Question for sanking: I am curious as to the magnitude of the effect. As I am sure you will agree, there are a lot of things that can be measured that don't affect print quality.
What method did you uses to test for that? I was thinking of using a series of evenly exposed frames of pyro-developed negatives as filters to determine the ISO(R) for each on Ilford MG paper. I think that would be pretty simple to do. Did you do something like that? Or did you measure spectral transmission of the stained negatives?
The problem with using individual frames to make VC filters and then measuring the contrast range of a print made using those frames/filters is that the stain a negative is not constant, it varies with the density of the negative. The shadow areas have less stain than the highlights, so one filter will not be able to make a proper comparision.
I did some tests a few years ago to figure this out, and my approach was to expose and process some stepwedges. I measured the density of the pyro processed negs along with Ilford MG filters using a color densitometer and scanned them for absorbance vs. wavelength using a spectrophotometer, which I bought just for this test.
Using these techniques, I was able to determine the amount of "filtration" in the pyro-stained negative. I then assumed that the layers in my VC paper (Ilford MG IV FB) were pretty blue only or green only sensitive. That's a big assumption, but it simplifies the rest. (I planned to make a wedge spectrograph to actually measure the response of the paper with various multigrade filters, but never made it... I have the parts, so I'll do it someday for testing home-made emulsions.)
Using those measurements, I figured that the stain in the lower Zones of the neg was acting as about 10CC Yellow filtration, and in the higher Zones, there was about 40CC Yellow or so. (I'll have to look the numbers up if you want more precise values.) If you subtract out the base+fog stain from the neg, that ends up being about 30CC or so Yellow filtration that the stain is adding to the negative. For my MGIV FB paper, that ends up being about 1/2 to 1 grade when printing without any filtration.
So that means the highlights have a compression of about 1/2 to 1 grade of paper compared to the shadows. Or, if you add 30CC Magenta, you can filter out the yellow filtration of the highlights and then have about 30CC Magenta in the shadows, thereby increasing the shadow separation relative to the highlights. Keep in mind, the middle Zones, and split between these two extremes so there is not a big difference in fitration from the middle Zones to the highlights or the middle Zones to the shadows.
The measurements also indicate that developers with a stain color similar to PMK should give greater highlight compensation than ones with a stain color similar to Pyrocat. There is less difference in green absorbtion relative to blue absorbtion in Pyrocat stain than PMK.(Also, there is more UV absorbtion relative to blue with Pyrocat than PMK.)