We usually refer to the F number to compare how fast is a lenses with another one. I understand that the optic quality should help as well — so would a L series in Canon (or whatever is the equivalent in Nikon) be faster than a non L equivalent? For example, we could consider the 100mm macro in Canon which exist both in L and non L. What about the Canon 24-104mm f4?
Would I gain a whole stop with an L lenses?
No, it won't be anything like a full stop, but there is probably a small advantage.
Our friends over at DXO mark test transmission of lenses, though only wide open, and they don't have a pair of L non L lenses at the same speed. What the results do show is that the wider the max aperture the more transmission you lose compared to the theoretical f stop. So the 50 f/1.2L is about a fifth of a stop further away from the theoretical value wide open compared to the 50 f/1.8
Since manufacturers (of SLR lenses) don't release t-stop information the data is scarce so most photographers refer only to the f stop value when they talk about lens "speed". Cine lenses are often specified by t-stop as this is more important to film since you can't (easily) change the shutter speed.
Old answer:
You are right that the f/ number tells you how fast a lens is. The lenses designated "L" tend to be faster, but that's because of the lower f/ number, not the L!
Case in point, the EF 100 f/2.8L is not faster than the EF 100 f/2.8 as they have the same f/ number. However the EF 35 f/1.4L is a stop faster than the EF 35 f/2, because the f/ number is smaller.
The f/ number relates to how large the aperture is in relation to the focal length. f/2 means the aperture diameter is the focal length divided by 2. Wider aperture means you let in more light and get more background blur which is generally desirable.
Being a stop faster means you let in twice as much light, which means the area of the aperture is twice as big. Hence if you square the f/ numbers and one is twice the other one, the smaller one is a stop faster.
Once a (fairly) long time ago, I found this interesting, so I did some testing. In my case, I was less concerned about whether something was labeled L-series, but about primes vs. zooms. Given the number of elements in many (most?) zoom lenses, I'd guess the light loss is even greater than differences in glass could account for.
Due to the timeframe, I was doing the testing on Velvia -- high contrast slide film that was notoriously picky about exposure, which should have maximized visibility of exposure differences.
The results varied somewhat. In a few of the best cases, I couldn't see any real difference between the zoom and the fixed-focal-length lens at all. In one case (an ancient hyperzoom, something like 50-250) there was probably about 1/4-1/3 stop loss at the center.
Toward the corners of the frame, the differences were much larger. In fairness, however, this probably doesn't mean all that much -- to get to the same apertures, the primes were stopped down quite a ways whereas the zooms were being shot within a stop or two of wide open. Greater light falloff when the lens is wide open (or nearly so) is extremely common. Zooms are generally worse in this respect than fixed-focal-length lenses as well.
The only way that's likely to apply in the case you care about is roughly the same one: at least in some cases, an L-series lens will be faster so to get the same aperture it'll have been stopped down further, which will usually reduce the light falloff at the corners. OTOH, this is primarily an issue when shooting full-frame -- if you're shooting APS-C, it's much less of an issue.
In short, as MattGrum says, no. Compare F-stop numbers for the amount of light (it's very close to the T-stop number which is the actual measure of the amount of light).
However, one might consider an L lens to buy you an extra stop if you're using lenses stopped down to get yourself a sharper image (i.e., instead of shooting at the widest aperture, you increase it a stop or two so that the image is sharper — not to be confused with increasing aperture to get a higher DOF).
A cheaper kit lens (e.g. Canon 18-55 f/3.5-5.6 IS USM) with a maximum aperture of, say, f/4 at a given focal length, will be much sharper if you stop it down to f/5.6 and slightly sharper again at f/8. An L lens in a similar focal range (say the 24-105 f/4L or 17-40 f/4L) will be sharper than the cheaper lenses for the same focal length and aperture, probably requiring one fewer "stops down" to get the same sharpness. This is a rule of thumb of course, check SLRgear for actual blur index measurements from 18-55, 24-105, 17-40.
The light transmission level may be better or less than claimed f-stop.
Take the 24-105mm f/4 at 105mm really transmits light like a f/5 lens. http://www.dxomark.com/Lenses/Canon/EF24-105mm-f-4L-IS-USM-mounted-on-Canon-EOS-5DS-R---Measurements__1009
Light loss may be a factor of multiple coatings, number of elements, vignette and marketing. Also, there is a large light loss apertures wider than f/1.4 related to the sensor micro-lenses over each pixel. So much so some cameras will 'adjust' the ISO sensitivity silently to compensate. http://www.dxomark.com/Reviews/F-stop-blues
Basically, don't expect a lens to be better just because it's more expensive. It may be designed to be better in one specific area and the resulting optical formula may actually be worse in others.
The optical quality of a lens relates more to its resolving power and (lack of) distortion. An L lens is not faster than standard lens in traditional sense, but it is certainly better because it lets the sensor capture the image at better resolution and with less distortion.
You would gain 1 stop only if for the given loss of light in the standard lens at a particular stop, its output matches a stopped down L lens. This is not true – you can try both the lenses on your camera in a shop to understand this behaviour.
Visually an MTF chart is a good summary of the optical quality of a lens.
