it. My guess was that being a result of the apple leftovers I used; the greens, which I love eating cuz they’re so crisp, probably don’t have that strong aromatic flavor in their flesh and skins like the best cooking apples do. And maybe if the sauce thickened, perhaps even reduced, a bit more, that may have helped.

But despite my negatives, there are plenty of positive results, and I understand these recipes and how I’d want to apply them in the future, and isn’t that the real goal of trying out new things in the kitchen? I know I can’t wait until my next casual ice cream, just to have more sweet deliciousness in the freezer, and finding an excuse to retry the milk syrup. Apples are apples though, I’ll just cut those up right before I need them next time.

20150809_192031Oh! One extra discovery! After draining out the caramel from the ‘curd,’ I actually tasted some of the solidified milk, and it was quite interesting. It actually tasted and had the texture of fresh cheese, but notably sweet and desserty; which is basically what it was. The texture isn’t super ideal, a little ‘springy’-er than I’d want, but I think I can do something with this, mixed with apples and nuts in a streusel maybe…