tl;dw
- Rich wants to wait for final code instead of using this examination as a final impression.
- Not quite "Next-gen today" but still doing a lot with its visual tech.
- Unexpected loading times, such as entering and exiting a corridor area.
- Visual parity save for some LODs.
- 1800p checkerboarding seems to be on all 3 platforms but requires more investigation by DF, including PC.
- Framerate drops across both platforms. Technically unlocked but hovers slightly above and more often below 30fps. Often into mid 20's on X.
- Pro runs slightly better than X in like comparisons, including during missions and combat.
- Frame timing issues across both platforms.
- 1080p output on pro alleviates some problems (40-50fps) but still no 30fps cap. "Hints" at 60fps ceiling in very minimalistic areas but isn't really relevant to overall gameplay.
- Anthem offers an upgraded experience for PC users, to come in a later video.
[UPDATE] PC Beta Performance and Analysis
New video focusing on PC performance was posted a short while ago by Dictator giving some details on the Anthem Beta as well. The video is linked above, with the tl;dw listed below:
- Targetting 1080p60 across multiple configurations to match the performance profile of the PS4 Pro at 1080p mode
- Mid-range testing consisted of i5 8400, GTX 1060/RX 580, 16GB DDR4 2667MHZ, SSD
- 60fps is off the cards at 1080p Ultra. Hub sees dips to high 40's/50's, drops to mid 40's in combat
- Below 60fps in the overworld
- Medium settings preset below 1080p60
- Low settings preset acheives some headroom above 60fps except in hub and firefights
- Didn't expect attaining 60fps to be this challenging
- Typically expect these cards to perform really well in Frostbyte games
- GTX 1060 performance profile even lower than expected based on other Frostbyte games (footage not shown as it was lost)
- CPU didn't seem to be a bottleneck in this scenario for the most part
- Higher mid-range graphics cards were Vega 56 / GTX 1070
- Vega 56 held above 60fps more consistently from the onset, but 1070 suffers heavily with drops below 60fps in firefights at Ultra settings, despite performing similarly to Vega while flying
- Vega dips below 60 during some scenes, including a giant enemy crab boss. GTX 1070 suffers similarly, probably tying in to massage damage
- Playing in the internal cave area yields much higher framerates due to lower scene complexity. Vega averages ~80fps allowing a lot of overhead for alpha effects in firefights, 1070 around ~60fps with some drops below
- Probably not worth dropping graphics settings with these cards due to low performance gains for image quality trade-off
- High-end testing uses RTX 2080 TI at Ultra Settings 4K
- Alex didn't expect locked 4K60
- Overworld sees drops into the 50's
- Firefights mainly in the 40's
- Dark1x played at 3840x1600 ultrawide resolution at 60fps in the overworld and hub but dropped frames heavily under 60 during firefights
- Richard says his platform of preference is PC, given all the circumstances
- AMD had an Anthem driver ready for this demo, Nvidia did not
- Alex states that there is rendering tech and a world size being used in this game that isn't commonly used in other console Frostbyte games
- Graphics options seemingly lacking or not having as much of an effect on fps as they should
- Ultra settings down to High doesn't seem to have much of an effect, only when dropping to Medium is there a tangible difference (HBAO>SSAO, shadow resolution etc.) with an 18% performance gain
- Low settings looks rough compared to Ultra, turns off a lot of settings, such as AA, SSAO, DOF and heavily affects shadows for a 21% performance gain
- Performance scaling is bad
- No DX12 options
- No HDR controls in game, disabling in Windows did not work - Had to disable HDR on the display
- Some graphical bugs on AMD hardware
- Need a deeper dive into specific Graphics options and what they do
- Firefights are the biggest offender for performance loss, probably due to heavy use of alpha effects
- Comparing gameplay type and performance to Warframe and Destiny 2 states that with such an emphasis on combat, more should have been done to optimise this area of gameplay
- CPU utilisation hit 100% and hit performance heavily, seemingly stopping by itself in Richard's case, but Alex needed to close the game for it to stop
- Loading screens seem to be random and occur during missions and traversal
- Alex needed to load an area twice whereas Richard only needed to load the area once.
Last edited:
