This section is from the book "A Treatise On Beverages or The Complete Practical Bottler", by Charles Herman Sulz. Also available from Amazon: A Treatise On Beverages.
Unfortunately, the vanillin prepared from the vanilla bean, or that prepared from other sources artificially, does not always represent what it pretends to do. There is quite a variance in commerce. If pure and of proper strength, about 150 grains should give a tincture equal in aroma to the tincture of one pound Of best vanilla beans, and on this basis the proportions given in the appended Formulae are calculated.
However, experiments have proved that most of the vanillin offered to the trade is not of the strength indicated by the manufacturers. About one ounce of vanillin to one pound of good vanilla beans is the ratio of the commercial vanillin - that is, one ounce of vanillin imparts to a solution the same aroma in strength as the extract or tincture of one pound vanilla beans would do. This proves that the commercial vanillin is about three times weaker than it could be prepared, and also proves that the manufacturers impose on the consumers. We therefore warn the carbonators not to buy such inferior preparations. Try a sample. Prepare a tincture of vanilla beans, using one ounce, and macerate in eight ounces diluted alcohol. Also dissolve about ten grains of vanillin in two ounces of strong alcohol, then dilute, and add six ounces of water, to make eight ounces. Now compare the strength of aromas. An admixture of coumarin may sometimes be detected by its peculiar and different aroma. The carbonator may sometimes get surprised in detecting none at all, or but a faint aroma of vanilla in this artificial sample, simply because the vanillin is too weak. Do not buy such a product, and if no better one can be obtained, stick to vanilla beans, and make a true preparation.