DuckStation vs HopStation

DuckStation is a free emulator for Sony PlayStation games, available for Android, Linux, macOS and Windows. Its hardware renderer supports D3D11, D3D12, OpenGL, Vulkan and Metal, with upscaling, texture filtering and 24-bit color among the graphics options. Enhancements include PGXP geometry precision, texture replacement, adaptive downsampling and post-processing shader chains. Play options include save states, rewind, runahead, CPU overclocking and multi-disc support, along with memory-card editing and save importing. DuckStation supports RetroAchievements and Discord Rich Presence. It accepts digital and analog controllers, rumble, and up to eight devices through multitap; GunCon and Justifier lightguns can be simulated with a mouse. Supported formats include bin/cue, MAME CHD, single-track ECM, MDS/MDF, CCD and unencrypted PBP. A PS1 or PS2 BIOS image is required but not supplied, and games are not provided. The maker says users should dump a BIOS from their own console. Desktop requirements include Windows 10/11 and macOS 13.3 or newer.

HopStation is a work-in-progress PlayStation emulator written in C++, shared chiefly as a reference for emulator developers rather than as a primary way to play games. It emulates the R3000 CPU, GPU, DMA and interrupt controllers, CDROM, GTE, SPU, MDEC, and memory cards. The project provides a windowed application and a headless command-line mode, with tools including a disassembler, executable sideloading, USB controller input, configurable logging, a VRAM viewer, and ImGui introspection windows. It accepts BIN/CUE disc images, but support is buggy: CUE parsing prevents many games from loading, and inaccurate CDROM emulation also limits compatibility. The SPU lacks features such as volume sweep and dynamic ADSR. A PlayStation BIOS is required; only SCPH1001, the original US launch version, has been tested. HopStation builds for Windows, Linux, and macOS and is free under the MIT License. Some builds need additional shader libraries. The author describes it as an unfinished personal learning project and says feature contributions are unlikely to be accepted.