Examine the following table.
Note that every composition contains strontium nitrate (Sr(NO3)2). That is because the strontium is responsible for either a red or red-orange color. However, strontium nitrate is not a great oxidizer - a mixture of Sr(NO3)2 and a fuel will hardly burn. So a better oxidizer such as potassium perchlorate or potassium nitrate or an energetic fuel such as aluminum or magnesium is added to give the extra energy needed for a fast combustion.
The rest of the components are either fuels, or a compound that serves as acombination fuel and binder, or parlon which is a binder but not really a fuel.
Some older flare formulations also had things such as pitch, asphalt, wax, tallow, potassium chlorate and black powder. Those are not likely to be part of modern flare formulations.
The flares are pressed dry.
Separation of the strontium nitrate from the other compounds is not trivial. In the following analysis let us assume we have extracted a kilogram (2.2 pounds) of road flares, and we have 1 liter (about a quart) of a concentrated solution of the salts. Since dextrin is not a common ingredient, let us assume that there is no dextrin. If dextrin is present it will make the separation much more difficult. If one cooled the solution to 0° C, up to 600 grams of the strontium nitrate would remain in solution, and a lot of the other salts would precipitate out. Only 7.5 grams of the potassium perchlorate would remain in solution, but a full 130 grams of potassium nitrate would remain in solution. If the crystals were removed by filtration, and the solution evaporated, one could obtain a material that might have the following composition:
Please note that the analysis above assumed that the liter of solution was properly concentrated to maximize seperation and retain strontium nitrate. Imagine what would happen if one used sub-optimal conditions. For example, if one used a flare with a composition similar to flare #3 (60% strontium nitrate and 23% potassium nitrate) and didn't make a properly concentrated solution.
Let us say the solution is under-concentrated: If the solution contained 500 grams of oxidizers, then it would contain about 361 grams of strontium nitrate and about 139 grams of potassium nitrate. When that would be cooled to 0°, then 9 grams of the KNO3 would come out of solution, and the final solution would contain 361 grams of strontium nitrate and 130 grams of potassium nitrate. The product would be about 73% strontium nitrate and 26% potassium nitrate (rounding errors).
On the other hand, if the solution is too-concentrated: If the solution contains 1100 grams of oxidizers, then it would contain about 790 grams of strontium nitrate, and 310 grams of potassium nitrate. When that solution is cooled, it will crystallize out BOTH strontium nitrate and potassium nitrate, giving a final solution with about 650 grams of strontium nitrate, and 130 grams of potassium nitrate. Than means the salt obtained by filtration and evaporation would be about 83% pure, BUT you have lost 140 grams of strontium nitrate (it's mixed with 180 grams of potassium nitrate crystals which you filtered out). Recovering that is possible, but very difficult.
So, the bottom line -
You will never get a pure product though a simple single step
If you don't do every thing right, you will get a very impure product or
you will end up losing some of the strontium nitrate.