We try to not only compare spectrum by its qualitative results differences, but we try to check the spectral efficacy of each, in terms of yield por unit light delivered (independent of watts used).
As for the moment, Ive checked that there is no an unique optimal spectrum. I would like to emphatize this. Not only some strians does better with some spectrum and others with different spectrum, sometimes with huge differences.
Same strain may do diferentially depending of the irradiance used. At low irradiances, spectrums with dominant red, some blue and only a little of other wl does very good (but some strains dont go fine into bloom this way). While at high irradiances, trend is very different, with spectrums with a good percentage of white LEDs (thus more green-yellow) does same or better. Never worse.
This fact has convinced me of the need of using a good number of white LEDs if wanting to have a LED lamp able to bloom in a variety of conditions. Its possible to use little white (and none!) on some conditions/strains, but in general it not work fine on a variety of conditions.
When I say white LED you can put perfectly any other broadband white light source, as fluorescents. Ive been unable to find any noticiable difference between results using white LEDs or fluorescents when percentage of white is similar.
I tend to use more white each day I check results using it or not. On some conditions is possible to obtain a better spectral efficiency using less white, but in general, more white means a guarantee of good results.
====
The main difference between LEDs and HIDs in this regard, and that makes them more similar to the lighting obtained by fluorescents, is that lighting achieved is way more evenly, meaning that there is small differences between the areas with max and min light density. Local light densities are closer to the average light density in LED grows.
This is a great advantage of LED lighting for growing, often overlooked. Over given light densities, which almost always we surpass by large on our bloom areas, plants goes reducing its efficacy using the light as the light density increases.
Thus for two grows with an average light density of 800uE/m2, one very uneven (HID) and one very even (LED), plants grows more for the even distribution. Because on first case one leaf is exposed to 200uE/m2, which is below the capacity of the leaf of process light optimally, and other at 1400uE/m2, at which efficacy processing light can be half of that at 800uE/m2. While the other setup with 600uE on one leaf and 1000 on other have higher efficacy, because less light is used below its potential. This effect is especially important if we manage to grow the plants using irradiances (local light densities) not over the point that plant process optimally (300-400uE/m2 for our loved plant).
====
The other reason is about light. LED spectums works with much higher efficacy than HPS ones at low irradiances. HPS growing only works fine at high light density levels, due its spectrum so rich on yellow-orange ask for it. With average densities below 500uE/m2, HPS usually are unable to get good buds. While with LED spectrums we are being able to get nice budding at 300uE/m2, and very good at 400uE/m2, with an efficacy of light near double.
But as we increase the light density of typical LED's spectrums, efficacy goes dropping. Its possible to grow with LEDs at high irradiances, but efficacy of light in this case, although still better than that of HPS spectrums at same irradiances, is far from being double. This is an empirical observation.
I believe its due the caracteristic of red photons of saturating photosynthetic systems of leaves, thus past some irradiance of red photons, they add little more. This happen too, at less degree, with blue, that anyway is not useful in large amounts to beat other spectrums as HPSs due its reduced photosynthetic efficacy, while green-yellow photons are way less affected and can continue increasing photosynthetic rate up to very high irradiances.
Practical consecuence is past some critical levels of irradiance, LEDs spectrums dont work much better than others, as HPS ones. While using half light we can get yields over 80%, if we want to yield same than a HPS, we need to use an amount of light near the one used for the HPS (say, 80%), and its expensive, apart of involve other problems, as the cited with elevated EC. And thats is distributing the light very evenly as we did, Im afraid that when using brick style LED panels, acheving same yield per sq meter than with HIDs will require to use near the same amount of light.