EVO Player (PPSA99039)

[Aryan-Solanki]

Existing thread

• I searched this category for the title ID and no thread exists for this title on any platform.

Title

EVO Player (PPSA99039)

Content version tested

01.012.000

AnyPS5 commit

main a08de68, built from source. No pull requests merged on top. Follow-up runs below were
on main 03c4cb5.

System tested

Windows 11 Home Single Language build 26200, NVIDIA GeForce GTX 1660 Ti (driver 32.0.15.5597), MinGW-w64 GCC 15.2.0 (WinLibs x86_64-ucrt-posix-seh, 15.2.0posix-14.0.0-ucrt-r7), CMake 3.31.8, Ninja 1.13.2, Release.

Furthest point

Menus

What is shown

Reaches its own menu UI: the Welcome screen with BROWSE, RECENT, FAVORITES, PROVIDERS,
SETTINGS and SCREENSHOT, a clock, a highlighted entry and the Select/Back prompts, running
steadily at about 10 FPS at 1088x612 and still up after 4000 frames. Its boot log passes
"EVO boot: frame 4: present done".

Working: Vulkan device selection and swapchain creation, the shader disk cache (171 hits,
0 misses, 79.9 MiB read in 68 ms), the filesystem mounts, and font loading from app0/assets.

Not exercised: controller input was not sent, so the menu is reported as rendered rather than
confirmed responsive. Audio, media playback, the network and Stremio features were not tested.

Current blocker

An intermittent fault in the NVIDIA driver on a thread whose stack is in libSceAgcDriver,
tracked in #6335: 6 crashes in 15 runs on main 03c4cb5 on 15 October, every one at the same
faulting address (nvoglv64.dll+0xa13a46, 0xc0000005, read of 0x10). It follows the line

[gpu] storage image 0x213800000 destroyed with GPU results pending

although that line also appears on runs that do not crash.

Not currently reproducible: 0 faults in 35 runs on the same commit the next day, across plain
main, main with #6414 merged, and a wiped shader cache. Not reproduced on RTX 4060 or RTX 3070
by other reporters. No open pull requests or local changes are known to address it.

How it was run

Both eboot.bin and sce_module/libc.prx ship as SELF containers and need unwrapping first:

python tools/unwrap_self.py PPSA99039/eboot.bin src/eboot.elf
python tools/unwrap_self.py PPSA99039/sce_module/libc.prx src/sce_module/libc.prx
relinker --windows src/eboot.elf run/app.exe

The relink reports 342 external prx references and 1 converted module, exit 0.

Beyond the usage guide: the title needs its original eboot.bin in app0/. Without it, it falls back to /data/homebrew/<title id> and reopens /data in a loop.

No environment variables are required. ANYPS5_GPU="GTX 1660" was used for some runs; it does not change the outcome.

Release used: https://github.com/sainsaji/EVO-PLAYER-PS5/releases/tag/v0.12.0, EVOPlayer-v0.12.0-PPSA99039.zip, sha256 98788be10b3d2fe2909f6e353e2e09409ad29e46c2bc37112e2f35d9baa00d8a

[rooklz]

Update with media playback, which the first report didn't exercise.

Setup: main a6dcea9279, EVO Player 0.12.5 from the catalog (sha256 79b18f57… matches). Linux x86-64 container (Ubuntu 24.04, GCC 13.3, Release) under Rosetta on an Apple M5, Xvfb with llvmpipe (Selected GPU name=llvmpipe (LLVM 20.1.2, 256 bits)). Converted with relinker --to-rosetta --registry; without --to-rosetta the title dies with Illegal instruction when the file browser opens, which is Rosetta, not AnyPS5. Rosetta's loader also refuses the output directly (first load segment not at file offset 0), so I started it through ld-linux-x86-64.so.2. Software rendering, so about 10 FPS throughout. import_audit.py: 0 absent imports.

Media: four 10 s clips I generated with ffmpeg (testsrc2 plus a 440 Hz tone, no third-party content), placed in mnt/usb0/ beside the executable. Input was scripted with tools/pad_input.py.

Clip	Result
H.264 High 1280x720, AAC stereo (MP4)	plays to the end; resident AVC decoder 1280x768, AAC through EVO adec native
HEVC Main10 1920x1080, HDR10 (PQ, BT.2020)	plays to the end; 10-bit first frame: 1920x1080 pitch=1920 pitch_bytes=3840 (expected 1920/3840), heartbeat decodes=300 framesout=298; the title switches scanout to HDR10 Bgr10A2Bt2100Pq (SubmitChangeBufferAttribute2 rc=0) and back to SDR when playback stops
VP9 profile 2, 10-bit 1280x720, Opus (WebM)	plays; EVO itself routes it to FFmpeg (VP9-2 prof=2 10-bit 1280x720 unsupported -> FFmpeg)
H.264 1280x720, AC-3 5.1 (MKV)	plays; audio: source 6ch -> port 8ch

Browse USB lists the four files with thumbnails decoded through Videodec2, and Back, the "Stop playback?" dialog and the subtitle panel respond to the scripted input. So the 10-bit HEVC output path that #6709 added decodes in a real title here. One note: after the switch to HDR10, sceVideoOutGetOutputStatus still reports dynamic_range=1 (SDR), which is probably right for an Xvfb display without HDR.

Not tested: a real GPU, Windows, audio output (SDL_AUDIODRIVER=dummy), network providers, and the console's own behaviour for any of it.