Qcumber

I managed to get a 5070ti at a decent price close to MSRP so went for that. The second hand 4080 market is inflated too. I was also considering a 9070xt.

Might be a frame time issue as I am using QVFR so it could have been spikes of CPU peaking!? Some of the cores were pushing 80% plus with my 5800x3d. I just upgraded to a 9800x3d. All cores are now about 40% max. Now I'm seriously limited by my GPU. 5070ti hopefully arriving the middle of next week.

Just upgraded my CPU motherboard and RAM and a clean w11 install. DLSS 4 is looking better. Less ghosting against the ground. Also updated to today's release of NVIDIA driver so not sure if it's this or hardware!?

Offtopic but I'm still finding some ghosting issues with DLSS/DLAA. That's why I was considering the rx9070xt. Similar performance to a 5070ti (according to AMD) but cheaper. Plus I couldn't get hold of a 5070ti and didn't want to wait months. I am now hoping that DLSS will continue to improve.

Very much so Just managed to get a 5070ti from Amazon for £799 which was very unexpected so given that this is DLSS4 etc I think this will suit me better than rx9070 xt. Good OC potential.

I'm seeing the same. If ASW kicks in it's no longer smooth as it used to be. I had not considered that it was not working. I need to try comparing ASW and forced to half refresh.

If the overclock potential for the Rx 9070xt is as good as the 7800xt then performance could equal the rtx 5080 FE base performance. Estimated based on information from tech-power-up reviews and the official AMD performance presentation. Starting point is an assumption that the FPS of a factory OC version of the Rx 9070 xt at 4k is similar to rtx 5070 ti (85 FPS). Overclocking results for the Rx 7800 xt range was about 13-15%. If this translates to the Rx 9070xt then it could give 96-97 FPS. Baseline performance for the rtx 5080 FE is 96.7 FPS. I hope this is the case

For BVR combat DLSS 4 with preset K is very good. For WW2 or cold war dog fights the ghosting is still very noticeable and significantly affects spotting. For this reason I have reverted back to MSAA. As I use QVFR the performance with MSAA is the same as DLSS (perf), and better than DLSS (quality).

I am waiting for the independent reviews in the 5th and then will look at local pricing and availability. If I can find a 9070xt for £600 I might be tempted. So far it looks similar to the 5070ti. Another factor is overclocking. The 50xx series so far have good OC potential (10%+). I am hoping the 9070xt will be the same. The three options for me are: 5080, 5070ti and 9070xt. In the UK the cheapest 5080 with any chance of availability is £1200 and that's using a tracker bot for Amazon spot sales. Most are £1400-1500. If I wait I might be able to get one for £1050. The cheapest 5070ti is £779 but there is no stock. If the 9070xt is as well stocked as AMD are saying and they are close to MSRP then this might be a good option. Half the price of a 5080 with 85% of the performance.

Thanks for the WW2 Marianas update. I am very much looking forward to this. Any "hint" about when we might see it? I am patiently inpatient.

Just use NVIDIA profile inspector. It's a lot easier than DLSSTweaks. I also find that using DLSSTweaks causes a lot of CTDs for me with preset J or K.

Use OXRTK or set centre and peripheral res to "1.0".

I find that preset F has more noticable ghosting than K.

K is the best but there is still ghosting of aircraft close the ground or sea. Distant spotting is difficult too, due to blurring, and there is some lost of contrast of objects against clouds. You can mitigate this to some degree by increasing the resolution but this then offsets any performance benefits of using DLSS. Expectations of DLSS varies depending on user and situation. For BVR modern jet combat it works very well. For WW2 or early cold war then the close nature of dogfights means the artifacts are a lot more noticable.

Mute all notifications in the headset. There is a single tab in the quick access menu.

Looks like you are using an old version of VD. The latest is 1.33.4 (5/2/25). I would advise updating and seeing if that helps.

I get this sometimes too. It is when I am running an intensive module such as the F4. When FPS remains persistently below 72 and VRAM is maxed out. It happens more with J than K.

Has anyone got any experience with an rtx 5080 in VR yet? Most reviews suggest a decent overclock which takes the 4K performance close to a 4090. Just wondered how this translates into VR performance.

Thanks for your replies but I am really looking for a direct comparison between the quest pro and crystal light. Someone who has experience of using both headsets.

Sorry. I should have checked your signature and translated.

Increasing the resolution helps reduce ground ghosting for me. I use QVFR so this is easy to do just for the foveated region. You could try increasing your resolution and using DLSS quality to offset the reduction in performance to see if this helps.

I don't think OXRTK frametime works when QVFR is enabled. It is mentioned somewhere by mbucchia either in the QVFR thread or on the git hub wiki.

I'm not seeing that with DLSS 4. The clarity of the cockpit and MFDs look better than with 3.8.10.

https://github.com/Orbmu2k/nvidiaProfileInspector/releases It has an option that you can select the most up-to-date DLSS release. Just need to toggle this and select preset K.

The latest version of NVPI is the easiest option now. I have stopped using DLSSTweaks.