With many attractive attributes including improved tissue characterization, reduced noise, and lower dose, photon counting detectors (PCDs) with multiple energy bins are being considered for clinical CT systems. A major problem with PCDs is the slow count rate, resulting in count rate loss and pulse pileup. Another challenge to PCDs with multiple energy bins is the large amount of projection data acquired that must be transferred in real time through slip rings to data storage subsystems, causing a bandwidth bottleneck problem. Our goal is to address these challenges with a dynamic bowtie filter, a pulse pileup modeling, and a compression algorithm for PCD projection data. A dynamic bowtie modulates flux distribution as a function of fan and view angles to reduce dose, scatter, and detector dynamic range. One approach, the piecewise-linear attenuator (Hsieh and Pelc, Med Phys 2013), consists of multiple wedges, each of which covers a different fan angle range and is moved in the axial direction to change the thickness seen in an axial slice. Our implementation of the filter with precision components and a control algorithm reduces the detected flux dynamic range to 42 for a chest and 25 for an abdomen, corresponding to a reduction factor of 5 and 11 from the object scans without the bowtie. The thickness profile of the piecewise-linear attenuator is proportional to wedge position and is sensitive to the precision of wedge control. Thus, we propose another dynamic bowtie design called a fluid-filled dynamic bowtie filter (FDBF), a two-dimensional array of small binary elements (filled or empty), that may be more reproducible due to digital control. On average, for simulated scans of chest, abdomen, and shoulder sections of an anthropomorphic phantom, the proposed FDBF reduces the maximum-count-rate required of a PCD to 1.2 Mcps/mm2, which is 55.8 and 3.3 times lower than the max-count-rate of the conventional bowtie and the piecewise-linear bowtie, respectively. For the pulse pileup effect, the performance of several analytical models are dependent on the assumptions used, including the assumed pulse shape which could differ from the actual one. From our sensitivity analysis of a pileup model by Taguchi et al. (Med Phys, 2010), the accuracy of some parameters such as the deadtime are more important than others, and they matter more with increasing count rate. For the bandwidth problem, we explored the use of data compression that could be performed prior to transmission through the slip ring. Utilizing the dependencies in the multi-energy data, our proposed compressor achieves an average compression ratio of 2.25:1 for lossless compression and 4.27:1 for lossy compression.