In a neutrino-less double-beta-decay (0νββ) experiment, an irremovable two-neutrino double-beta-decay (2νββ) background surrounds the Q-value of the double beta decay isotope. The energy resolution must be improved to differentiate between 0νββ and 2νββ events. CAlcium fluoride for studies of Neutrino and Dark matters by Low Energy Spectrometer (CANDLES) discerns the 0νββ of 48 Ca using a CaF2 scintillator as the detector and source. Photomultiplier tubes (PMTs) collect scintillation photons. Ideally, the energy resolution should equal the statistical fluctuation of the number of photoelectrons. At the Q-value of 48 Ca, the current energy resolution (2.6%) exceeds this fluctuation (1.6%). Because of CaF2's long decay constant of 1000 ns, a signal integration in 4000 ns is used to calculate the energy. The baseline fluctuation (σ baseline ) is accumulated in the signal integration, degrading the energy resolution. Therefore, this paper studies σ baseline in the CANDLES detector, which has a severe effect (1%) at the Q-value of 48 Ca. To avoid σ baseline , photon counting can be used to obtain the number of photoelectrons in each PMT; however, a significant photoelectron signal overlapping probability in each PMT causes missing photoelectrons in counting and reduces the energy resolution. "Partial photon counting" reduces σ baseline and minimizes photoelectron loss. We thus obtain improved energy resolutions of 4.5-4.0% at 1460.8 keV (γ-ray of 40 K), and 3.3-2.9% at 2614.5 keV (γ-ray of 208 Tl). The energy resolution at the Q-value shows an estimated improvement of 2.2%, with improved detector sensitivity by factor 1.09 for the 0νββ half-life of 48 Ca.