I compared phi reconstruction efficiencies from the phi embedding and the single electron embedding.

This is calculated from (Update: please ignore those two plots below. This is different isue. I'm sorry about this confusion.)

I used the same cut for both calculation.

This difference end up with

Those efficiencies are without PID cut efficiency.

The relative difference are 15% at low momentum.

Update: same single reconstraction efficiency with |eta|<0.1

There is something wrong with the single electron embedding because it should be pt flat.

I used embedding minimc from

<input URL="file:/eliza15/star/starprod/hpss/staging/embedding/AuAu200_production/Electrons_AuAu200GeV/Electrons_*_20103110/P10ik.SL10k/2010/*/st_physics_*.minimc.root"/>

http://www.star.bnl.gov/HyperNews-star/get/embedding/1402/1/1/1/2/1/1/1/1/1.html

My cut for MATCHED pair is

            float ratio = (float)retrack->fitPts() / (1.0*retrack->nPossiblePts());
            if(retrack->geantId()==3 && retrack->commonHits() > 10 && retrack->fitPts() >= 15
                    && fabs(retrack->dcaGl()) < 1.5 && fabs(retrack->ptPr()) > 1e-6 //&& fabs(retrack->phiPr()) > 1e-6
                    && ratio >= 0.52
                )

For Mc tracks

if(mctrack->geantId()==3 && fabs(mctrack->etaMc()) < 0.1 && mctrack->ptMc() > 0.0)

The basic QA is in this post

http://www.star.bnl.gov/HyperNews-star/get/embedding/1402/1.html