Someone asked me what's the difference between filtered and unfiltered espresso on the refractometer and I had to admit I hadn't thought "out of the cup" yet and never tried.
Now I did: I took 18g of a dark (Tonino #63) blend in the Reneka basket, ground it a bit too fine I guess and had a very slow extraction which I helped a bit by pushing up the lever on my Londinium. 29g espresso out.
Calibrated the VST refractometer at 19.9°C ambient and temp climbed to 20°C in the process so no need to recalibrate before I finished.
I poured the espresso into a large heavy cup and swirled to get it to cool down relatively quickly, then took measurements, after that put 7g in the centrifuge and repeated the measurements, adding the centrifuged one.
Before using the filters, I sucked up and pushed out the espresso with the syringe to get an optimal mix and I pressed out the drops straight onto the VST prism. Pressed GO three times on all samples.
Shortly after half-cooling down:
VST filter: 9.5%
NO filter: 9.6%, 9.7%, 9.8% then staying 9.8%
SA filter: 9.6%
Then I did the centrifuge and after 10 minutes when the centrifuge was ready, using fresh filters on the more completely cooled sample, stirred well again:
VST filter: 9.8%
NO filter: 9.9%
SA filter: 9.8%
Centrifuged sample: 9.8%
I would need to do this more often, and preferably hear from others who are able to repeat this, but to me it indicates that a filter is useful, the VST filter assures you you are measuing the same way as the next guy, the cheapo filter I found is okay, but even NO filter works with just a small error if you stir very well and use a thoroughly cooled sample (and clean the refractometer well).