Four Thirds lenses do start decreasing in resolution after f/5.6, but they are not "diffraction-limited" by f/5.6, not if by that you mean "they perform at the diffraction limit there and beyond" (although maybe diffraction is the main cause of decreased performance). I actually calculated some data on his from dpreview's tests of the Olympus Zuiko Digital 50/2 macro lens, which I own (I had to convert their numbers to lp/mm, of which here are the results):

**Olympus Zuiko Digital Macro 50mm f/2**

*f/stop (4x5 equiv) -- resolution -- theoretical diffraction limit -- % Diffraction Limit*

f/2 (f/14.2) -- 92 lp/mm -- 750 lp/mm -- 12%

f/2.8 (f/20) -- 112 lp/mm -- 536 lp/mm -- 21%

f/3.3 (f/23) -- 119 lp/mm -- 455 lp/mm -- 26%

f/5.6 (f/40) -- 116 lp/mm -- 268 lp/mm -- 43%

f/8 (f/57) -- 105 lp/mm -- 188 lp/mm -- 56%

f/9 (f/64) -- 96 lp/mm -- 167 lp/mm -- 57%

f/11 (f/78) -- 87 lp/mm -- 136 lp/mm -- 64%

f/16 (f/114) -- 72 lp/mm -- 94 lp/mm -- 77%

f/22 (f/156) -- 54 lp/mm -- 68 lp/mm -- 79%

some similar large format lenses for which there is data (I've averaged the results for center, middle, and midline): http://www.hevanet.com/cperez/testing.html/

**Schneider G-Claron 305mm f/9** (equivalent to 43mm f/1.3 in 4/3rds for AOV and DOF)

*f/stop (4/3rds equiv) -- resolution -- theoretical diffraction limit -- % Diffraction Limit*

f/16 (f/2.3) -- 63 lp/mm -- 94 lp/mm -- 67%

f/22 (f/3.1) -- 55 lp/mm -- 68 lp/mm -- 81%

**Fuji C-Series 450mm f/12.5** (equivalent to 63mm f/1.8 in 4/3rds for AOV and DOF)

*f/stop (4/3rds equiv) -- resolution -- theoretical diffraction limit -- % Diffraction Limit*

f/12.5 (f/1.8) -- 40 lp/mm -- 120 lp/mm -- 33%

f/16 (f/2.3) -- 43 lp/mm -- 94 lp/mm -- 46%

f/22 (f/3.1) -- 43 lp/mm -- 68 lp/mm -- 63%

f/32 (f/4.5) -- 41 lp/mm -- 47 lp/mm -- 87%

Some interesting things worth nothing:

(1) By f/22, 4x5 lenses can perform as well as 4/3rds lenses at f/22. Look at other lenses in the link I provided, some of them are getting around 50-55 lp/mm at f/22.

(2) However, at f/22, large-format lenses that may be performing at 63-81% of the theoretic diffraction limit are equivalent to f/3.1 on 4/3rds in terms of DOF. Around there, 4/3rds lenses are getting 21-26% of the theoretical diffraction limit (actually, not all 4/3rds lenses...the Oly 50/2 is one of the sharpest lenses ever made).

(3) For portraiture work at f/2 on Olympus and assuming 5 lp/mm is desired for final print sharpness, the Olympus 50/2 can produce an enlargement of 92/5 = 18.4x enlargement over the 17.3 x 13.0 mm sensor size, for final prints of 9.4 x 12.5 inches. This is about on par with my experience of what Olympus can produce.

(4) For portraiture work at f/16, which has equivalent DOF using the same composition as f/2 in 4/3rds, using say the 305/9 G-Claron, you can produce an enlargement of 63/5 = 12.6x over the 3.75 x 4.75 in film area (the edges are unusable due to the loading notches) for final prints of 47 x 60 inches! Now, granted, you'd need a lot accessory light for this. Also, most people don't blow portraits up that large, but even if not, you'd still have a much greater tonality and smoothness. (and if the pictures were *too* sharp, you'd have plenty of room to do some blurring to eliminate unwanted facial detail and provide a smoother image instead).

(5) If you want to shoot with less intense light, there are the Xenotar 150/2.8 and 135/3.5. The Voightlander Apo Lanthar 300/4.5 has a longer focal length and some may prefer it for portraiture -- but its $5,900 on eBay right now (and it covers up to 8x10, so may not be sharp enough for 4x5)! In any event, assuming my Xenotar 135/3.5 for portraiture work wide open, it gets 20 lp/mm at the center wide open and 10 lp/mm at the edge for 15 on average. That's an enlargement of 3x over 4x5 for final print sizes of 12x15. Somewhat larger than 4/3rds and much much narrower depth of field (equivalent to f/0.5 on 4/3rds), along with much much greater tonality and smoothness. By f/8 (equivalent to f/1.1 on 4/3rds) the 135/3.5 is already getting 45 lp/mm and you can make 34 x 43 in prints!

(6) For landscape work, unfortunately 4x5 lenses don't stop down below f/64. You won't be able to match the extreme DOF you'd get with 4/3rds at f/22, although by then you could think about pinholes or putting a pinhole aperture in front of a 4x5 lens. In any event, because 4x5 has the ability to tilt the standards, I've often achieved f/32 or even f/22 for landscape work (f/11 for some architectural work). But even in the worst case, at f/64, 4x5 still excels. I haven't seen tests, but from graphs I've plotted of resolution of 4x5 lenses vs. f-stop, I'm pretty sure 4x5 is going to resolve just about at the diffraction limit at f/64 (heck, according to Kerry Thalman and Mike McDonald, 4x5 lenses are already performing at or nearly at the diffraction limit at f/32 and the data backs that up; see my link and the link to discussion).

Anyhow, at f/64, the diffraction limit is 23 lp/mm. If 4x5 lenses perform near there, that's a 4.6x enlargement over 4x5 for 22 x 17 in prints. How large a print can we make from 4/3rds at f/9, equivalent to f/64 on 4x5? The Oly 50/2 gets 96 lp/mm, so that is a 19.2x enlargement for a 13 x 10 in print.

(7) All this is just a really complicated way of saying that at any given f-stop on 4/3rds, the equivalent f-stop on 4x5 for DOF in the same composition (AOV, framing) is going to be far closer to performing at the diffraction limit. The difference in the size of enlargements at equivalent f-stops is equal to the fold-difference in the % of theoretical diffraction limit achieved.

And when you can benefit from tilting or swinging the standards on 4x5 -- or when you want narrow DOF -- then you can really bang out incredibly sharp negatives of transparencies capable of enlargement to enormous sizes.