Research shows that spectrum has little to do with yield. As the percentage of blue photons increases, yield decreases (see attached).
In terms of LED's increasing yield, I think the attached paper documents that HPS results in higher yield but the cost of running an HPS over N years is so much higher than an LED light that using HPS can't be justified on a cost basis.
The main issue with spectrum its impact on plant morphology—blue makes plants short and compact, red "encourages growth" (red diodes are very electrically efficient), and far red tends to increase stem elongation so it can help offset the impact of blue photons in a "white" LED.
Many LED are sold as "HPS replacements" and they have a PPFD map with a lot of light in the center but a rapid fall off once you move off center. That allows a company to create a "750 watt" LED at a low price. Growers coming from HPS don't know about PPFD maps so they don't know the difference.
Companies light Spider, Mars, and Migro (and Growcraft, before they went belly up), put the engineering into creating lights that have a more even PPFD map. The engineering and the additional diodes are cost drivers but they, unquestionably, can produce a higher yield.
Most growers don't get the extra yield because they don't give their plants enough light to maximize yield (bud quality and yield quality increase as PPFD increases, though they max out at lower levels). That's a huge issue—if you want to just grow some weed, a lower priced light, with. "hot spot" PPFD map will save you a few hundred dollars. If you want to go for a light with an even PPFD map and will turn up the dimmer, assuming that your grow environment is sound, your chances are very high of at least meeting the seed seller's estimates, which tend to be in the 450-650gm/sqare meter range.
There is no magic to that. We know how cannabis reacts to light, in terms of yield. There's scads of research on that and the results are consistent. It takes 9 photons to make 1 molecule of carbon so if you only provide a modest number of photons, a cannabis plant cannot grow as well as a plant that gets a lot of photons (as long as light is the limiting factor and all other things being equal).
The Frontiers paper is a few years old and is unusual because the research was done in ambient CO2.
To your question—I think in the "Decreasing Blue Photons…" paper HPS will produced more weed but the electricity required is much higher. (I've never even entertained the idea of using HPS.)
I grow for crop yield, crop quality, and bud quality. My buds plants are big, the yields are high but the buds are not the small, cute bud of the month buds. If I was growing in a 4' x 4' tent, in ambient or enhanced CO2, I'd go with a a pair of Spider Farmer G4500's, which are for 2' x 4' tents, and would add a set of Spider Glow R80's to the G4500.
The two light setup gives you some redundancy, they're light and easy to maneuver, and they allow you to improve the light cast because you can set the lights at different heights. The R80's add 80 watts of 660nm light which helps boost the percentage of red light. I've run a few sets of lights through chatGPT and the spectral balance of those lights is really quite something.
I added the R80's to my Growcraft X3 flower light last year and got just the light I was looking for. If I were to continue to grow, I'd DX my Growcraft and go with the G4500. I simply can't find a better combination of spectrum (to shape the plants), PPFD output, and PPFD map unless I go to something like the behemoths at grandmaster.com which are far more expensive (though I'd go with them if I was going with CO2 and wanted to put that kind of money into a grow setup).
Long answer but we're all well served by a significant discussion.
