Alan Wake 2 [PC, PS5, XBSX|S] 17th october 2023

One other odd thing I noticed in TPU's benchmark data...
4060Ti 16GB beating all the other cards has to be from the extra 4GB of VRAM.

Odd to see the 7700XT hanging in tough though, as if it's not running into VRAM pressure the way the other cards are, especially since AMD cards tend to use more VRAM for the acceleration structures and such when raytracing. I wonder if this title has the same issue a lot of recent ones do, where if you change settings 'on the fly' without restarting the game it uses extra VRAM over what it normally would?

Fun to see the 2080Ti beating the 3080 10GB at 1440p too, that extra 1GB of VRAM's coming in clutch.


View attachment 9898

If you scroll up you'll see I linked their VRAM usage chart which shows 4k max settings with path tracing with use 16GB VRAM. Honestly, I don't see why anyone would ever run native 4k at that performance anyway.
 
The digital foundry video dropped on PS5.

Quality is 4k with internal res of 1270p30
Performance is 1440p with internal res of 847p60

Both modes stick close to their targets but there are a few areas with performance issues (27+ on quality, 50+ on performance)

Outrage cycle of 480p was bogus. Who knew? :ROFLMAO:

Doesn’t line up with the performances their spec sheet advertised but I’m happy to be wrong.
 
DLSS3.5 in Alan Wake 2 can be combined with DLAA for a native implementation. It also produces stellar image quality and increased performance as well.

FSR2 is so bad in Alan Wake 2 even in its native mode, that it directly compromises the experience in a big negative way.

 
Alan-Wake-2-NVIDIA-RTX4090-benchmarks.png
 
DLSS3.5 in Alan Wake 2 can be combined with DLAA for a native implementation. It also produces stellar image quality and increased performance as well.

FSR2 is so bad in Alan Wake 2 even in its native mode, that it directly compromises the experience in a big negative way.

On consoles FSR2 is bad in all games so far. That shimmering in motion is so noticeable even in highly compressed feeds. It's really the new FXAA experience (terrible image quality industry standard) on consoles.
 
IQ is indeed terrible. Other than that the game looks quite good. Reaffirms my opinion that blowing your render budget on RT reflections is not the best use of these consoles.
 
Doesn’t line up with the performances their spec sheet advertised but I’m happy to be wrong.

It kind of does. They basically gave numbers they think they can guarantee. So if it says you can get 30 fps, you're probably running 40 most of the time. They're giving numbers that should be a worst case.
 
Right, the min FPS for the 3060 is 25.4 in the TPU benches, so the 2060 would be 20.5 if it scaled as per the average performance. Then the next rung down for DLSS is Quality, which is 720p. But you could probably get close to 30 minimum with smart optimisations instead of DLSS. So then the 847p60 mode with further cutbacks looks more reasonable, especially since it is not locked in outdoor areas.

The chart for Alan Wake 2 requirements shows the 2060 at preset low for 1080p30 with DLSS Quality. I think it actually runs better than that, but in areas in the game that are not likely to be the worst the game has to offer. Most likely they're giving you a chart that's basically what you could lock to and not have to worry about dips, or something close to that.

Maybe if they'd worded it as "30 fps minimum" or something like that, it would have alleviated some of the concerns. I don't think performance is massively higher than the numbers they gave in difficult areas, so they're not way way off. I think matching PS5 settings will give performance better than the low preset.
 
The chart for Alan Wake 2 requirements shows the 2060 at preset low for 1080p30 with DLSS Quality. I think it actually runs better than that, but in areas in the game that are not likely to be the worst the game has to offer. Most likely they're giving you a chart that's basically what you could lock to and not have to worry about dips, or something close to that.

Maybe if they'd worded it as "30 fps minimum" or something like that, it would have alleviated some of the concerns. I don't think performance is massively higher than the numbers they gave in difficult areas, so they're not way way off. I think matching PS5 settings will give performance better than the low preset.
I was looking at the ray tracing figures by accident. The 2060 should be slightly above 30 in the TPU benchmarks, but those might not be in the most demanding part of the game.
 
Last edited:
It kind of does. They basically gave numbers they think they can guarantee. So if it says you can get 30 fps, you're probably running 40 most of the time. They're giving numbers that should be a worst case.
The spec sheet said a PS5 equivalent GPU was good for 1440p/30 medium settings no RT with FSR 2 balanced which is 864p.
 
IQ is indeed terrible. Other than that the game looks quite good. Reaffirms my opinion that blowing your render budget on RT reflections is not the best use of these consoles.

IQ has been terrible in Remedy games since Alan Wake, that ran at 540p on the 360 back when Xbox 1 was still 3 years away. So, nothing changed there, it's too bad Intel never open sourced Xess though.
 
On consoles FSR2 is bad in all games so far. That shimmering in motion is so noticeable even in highly compressed feeds. It's really the new FXAA experience (terrible image quality industry standard) on consoles.
It looks like a repeat of what happened in Forspoken, FSR Native is broken, and you get terrible sparkling, noise and shimmering. Since the game supposedly doesn't have TAA, you're basically forced to use upscaling, and have to deal with the bad stability and usual FSR upscaling cons.
 
Back
Top