Plants are designed to grow in the SUN... Any artificial light that comes closer to that will give better results.
High pressure sodium works because of it's raw power and efficiency. It concentrates nearly all it's energy as light into that ugly orange color peak around 589nm wavelength. Plants only absorb that type of light moderately, because it's not in the "butter zone" of photosynthesis. It is however, fairly close which is why it's used for flowering, and like I mentioned, it's raw power makes up it's lack of absorption.
CMH lamps such as the Philips MasterColor series give off a little bit more red then the Hortilux Blue. In my opinion, CMH bulbs have the highest percentage of light absorbed by the plants when compared to all other grow light technologies. They are less efficient then HPS, more of their light is useful light within the 400-450nm blue and 600-700nm red range.
It's mathematically calculable how much useful energy to a plant each type of lighting technology gives out. I did the calculations;
I compared an EYE Hortilux Super HPS 400w bulb to a Philips MasterColor 400w CMH. The HPS bulb gives 55,000 initial lumens for 137 lumens/watt and the CMH bulb gives 34,800 initial lumens for 87 lumens/watt.
Assuming all the factory specs, and rounding lumens down, I calculated the relative energies at 10 nanometer wavelength intervals from 400nm-700nm within the visible light spectrum; A total of 31 different measurements. I then compared Clorophyll A and B photosynthesis at each wavelength interval, and ran the numbers.
In the end it turned out the HPS bulb had 12,318 lumens worth of useful photosynthetic penetration. This beat the CMH bulb with 9,903 lumens worth of useful photosynthetic penetration; The HPS beating the CMH bulb by roughly 124% overall. Predictably though, the CMH had a higher percentage of it's light used for photosynthesis, 28% of it's light used versus only 22% from the HPS.
Looking at a narrower scope, just veg/flower colors, the HPS is 141% better at flowering then the CMH, producing 4935 useful lumens within the 600-700nm "butter zone" compared to the CMH's 3478 lumens. The CMH had the HPS beat though for vegetative stage, 850% better at the 400-450 blue spectrum. The HPS only mustered 416 useful blue spectrum lumens, compared to the CMH 3514 lumens. Keep in mind though, this is a horticultural bulb with an enhanced spectrum, standard security light HPS would lose hands down to CMH.
The conclusion? It's still not clear. If you had to use a single bulb throughout the whole grow, the CMH may have the advantage because it handles the veg. stage so much better, while only trailing HPS during flowering. On the other hand, if you have the option of swapping out bulbs, you'll want to use the CMH for veg. and HPS for flowering. If your looking for the best possible light conditions, I would suggest one of each. What's interesting is the normal metal halides can easily push 110 lumens/watt. If they are to improve CMH technology in the future to make it more efficient while keeping the same spectrum, at that level it will beat out the HPS in all circumstances.
My personal opinion about CMH bulbs; I think they're the best artificial light source for growing yet, beating out HPS. The balanced spectrum plus UV light emitted from CMH bulbs has additional, unmeasured benefit for flowering cannabis. I'm convinced, but until someone compares the two in laboratory conditions a lot of people won't be. Your welcome to make your own conclusions.
If anyone wants to see my spreadsheet with all the numbers let me know i'll post it.