@Rich_Leadbetter If a similar situation happens in the future (i.e. NVIDIA previews DLSS 5.5) how would you balance the need to release timely content vs the difficulties of assembling the team on short notice for reactions?
I thought the Q&A addressed many of my concerns from the initial video. I don't know if waiting for audience & developer reactions is the right takeaway here, because I value DF's editorial takes and I wouldn't want, as a hypothetical example, Oliver's enthusiasm dampened to avoid risk of internet blowback. But I think balancing that enthusiasm with reactions from John & Alex would result in a better discussion & a better video. More than anything, I hope the team is doing okay and that everyone is still gelling after this episode.
DF covered this angle in their Q&A, but just re-stating it here:
Game studio approving their IP for use in a tech preview and providing a marketing quote is NOT proof that everyone in the studio feels that way. To me, those quotes speak to studio leadership's (understandable) excitement over the tech as a path to lower development costs.
I'd be surprised if leads & principals on art teams at those studios shared that excitement, but they weren't quoted here.
This is the first DLSS feature I've ever had a negative reaction to. On a purely technical basis, I think it's cool and interesting that we've reached a point where something like this is possible in real-time rendering.
But that doesn't mean it's good. I'm sure a deeper dive on technical specifics of DLSS5 is coming at some point, but it doesn't FEEL accurate to describe what happened with the RE9 Grace examples as a simple change in scene lighting. It might well BE accurate, to the extent that scene lighting also includes things like subsurface scattering, radiance caches, etc. Whatever the method, she was no longer recognizable as the same character and that's a bad result.
Overall this feels like an overreach by NVIDIA and a step in the wrong direction. DLSS SR, FG, RR are great models and I wish NVIDIA's next ML model had been something along those lines. With DLSS5, it feels like their working definition of "graphical fidelity" changed from "fidelity to the game's core render" to "fidelity to whatever version of reality NVIDIA trained its world models on". If that's true, it's a major shift and one that's not good for anyone.
Alex alluded to this in DF Direct, but I would absolutely welcome a video (or series of videos) on transitioning from Win11 to Linux, with a special focus on users with NVIDIA GPUs. I'm vaguely aware of the 'Nvidia tax' on Linux and would be willing to accept the performance penalty on DX12 games.. but what's turned me off about making the switch is the unclear support for other features I care about, like HDR, VRR, and DLSS.
Comments 5
Re: DLSS 5: Game-Changing Tech That Poses Big Questions For The Future Of Gaming
@Rich_Leadbetter If a similar situation happens in the future (i.e. NVIDIA previews DLSS 5.5) how would you balance the need to release timely content vs the difficulties of assembling the team on short notice for reactions?
I thought the Q&A addressed many of my concerns from the initial video. I don't know if waiting for audience & developer reactions is the right takeaway here, because I value DF's editorial takes and I wouldn't want, as a hypothetical example, Oliver's enthusiasm dampened to avoid risk of internet blowback. But I think balancing that enthusiasm with reactions from John & Alex would result in a better discussion & a better video. More than anything, I hope the team is doing okay and that everyone is still gelling after this episode.
Re: Nvidia's new DLSS 5 Brings Photo-Realistic Lighting To RTX 50-Series
DF covered this angle in their Q&A, but just re-stating it here:
Game studio approving their IP for use in a tech preview and providing a marketing quote is NOT proof that everyone in the studio feels that way. To me, those quotes speak to studio leadership's (understandable) excitement over the tech as a path to lower development costs.
I'd be surprised if leads & principals on art teams at those studios shared that excitement, but they weren't quoted here.
Re: Nvidia's new DLSS 5 Brings Photo-Realistic Lighting To RTX 50-Series
This is the first DLSS feature I've ever had a negative reaction to. On a purely technical basis, I think it's cool and interesting that we've reached a point where something like this is possible in real-time rendering.
But that doesn't mean it's good. I'm sure a deeper dive on technical specifics of DLSS5 is coming at some point, but it doesn't FEEL accurate to describe what happened with the RE9 Grace examples as a simple change in scene lighting. It might well BE accurate, to the extent that scene lighting also includes things like subsurface scattering, radiance caches, etc. Whatever the method, she was no longer recognizable as the same character and that's a bad result.
Overall this feels like an overreach by NVIDIA and a step in the wrong direction. DLSS SR, FG, RR are great models and I wish NVIDIA's next ML model had been something along those lines.
With DLSS5, it feels like their working definition of "graphical fidelity" changed from "fidelity to the game's core render" to "fidelity to whatever version of reality NVIDIA trained its world models on". If that's true, it's a major shift and one that's not good for anyone.
Re: Discounts Available on the DF Supporter Program
Cheers Rich, I just made the switch to annual - I hope enough users take this up to make forecasting a bit easier!
Re: Proof That Recent Windows Update Harms Gaming Performance
Alex alluded to this in DF Direct, but I would absolutely welcome a video (or series of videos) on transitioning from Win11 to Linux, with a special focus on users with NVIDIA GPUs. I'm vaguely aware of the 'Nvidia tax' on Linux and would be willing to accept the performance penalty on DX12 games.. but what's turned me off about making the switch is the unclear support for other features I care about, like HDR, VRR, and DLSS.