Measuring your Sake – Part Two: Percent Alcohol by Volume

Measuring your Sake – Part Two: Percent Alcohol by Volume

In Part One we talked about measuring the SMV (Nihonshu-do) of your sake. In this part we will discuss two ways to measure the percent alcohol by volume for your sake. Because Heikou Fukuhakkou (並行複醗酵) or multiple parallel fermentation is used to make sake, we cannot employ the same simple method for determining the percent alcohol as is used for other fermented beverages. In particular, the simplest methods used for both beer and wine depend on knowing the initial specific gravity prior to fermentation. For sake, there is no point prior to fermentation when all the sugar is available for such a measurement. Rather, koji enzymes work side by side with yeast in the fermenting mash. Enzymes create sugar and yeast creates alcohol using the sugar; this is multiple parallel fermentation.

The first way to measure the alcohol level we will discuss uses what is known as the boiling method. It is a modified version of the distillation method that can be done at home with relative ease. The basic idea behind this method is that the alcohol in sake has a known specific gravity and a known effect on the specific gravities of substances it is in. Given this we can measure the specific gravity of our sake to get an initial state, replace the alcohol with distilled water and then take a new specific gravity reading. The difference between these two specific gravities can then be used to determine the percentage of alcohol by volume.

The difference between these two specific gravities multiplied by 1000 is known as the spirit indication. If the specific gravity taken after replacement is SG2 and the initial specific gravity is SG1, then the equation is:

Spirit indication (SI) = (SG2 – SG1)*1000

The best formula1 I have found to give the percentage alcohol by volume (%ABV) is:

%ABV = (0.008032927443 * SI^2) + (0.6398537044 * SI) – 0.001184667159

Note: There is an equation that has been picked up on the net but is incorrect. I mention it only so you will not be confused by it. Please do not use %ABV =(SI/1000)/2.11*1000.

Jon Musther’s site at has an online calculator that will handle the actual calculations for you so you don’t really need to create your own if you don’t want to.

OK, so that was all well and good but how do you replace the alcohol with distilled water? The way this is done is to begin with the sample of sake, place it in a beaker, flask or other container that can be used to slowly boil the sake to close to ½ its starting volume. It is important that you can precisely return the level of sake to the pre-boil level by adding distilled water. The pre-boil amount should be close to 250ml in order to be able to properly take the specific gravity with a hydrometer. Recall from part one how we used the hydrometer to measure the specific gravity of sake.

Be careful while boiling. Sake, in a small flask, will very easily boil over. It will look like it is not boiling at all and then a series of bubbles will come up the side of the flask and out the top. It may then, also catch on fire. In addition you may have to start over because some of the non-alcohol materials that affect the specific gravity can be lost.

The reason for using distilled water is to ensure the water being added has a specific gravity of 1.0. If it does not the calculations will be thrown off and you will not get accurate results. Also, to ensure accuracy, the post-boil level needs to be exactly the same level as was there pre-boil. The degree to which you don’t get back to the same level drives a difference in the ratio of material that produces a deflection in specific gravity to water which will throw off the final measurement. If we do bring the level back precisely then we will have replaced the water boiled off with water and the alcohols boiled off with water. Hey, that is it; replacing the alcohol with water.

So, to recap, what we need to do is choose a sample large enough to take a hydrometer reading (SG1), place these sample in a flask and boil off ½ the amount, precisely back up to the pre-boil level with distilled water, take the second hydrometer reading (SG2), plug these values into the equation above and voila out pops %ABV.

While not overly challenging the boiling method takes a considerable amount of time. A much faster method that is still quite accurate is done using a hydrometer in concert with a refractometer.  As the hydrometer uses the relative density, refractometers use the relative speed of light as it travels through a medium. The difference in speed between the two medium is the refractive index.

Many inexpensive refractometers are used for determining the amount of sugar in a solution. They can do this because the refractive index changes with the amount of sugar in solution just as the density of the solution changes. These refactometers often use a scale marked in degrees Brix. Degrees Brix is the measure of sugar in a solution. 1 degree Brix means that 1% of the solution is sugar. A refractometer can be scaled in degrees Brix using a one to one mapping between degrees Brix and the refractive index. Solutions used for producing alcoholic beverages have other components that also affect the refractive index reading so in these solutions the Brix reading is not entirely percent sugar but includes other dissolved solids.

Inexpensive Refractometers

However, the refractive index for solutions that have both sugar and alcohol is more complex. For any given percent sugar + other dissolved solids, we get a mapping from the refractive index to percentage of alcohol. So to make sense of the refractometer reading we need to know both the specific gravity and refractive index. We can use a hydrometer to measure the specific gravity and then use the refractometer to estimate the alcohol. The chart below shows how a mapping for the refractive index to %ABV for various specific gravities can be done.

%ABV  by RI (in Brix Eq) given Specific Gravity
%ABV by RI (in Brix Eq) given Specific Gravity

The equation for this is given by:

%ABV = 1.646 * RI – 2.703 * (145 – 145 / SG) – 1.794

Refractometers like the one in the picture above have Automatic Temperature Correction (ATC). They are used by placing a drop of sake on the plate and closing the cover. You then look through the eye piece and read off the value where the shade changes colors.

After measuring both the specific gravity with a hydrometer and the refractive index (given in Brix equivalents) with a refractometer, plug these values into the above equation and again, voila we get the %ABV for our sample.

Both of these methods work well with the former taking more time and the latter requiring additional equipment (the refractometer).

In the next part (Part Three) I will discuss how to measure the Sando ( 酸度 ), that is the acidity of sake.

  1. I got this equation from Jonathan Musther at I don’t know where he got it or if he created it but it matches the table values from William Honneyman,B.Sc.,Ph.D. discussed in
Click to Enter a Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

10 thoughts on “Measuring your Sake – Part Two: Percent Alcohol by Volume”

  1. Hi Will,

    I am about to rack my first attempt at brewing Sake, so starting to look at how I will analyse the product. I just made some measurements of a store bought Sake to get a feel for the calculation. It is labeled as +5 and either 16 or 18% ABV (the printing on the label is not clear). The refractometer Brix reading is 10 and the specific gravity is o.999. When I plug this into your equation I get -23.2. I get that the right hand part of the equation gets numerically bigger with SG reading less than one, so I guess that part is OK and we should disregard the negative value. However, the value is pretty far off. Where do you think the error is coming from? I am guessing in the specific gravity reading, but if it’s really a +5, this value should be negative, no? I adjusted the temperature to 60 F and am using the standard beer/wine hydrometer. Is there some more accurate way of making this reading or are some hydrometers better than others?

    Thanks in advance for any advice you can give me.


    1. Hey Greg,

      Trying this out of a commercial sake first is a great way to get your feet wet. You know what it should be and if it isn’t you know something is wrong.

      When I plug your measured values Brix = 10 and SG = 0.999 to the %ABV and SMV equations I get 15.05833 %ABV and 1.444444 SMV respectfully. This does not agree with your outcome so I suspect that there is an issue with calculation order or grouping. Just to try to be more explicit I’ll write it out here with extra braces to show what needs to be calculated together:
      %AVB = (1.646 * 10 Brix) – (2.703 * (145 – (145/0.999))) – 1.794 = 15.05833
      SMV = (1443/0.999)-1443 = 1.444444
      The 15.05833 %ABV seems reasonably close if the bottle is labeled 15 or 16, but the 1.44.. SMV is quite low compared to the +5 on the label.


      1. Hi Will,

        Thank you. Math skill rusty, so thanks for the clarification. For the SMV, I think my value from the hydrometer may be high, because was not fully down to to 60 F. So, I think it is all good to go.

        I was thinking of trying a very subtle dry-hopping of a fraction of my product. I think it could interesting if it doesn’t dominate or clash with other features of the aroma profile.

        What do you think? Crazy idea? Do you know of anyone doing this before?

        Best, Greg

  2. Hi Will,

    Thank you so much for this.

    I’m an absolute beginner and playing around a bit. So, please excuse the simple question:

    In the equation you’ve given, where do your numerical constants come from? – %ABV = 1.646 * RI – 2.703 * (145 – 145 / SG) – 1.794



    1. Hey Robin,

      Ultimately they come from a curve fitting exercise with known quantities. However, that is not where / how I got them. I looked to authoritative sources in books, papers and other websites that have this information. I recall that the values in this article were the same from several sources I was using at the time of writing, including which I reference in the article. Hope that is helpful.


      1. Ocht, you’re the man!

        I always like to know the fundamentals so I can try to diagnose if this start to look weird.

        Thanks a million!


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.