1 00:00:00,040 --> 00:00:02,520 We've got a new saying. Frame Gen. 2 00:00:02,520 --> 00:00:06,640 I didn't think you were going to start with it. 3 00:00:07,600 --> 00:00:12,240 Bam, slap him in the face. All right, so we've hated on ray 4 00:00:12,240 --> 00:00:13,960 tracing. We have, yeah. 5 00:00:13,960 --> 00:00:16,720 In the past, yeah, we hate on what Acer are doing with the 6 00:00:16,720 --> 00:00:19,120 Nitro Blaze, and now we're saying it. 7 00:00:19,160 --> 00:00:21,680 Yeah, yeah. We hate frame Gen. 8 00:00:21,680 --> 00:00:24,280 Yeah, well, I I think. OK, let's start off. 9 00:00:24,640 --> 00:00:27,920 Hating is a big word, all right? Because it is useful. 10 00:00:28,040 --> 00:00:32,040 No, it is good. And it is getting better than it 11 00:00:32,040 --> 00:00:33,760 was. Yes, yeah, yeah. 12 00:00:33,920 --> 00:00:37,520 But is at what cost? Yes, yeah, yeah. 13 00:00:37,520 --> 00:00:39,240 Is the problem with frame Gen. Yes. 14 00:00:39,240 --> 00:00:40,680 Exactly. I, I think there's, like you 15 00:00:40,680 --> 00:00:43,200 said, there's a definitely like use case for it. 16 00:00:43,200 --> 00:00:44,520 Yeah. And I think it's a good 17 00:00:44,520 --> 00:00:47,080 technology to expand upon. Yeah. 18 00:00:47,560 --> 00:00:50,040 But currently, there's definitely some quirks. 19 00:00:51,520 --> 00:00:52,920 Where do we start with the? I know it's. 20 00:00:53,280 --> 00:00:55,640 So difficult isn't? It there's so many. 21 00:00:55,640 --> 00:00:58,520 Issues so frame generation there's there's different kinds 22 00:00:58,720 --> 00:01:00,320 of course. So obviously we've got the new 23 00:01:00,320 --> 00:01:03,680 NVIDIA 50 series which has introduced the new multi frame 24 00:01:03,680 --> 00:01:04,519 Gen. multi. Frame. 25 00:01:04,720 --> 00:01:08,440 Gen. that's different, I would say because it's a whole new 26 00:01:08,440 --> 00:01:10,520 technology. They've revamped it from the 27 00:01:10,520 --> 00:01:12,480 ground up and it's specific for these cards. 28 00:01:13,160 --> 00:01:16,480 The 40 series obviously have the two times which has been 29 00:01:16,520 --> 00:01:21,400 upgraded as well. So that's cool, but that's still 30 00:01:21,400 --> 00:01:25,680 has issues just I would say less issues than other stuff like 31 00:01:25,680 --> 00:01:30,240 we've seen with FSR 3 frame generation also stuff like AFMF. 32 00:01:30,560 --> 00:01:35,920 So AM DS fluid motion frames. And then we've also got lost 33 00:01:35,920 --> 00:01:38,040 with scaling, which is a paid app on Steam. 34 00:01:38,360 --> 00:01:43,080 And effectively what they do is try to generate frames. 35 00:01:43,080 --> 00:01:47,520 So it will take 2 frames and then artificially fake one in 36 00:01:47,520 --> 00:01:50,960 the middle or or multiple, you know, because even scaling can 37 00:01:50,960 --> 00:01:53,640 do up to four times frame generation. 38 00:01:53,920 --> 00:01:57,520 So you can take 2 frames and then put, you know, four times 39 00:01:57,520 --> 00:02:00,600 as amount in in between. But obviously they're they're 40 00:02:00,600 --> 00:02:03,080 fake. You're downloading more frames. 41 00:02:03,520 --> 00:02:06,560 So how can I download more FPS? Well, it's it's funny because 42 00:02:06,560 --> 00:02:08,840 that used to be a big thing and now you literally can. 43 00:02:08,919 --> 00:02:16,000 So it's a way of artificially increasing what the what the FPS 44 00:02:16,000 --> 00:02:17,880 is showing. And I want to be, I was trying 45 00:02:17,880 --> 00:02:21,200 to be very careful of how I say that because it's not the the 46 00:02:21,200 --> 00:02:24,160 frames that you're feeling or playing, it's just the frames 47 00:02:24,160 --> 00:02:26,480 that it's showing. Yeah, it's what your eyes are 48 00:02:26,480 --> 00:02:28,320 seeing. So it's artificially creating 49 00:02:28,320 --> 00:02:34,440 frames inside through some kind of AI magic by replicating. 50 00:02:34,440 --> 00:02:37,920 It's a bit like before in video editing we had a thing called 51 00:02:37,960 --> 00:02:41,200 optical flow where you could slow footage down that wasn't a 52 00:02:41,200 --> 00:02:42,480 high frame rate. Yes. 53 00:02:42,480 --> 00:02:44,560 And it would artificially create frames inside. 54 00:02:44,560 --> 00:02:46,320 Yes, it's kind of. Exactly the same. 55 00:02:46,320 --> 00:02:50,520 Make it look less stuttery yes. So when you slowed down like 5 56 00:02:50,520 --> 00:02:54,640 FPS, instead of it being like a slide show, it would still look 57 00:02:54,640 --> 00:02:57,240 like a video even though there's not enough information. 58 00:02:57,240 --> 00:03:00,240 But obviously with artificially creating frames, you're relying 59 00:03:00,240 --> 00:03:04,040 on either the driver base or the hardware base to guess what the 60 00:03:04,040 --> 00:03:06,440 frames would be inside. That, yeah, yeah. 61 00:03:06,440 --> 00:03:09,440 So there's some sort of algorithm that like thinks OK. 62 00:03:09,560 --> 00:03:12,600 Between A&B that it's moving in that direction. 63 00:03:12,600 --> 00:03:14,640 That should happen. Yes, exactly that. 64 00:03:14,640 --> 00:03:17,120 So there's those two as well. Just round up there's as you've 65 00:03:17,120 --> 00:03:19,760 just touched on, there's hardware based and then there's 66 00:03:20,040 --> 00:03:22,680 like driver based. So the driver based ones are 67 00:03:22,680 --> 00:03:27,120 stuff like AFMF. So AM DS side lossless scaling, 68 00:03:27,120 --> 00:03:30,200 I would cast as that as well because it's just like a 69 00:03:30,240 --> 00:03:32,440 software based, software based exactly. 70 00:03:32,720 --> 00:03:38,520 And then you've also got still software based, but it's AMD FSR 71 00:03:38,520 --> 00:03:42,360 3 frame generation and then you've got DLSS frame generation 72 00:03:42,360 --> 00:03:45,720 which is hardware based, which is that sort of tear up the 73 00:03:45,720 --> 00:03:48,240 better one, so. Before we go into the actually I 74 00:03:48,240 --> 00:03:50,680 was I was going to start with the problems, but actually I'll 75 00:03:50,760 --> 00:03:52,480 start with a good thing. Do this as a shit sandwich 76 00:03:55,440 --> 00:03:58,720 there. There is good reason to use it. 77 00:03:58,960 --> 00:04:01,560 Yes. And like you're saying, adding 78 00:04:01,560 --> 00:04:05,080 longevity to ageing hardware with newer game releases that 79 00:04:05,080 --> 00:04:08,280 are more demanding. It's definitely a great use case 80 00:04:08,280 --> 00:04:12,200 on paper. And it also seems apparent that 81 00:04:12,480 --> 00:04:15,760 games over the last year and definitely going forward into 82 00:04:15,760 --> 00:04:19,519 25, all games are going to be using some form of frame 83 00:04:19,519 --> 00:04:22,960 generation and upscaling or a combination of both. 84 00:04:23,920 --> 00:04:27,200 So we're going to see more and more of this being relied upon 85 00:04:27,200 --> 00:04:30,440 for games, especially on in the PC World where we have such 86 00:04:30,440 --> 00:04:31,880 varying hardware. Yeah. 87 00:04:31,880 --> 00:04:35,640 And especially with that being Nvidia's like ultimate push now. 88 00:04:35,720 --> 00:04:39,480 Well, it's the big, really the big performance upgrade with the 89 00:04:39,480 --> 00:04:42,240 50 series cards is, is these features. 90 00:04:42,240 --> 00:04:44,640 Yeah, yeah, the new frame. Rather than the pure kind of 91 00:04:44,640 --> 00:04:45,600 hardware. Kind of, yeah. 92 00:04:45,640 --> 00:04:47,440 Yeah. You're not RAW powering through 93 00:04:47,440 --> 00:04:50,120 it anymore. You're relying on upscaling a 94 00:04:50,120 --> 00:04:52,080 lower resolution and then creating fake frames. 95 00:04:52,200 --> 00:04:54,840 Creating freight frames. So I mean, if you've got 96 00:04:54,840 --> 00:04:57,880 underperforming hardware, maybe you've got an older PC handheld, 97 00:04:57,880 --> 00:05:01,760 maybe a Steam Deck was starting to see some janky, yes, 98 00:05:02,440 --> 00:05:05,200 workarounds that. Or even just full on PCs. 99 00:05:05,200 --> 00:05:06,600 You know this works on those as well. 100 00:05:06,840 --> 00:05:10,120 Like again, if you you've got AMD, you've probably got AFMF or 101 00:05:10,120 --> 00:05:13,040 lossless scaling doesn't care what system you've got, it will 102 00:05:13,040 --> 00:05:15,840 just work. So you can if your graphics card 103 00:05:15,840 --> 00:05:18,200 struggling or if your handheld struggling or whatever, or if 104 00:05:18,200 --> 00:05:20,520 you just want to play at lower power settings and still have a 105 00:05:20,520 --> 00:05:24,280 higher FPS, you can utilize these features to potentially 106 00:05:24,280 --> 00:05:28,800 increase your FPS for a visual. Smoothness, very well put, 107 00:05:28,800 --> 00:05:30,680 visual speeders, and that's where I want to kind of really 108 00:05:30,680 --> 00:05:35,680 put a pin right there because it's current technology anyway. 109 00:05:35,680 --> 00:05:39,640 Who knows what we're going to see very soon, but right now 110 00:05:40,080 --> 00:05:43,920 what currently happens is that visually it looks like you're 111 00:05:43,920 --> 00:05:46,640 getting 90FPS. You get for example. 112 00:05:46,800 --> 00:05:50,560 Yeah, as an example. But in truth, what it feels 113 00:05:50,560 --> 00:05:53,840 like, and this for me is really important, it still feels like 114 00:05:53,840 --> 00:05:55,720 the base frame rate. Yes, yeah. 115 00:05:55,720 --> 00:05:58,240 So you could, you could be outputting AT30FPS and you're 116 00:05:58,240 --> 00:06:00,400 using three times. Yeah. 117 00:06:00,400 --> 00:06:03,320 And it's showing you, you're getting 90FPS on this, right. 118 00:06:03,360 --> 00:06:05,840 And, and visually you look at it, you go, yeah, that's kind of 119 00:06:05,840 --> 00:06:08,160 90 FPS. But to you to you're controlling 120 00:06:08,160 --> 00:06:10,920 to how it feels and the feedback you get, you're not getting that 121 00:06:10,920 --> 00:06:14,360 fluid high speed feeling that you get with a higher frame 122 00:06:14,360 --> 00:06:15,200 rate. Yeah, yeah, 'cause. 123 00:06:15,200 --> 00:06:17,800 Still feels like 30. Yes, yeah, 'cause typically the, 124 00:06:18,000 --> 00:06:21,480 the higher the frame rate, the lower the latency, yes, Because 125 00:06:21,480 --> 00:06:23,160 obviously the frames are coming out like a machine gun. 126 00:06:23,160 --> 00:06:27,440 So if you've got a high polling rate input, it can sort of pick 127 00:06:27,440 --> 00:06:30,720 that up and the faster frames. This is why like a competitive 128 00:06:30,960 --> 00:06:34,560 Counter Strike players play at like 400 FPS because they want 129 00:06:34,840 --> 00:06:39,000 the absolute maximum amount of control so that as soon as you 130 00:06:39,000 --> 00:06:43,040 touch that stick or whatever it is mouse, then it's going Yep, 131 00:06:43,040 --> 00:06:46,000 that's inputted. But obviously with fake frames 132 00:06:46,000 --> 00:06:49,120 with frame Gen. that's not happening because instead of 133 00:06:49,120 --> 00:06:53,920 your GPU raw outputting these frames, it's taking that 30 FPS 134 00:06:54,080 --> 00:06:58,880 that's taking it to 60 or 90. But really, the interval is 135 00:06:58,880 --> 00:07:02,640 still the same and therefore it introduces latency. 136 00:07:02,680 --> 00:07:08,120 Now, just before the latency part with like great single 137 00:07:08,120 --> 00:07:12,320 player games or non competitive multiplayer games, there's a 138 00:07:12,320 --> 00:07:15,960 kind of once you get to like 60 FPS, everything kind of feels 139 00:07:15,960 --> 00:07:17,880 great. Yeah, Feels great, yes. 140 00:07:17,880 --> 00:07:20,680 You get higher than that and it gets like 90 like, yeah, OK. 141 00:07:21,400 --> 00:07:24,800 But once you're starting to get above that, yeah, it's kind of 142 00:07:24,800 --> 00:07:28,800 very little gain to it. The big numbers, the big FPS 143 00:07:28,800 --> 00:07:33,480 counts that that we see in like FPS counters are typically for 144 00:07:33,480 --> 00:07:36,240 FPS games like you were saying. Like competitive stuff 145 00:07:36,640 --> 00:07:38,000 Competitive. Shooters and stuff, you want 146 00:07:38,000 --> 00:07:40,880 that super high? Yeah, super quick reaction. 147 00:07:42,000 --> 00:07:47,920 Well, we're seeing NVIDIA kind of like showcasing the 50 series 148 00:07:47,920 --> 00:07:51,880 of his multi frame Gen. at like 3-4 hundred or above or above, 149 00:07:52,200 --> 00:07:55,640 which is I would argue meaningless for single player 150 00:07:55,640 --> 00:07:58,360 games, especially if you can't feel it. 151 00:07:59,320 --> 00:08:01,600 And really really most attractive those super high 152 00:08:01,600 --> 00:08:06,480 frame rates for competitive gamers, but it'd be useless to 153 00:08:06,480 --> 00:08:09,760 them because a the frames of are fakely generated. 154 00:08:09,760 --> 00:08:12,280 Yes, it doesn't feel like that speed. 155 00:08:12,720 --> 00:08:15,680 And most importantly, you've introduced a bucket load of 156 00:08:15,680 --> 00:08:19,040 latency in order to get there. Yeah, I mean, I can't speak for 157 00:08:19,040 --> 00:08:22,880 the NVIDIA 50 series because I believe their new algorithm and 158 00:08:22,880 --> 00:08:25,880 like the whole thing that they base their multi frame Gen. off 159 00:08:26,600 --> 00:08:29,280 kind of fixes a lot of that. I still wouldn't say it would be 160 00:08:29,280 --> 00:08:34,960 the same as playing at 500 like FPS by taking 350 nineties and 161 00:08:35,039 --> 00:08:38,880 strapping them together, you know, and then playing frame 162 00:08:38,880 --> 00:08:40,320 Gen. I can't really say with that, 163 00:08:40,320 --> 00:08:46,040 but for sure if you were using like AFMF, loss of scaling or 164 00:08:46,040 --> 00:08:49,320 like AMD's frame generation, that kind of thing, the software 165 00:08:49,320 --> 00:08:51,920 based stuff, for sure. You're absolutely right. 166 00:08:51,920 --> 00:08:55,280 But I mean, I've not tested A50 series so I can't say on that 167 00:08:55,280 --> 00:08:58,480 side, but you are absolutely right on that. 168 00:08:58,480 --> 00:09:01,560 I would say for like competitive, just no, like 169 00:09:01,560 --> 00:09:05,200 that's a big no no using frame Gen. who knows about the new 170 00:09:05,200 --> 00:09:08,520 NVIDIA side, but all the previous versions, yeah, don't 171 00:09:09,200 --> 00:09:12,640 like you're you're better off playing at 60 FPS and knowing 172 00:09:12,640 --> 00:09:16,040 how and expecting how that's going to feel like without fake 173 00:09:16,040 --> 00:09:18,600 frames. Upscaling is fine because you're 174 00:09:18,600 --> 00:09:21,000 still getting the correct amount of frames, you know? 175 00:09:21,000 --> 00:09:23,680 It's just a visual thing. But fake frames, obviously frame 176 00:09:23,680 --> 00:09:26,360 Gen. is is is falsified, you know. 177 00:09:26,440 --> 00:09:28,280 And introducing latency into your mix. 178 00:09:28,280 --> 00:09:32,480 So you like you were saying the feeling is a big thing and and 179 00:09:32,480 --> 00:09:37,920 that I think introduces that latency issue because when I 180 00:09:37,920 --> 00:09:40,600 started testing frame Gen. like ages ago, whenever it came out, 181 00:09:41,000 --> 00:09:42,920 I was like, I was trying to make it work magic. 182 00:09:42,960 --> 00:09:45,280 You know, I was like, sweet, I'm going to push my game to the 183 00:09:45,280 --> 00:09:47,720 Max. Like, like burn my computer to 184 00:09:47,720 --> 00:09:48,960 the ground. Let's output. 185 00:09:49,080 --> 00:09:53,480 Yeah, yeah. 4K let's go and you know I've got ray tracing on 186 00:09:53,480 --> 00:09:56,560 like maxed out. Base frame rates 15, Yeah, yeah. 187 00:09:56,640 --> 00:09:59,480 Yeah, so like 20 FPS, you know, and I'm like, sweet, I can make 188 00:09:59,480 --> 00:10:03,120 that 60, you know, and then it still feels like, you know, 189 00:10:03,160 --> 00:10:05,240 20FPS. And it's like, oh, what's going 190 00:10:05,240 --> 00:10:06,960 on here? And that was the same with when 191 00:10:07,120 --> 00:10:10,640 AM MFA FM F2 drops. Those people in the comments 192 00:10:10,640 --> 00:10:12,240 were like, still it's rubbish. Like what you want about. 193 00:10:12,480 --> 00:10:16,000 It's like, well actually you want a minimum base rate of like 194 00:10:16,000 --> 00:10:19,920 45 FPS minimum ideal. You want 60 already. 195 00:10:20,000 --> 00:10:22,680 So it feels like 60, yeah. Exactly so that the game's 196 00:10:22,680 --> 00:10:27,760 running well already and then you're just falsifying up to 120 197 00:10:27,760 --> 00:10:29,160 for. Example, a really important 198 00:10:29,160 --> 00:10:31,760 point, and I've seen you make that a few times in the guides 199 00:10:32,160 --> 00:10:34,440 videos you've done recently with like best settings and how to 200 00:10:34,440 --> 00:10:37,960 use a FM F and stuff. Yeah, that you want to target a 201 00:10:37,960 --> 00:10:43,080 really stable high base rate. If you can get to like a minimum 202 00:10:43,080 --> 00:10:46,040 40FPS and you're really happy 45 FPS and you're really happy with 203 00:10:46,040 --> 00:10:50,000 that look, that's a good target. And then you can apply frame 204 00:10:50,000 --> 00:10:53,200 generation and you'll you'll get a noticeable visual improvement. 205 00:10:53,200 --> 00:10:56,120 You still feel like 45. FPS yeah, yeah, but really 206 00:10:56,120 --> 00:10:59,440 you're seeing like 90 and it just smooths it out, you know, 207 00:10:59,440 --> 00:11:03,200 like it smooths everything out and it will still feel good 208 00:11:03,200 --> 00:11:07,920 because 45 FPS is actually OK. But I'd noticed we've I had 209 00:11:07,920 --> 00:11:11,040 tried games where I was pushing it to the Max and I put it at 210 00:11:11,040 --> 00:11:14,440 like 15 FPS and I'm like, let's try and double it or whatever or 211 00:11:14,440 --> 00:11:16,920 triple it, you know, and it was horrible. 212 00:11:17,160 --> 00:11:20,360 And even though the scream was saying 60 FPS, it's still 213 00:11:20,560 --> 00:11:24,760 juttering because really it's 15 FPS, you know, and I was, I was 214 00:11:24,760 --> 00:11:26,400 saying I couldn't get my head round it. 215 00:11:26,400 --> 00:11:28,080 So I don't get this. And that's when I figured out 216 00:11:28,080 --> 00:11:30,880 myself before I realized everyone else recommends all the 217 00:11:30,880 --> 00:11:32,960 companies say like make sure your frame rate's high. 218 00:11:33,680 --> 00:11:35,720 They don't really like push that enough, I feel. 219 00:11:35,920 --> 00:11:40,080 But once I'd got the the base game to around that 45 mark, 220 00:11:40,560 --> 00:11:43,520 then it was super smooth. It did it's magic because if you 221 00:11:43,520 --> 00:11:44,880 took it off, it was already smooth. 222 00:11:44,880 --> 00:11:45,960 Yeah. Do you know what I mean? 223 00:11:45,960 --> 00:11:47,760 Like, so it's just making it feel even better. 224 00:11:47,760 --> 00:11:49,640 I think that's a. Probably a really important 225 00:11:49,640 --> 00:11:53,520 point when you talk about like frame rates and how it looks and 226 00:11:53,520 --> 00:11:57,120 feels, frame generation is being taught. 227 00:11:57,120 --> 00:11:59,480 And I think This is why I'm a little bit kind of like, oh God, 228 00:11:59,480 --> 00:12:01,560 I'm sick of hearing it. It's being talked about as if 229 00:12:01,560 --> 00:12:03,160 it's like some kind of magic. Yes. 230 00:12:03,240 --> 00:12:05,240 And this is going to solve everyone's problems regardless 231 00:12:05,240 --> 00:12:07,200 of what rig you use. This will solve everything. 232 00:12:07,840 --> 00:12:11,320 It's not. It's a great tool to add a tiny 233 00:12:11,320 --> 00:12:13,880 bit of gains on top of what you already got. 234 00:12:13,880 --> 00:12:18,760 Yes, Yeah, it's not a magic wand that's gonna fix a 15 FBS output 235 00:12:18,920 --> 00:12:21,480 up to 60 and it'd be great. Yes, yeah, yeah, precisely. 236 00:12:21,520 --> 00:12:24,200 It's it's like a little tweak, it's a bit of seasoning to. 237 00:12:24,200 --> 00:12:25,440 Yeah, yeah, yeah, it is. Yeah. 238 00:12:25,440 --> 00:12:28,880 It is just that visual seasoning, seasoning in terms of 239 00:12:28,880 --> 00:12:30,960 fluidity. Treat it like sharpening. 240 00:12:31,080 --> 00:12:33,480 Yes, yeah. Yeah, yeah, a little bit great. 241 00:12:33,480 --> 00:12:35,680 Yes 100% probably don't do it on your 4K screen. 242 00:12:35,680 --> 00:12:38,960 Yes, Yeah, precisely that. Yeah, a bit of it is fine, Yeah. 243 00:12:39,800 --> 00:12:42,440 And it really depends on the app as to if it gives you that 244 00:12:42,680 --> 00:12:45,160 option to be able to do like a little bit or not. 245 00:12:46,080 --> 00:12:47,640 Most of the time it's like 2 times minimum. 246 00:12:48,480 --> 00:12:51,360 But if you've already got that high frame rate, then it can 247 00:12:51,360 --> 00:12:56,200 certainly look good. But I know it's the feel is 248 00:12:56,200 --> 00:12:58,360 different. It doesn't feel different. 249 00:12:58,400 --> 00:13:00,960 Not quite there, you know, like it just feels. 250 00:13:00,960 --> 00:13:03,640 I can feel that latency no matter how small it is. 251 00:13:03,840 --> 00:13:07,000 That's one thing I like about air from F2 is it tells you the 252 00:13:07,160 --> 00:13:09,920 the latency. And the good thing is, is if you 253 00:13:09,920 --> 00:13:13,000 start with a crap frame rate, your latency is going to be 254 00:13:13,000 --> 00:13:14,440 high. So it'll be like 30 255 00:13:14,440 --> 00:13:16,840 milliseconds. Whereas if you get your base 256 00:13:16,840 --> 00:13:18,760 frame rate like drop your settings down, drop your 257 00:13:18,760 --> 00:13:22,360 resolution, whatever, right and get the game let's say to 60, 258 00:13:22,360 --> 00:13:24,080 it'll be like 10 milliseconds delay. 259 00:13:24,480 --> 00:13:27,000 And that's because like we said earlier, there's so many frames, 260 00:13:27,000 --> 00:13:31,520 it's, it's already at a good, you know, stable area that your 261 00:13:31,520 --> 00:13:33,880 latency is going to be lowered and then you can double it 262 00:13:34,120 --> 00:13:37,920 without feeling it too much, but it'll never be the same. 263 00:13:37,920 --> 00:13:40,400 It's just raw power. As RAW powering it out, yeah. 264 00:13:41,360 --> 00:13:45,480 So beyond it, beyond using it sparingly to get a kind of a 265 00:13:45,480 --> 00:13:51,240 marginal visual improvement, there is then the opposite end 266 00:13:51,240 --> 00:13:54,160 of the scale. The cost of this isn't just 267 00:13:54,360 --> 00:13:57,840 input delay, it's the visual artefacting. 268 00:13:58,280 --> 00:14:03,040 And unless there's a, unless there's a generational leap in 269 00:14:03,040 --> 00:14:07,040 the technology and how it's kind of looking ahead as to what's 270 00:14:07,040 --> 00:14:12,480 happening in a frame, I can't see how they get around some of 271 00:14:12,480 --> 00:14:13,760 the visual artefacts that happened. 272 00:14:15,680 --> 00:14:17,160 And why I said earlier, but I can't? 273 00:14:17,440 --> 00:14:19,400 Can't speak to the. 50 series, 50 series. 274 00:14:19,400 --> 00:14:21,920 Because it looks like NVIDIA are doing just this. 275 00:14:22,280 --> 00:14:25,800 They are looking at completely revamping how frame Gen. works 276 00:14:26,200 --> 00:14:31,320 and how you know the algorithm sees what's happening and and 277 00:14:31,320 --> 00:14:34,560 predicts those frames. It just feels like there's 278 00:14:34,960 --> 00:14:38,040 there's always going to be a cost involved in this at one 279 00:14:38,080 --> 00:14:39,640 point in it. It's a trade off, isn't? 280 00:14:39,640 --> 00:14:44,000 It there's a trade off and if they can get, if they can fix 281 00:14:44,000 --> 00:14:46,240 the artifacting, we'll talk about the kind of issues that 282 00:14:46,240 --> 00:14:49,280 you tend to see. But if they can fix, fix that, 283 00:14:49,280 --> 00:14:51,960 that's obviously going to cost time. 284 00:14:52,200 --> 00:14:54,720 Yeah, because there has to be some process that happens. 285 00:14:54,720 --> 00:14:57,120 It's going to introduce more time cost. 286 00:14:57,120 --> 00:14:59,800 Yes, and I think that's why with the NVIDIA stuff specifically, 287 00:14:59,800 --> 00:15:02,520 that's why they just ditched the old model and basically start 288 00:15:02,520 --> 00:15:06,400 again because they couldn't figure out a way of fixing it 289 00:15:06,440 --> 00:15:07,800 until they've done what they've just done. 290 00:15:07,800 --> 00:15:10,720 And that's why I'd love to try the 50 series to see how it 291 00:15:10,760 --> 00:15:12,040 feels. Is this a plug for? 292 00:15:12,040 --> 00:15:15,240 It is. This is our two NVIDIA. 293 00:15:15,240 --> 00:15:18,080 Can you send us 249 fifty 90s please? 294 00:15:18,080 --> 00:15:19,560 You can. Keep your 40. 90s yeah, well 295 00:15:19,600 --> 00:15:22,480 I'll have that instead if you want to just throw away yeah you 296 00:15:22,520 --> 00:15:27,080 know how could we paint? But you know I would love to try 297 00:15:27,080 --> 00:15:30,480 that just to see how it feels and fares against traditional 298 00:15:30,600 --> 00:15:31,880 frame jet, if you want to call it that. 299 00:15:32,120 --> 00:15:35,160 But you're absolutely right, there's going to be a trade off 300 00:15:35,200 --> 00:15:40,520 and I I think the software based stuff is noticeably worse than 301 00:15:40,520 --> 00:15:44,640 even Nvidia's 40 series frame Gen. which I still had an issue 302 00:15:44,640 --> 00:15:46,920 with because that that artifacting. 303 00:15:46,920 --> 00:15:49,160 It's funny isn't it? Because the whole point of frame 304 00:15:49,160 --> 00:15:54,880 Gen. is for a visual increase in frame fluidity, right? 305 00:15:54,880 --> 00:15:58,920 And how the game looks, Yeah, and yet the trade off here. 306 00:15:59,560 --> 00:16:05,000 And also is enables you to run visually high fidelity textures 307 00:16:05,000 --> 00:16:06,680 and stuff. Yes, yeah, yeah, with a higher 308 00:16:06,680 --> 00:16:08,760 frame rate. But the trade off, the biggest 309 00:16:08,760 --> 00:16:12,440 trade off is it looks awful. That you're you're getting 310 00:16:12,480 --> 00:16:15,640 artifacting, which then ultimately lowers the visual 311 00:16:15,640 --> 00:16:17,280 fidelity pulls. You out of the. 312 00:16:17,320 --> 00:16:19,120 Experience and it's jarring, yeah. 313 00:16:19,120 --> 00:16:22,880 I mean I'm I'm sure there's plenty of people that don't even 314 00:16:22,880 --> 00:16:28,600 notice, but for me I'm very particular with stuff looking 315 00:16:28,600 --> 00:16:30,880 weird. Well, I, I would argue right. 316 00:16:30,880 --> 00:16:35,880 If you're in this space enough to be interested in gaining 317 00:16:35,880 --> 00:16:39,240 performance and tweaking your settings to gain performance, 318 00:16:39,920 --> 00:16:42,920 then how things look is really important to you. 319 00:16:42,920 --> 00:16:45,160 And you are going to be the type of person that picks or peeps 320 00:16:45,160 --> 00:16:47,520 probably and notices. And I'm not talking about. 321 00:16:47,720 --> 00:16:49,720 Oh, there's a bit of like compression, artifacting and. 322 00:16:49,720 --> 00:16:52,120 Stuff like that, yeah. Yeah, these are like major, 323 00:16:52,400 --> 00:16:54,640 major artefacts that you. Get to see. 324 00:16:54,640 --> 00:16:57,560 Yeah, to the point where I will disable it if I can. 325 00:16:57,840 --> 00:17:00,240 Well, for me, it pulls me out of the experience. 326 00:17:00,240 --> 00:17:02,880 Exactly that, yes. And so I'd rather run it at a 327 00:17:02,880 --> 00:17:05,640 lower frame rate and be in the experience than be pulled out of 328 00:17:05,640 --> 00:17:08,920 it because it's trying to clone steps as you're walking up with 329 00:17:08,920 --> 00:17:11,640 them. And it looks like a weird I 330 00:17:11,680 --> 00:17:13,800 don't even I don't know how to explain or you get an awful 331 00:17:13,800 --> 00:17:18,319 tiring or the worst that they you were saying that they might 332 00:17:18,319 --> 00:17:21,960 be fixing is the ghosting. Yes, that that again is down to 333 00:17:21,960 --> 00:17:25,319 the NVIDIA side. So they've already shown with 334 00:17:25,319 --> 00:17:29,000 their latest frame Gen. amount that they did like a comparison 335 00:17:29,000 --> 00:17:32,480 between the old model and the new model and the ghosting is 336 00:17:32,480 --> 00:17:37,000 massively reduced, the image quality and sharpness is 337 00:17:37,000 --> 00:17:40,880 massively increased. So everything all round has been 338 00:17:40,960 --> 00:17:45,960 like better, like more stuff flickering, like the anti 339 00:17:45,960 --> 00:17:48,280 aliasing and stuff has all been tidied up. 340 00:17:48,600 --> 00:17:52,360 So specifically in like background images and stuff out 341 00:17:52,360 --> 00:17:55,000 of the centre of the screen, which is always where the 342 00:17:55,200 --> 00:17:58,320 problem is really. So NVIDIA seems to be on to 343 00:17:58,320 --> 00:18:01,480 something. Whether or not it's good I've no 344 00:18:01,480 --> 00:18:03,920 idea. But I've already seen, I think 345 00:18:03,920 --> 00:18:07,440 Digital Foundry did a video and I could already see some of the 346 00:18:07,440 --> 00:18:10,840 artifacts are still there, but obviously it's just better than 347 00:18:10,840 --> 00:18:13,600 it was prior. But like you were saying, some 348 00:18:13,600 --> 00:18:16,160 of the visual artifacts are horrendous. 349 00:18:16,160 --> 00:18:22,440 I was playing Ghosts of Satsuma and I turned on frame generation 350 00:18:22,440 --> 00:18:24,960 on that and I was playing a handheld for a while on the Ally 351 00:18:24,960 --> 00:18:28,320 X with frame Gen. on. I was saying, this is right, I 352 00:18:28,520 --> 00:18:30,040 don't know, there's something going on. 353 00:18:30,040 --> 00:18:32,760 I thought my screen was broken genuinely, and I was, I was a 354 00:18:32,760 --> 00:18:34,280 bit weird. Turned it off, it was gone. 355 00:18:34,400 --> 00:18:35,880 Was that OK? So it's frame Gen. 356 00:18:36,680 --> 00:18:39,520 Put it on the TV where you're stretching it out to a massive 357 00:18:39,520 --> 00:18:42,200 image. And I was like, oh OK, still 358 00:18:42,200 --> 00:18:44,640 happening when it has a problem with the like the tops and 359 00:18:44,640 --> 00:18:47,840 bottoms of the screen. So I had probably an inch from 360 00:18:47,840 --> 00:18:51,280 the bottom if I'm on the TV, if not like 5 millimetres like from 361 00:18:51,280 --> 00:18:56,320 the bottom of my screen, like a whole line from corner to corner 362 00:18:56,600 --> 00:18:58,960 of the entire screen where there was just a tear. 363 00:18:59,200 --> 00:19:01,480 Constantly tearing in and out, Yeah. 364 00:19:01,640 --> 00:19:05,600 In that one area whilst there was like ghosting or flickering 365 00:19:05,600 --> 00:19:08,520 somewhere else and I was like, what's going on here? 366 00:19:08,520 --> 00:19:11,120 Like it's so. Odd bordering tear. 367 00:19:11,160 --> 00:19:14,520 I've seen multiple like watching people's videos on this. 368 00:19:15,080 --> 00:19:17,960 Multiple examples of that, like a really common thing. 369 00:19:17,960 --> 00:19:19,760 Yeah, yeah. It's always on the top or the 370 00:19:19,760 --> 00:19:21,120 bottom or both. Yeah, yeah. 371 00:19:21,200 --> 00:19:23,480 You know it's it's. Either that that straight 372 00:19:23,480 --> 00:19:25,240 because that's so big and visual. 373 00:19:25,240 --> 00:19:27,880 That's like what's going. On yeah, if you notice it, yeah, 374 00:19:27,880 --> 00:19:29,120 you can't Unsee it. Yeah. 375 00:19:29,120 --> 00:19:31,640 So I didn't notice it for a while because a lot of the time 376 00:19:31,640 --> 00:19:32,960 your eyes are in the centre of the screen. 377 00:19:32,960 --> 00:19:34,760 Which is what it's relying on, yes. 378 00:19:34,760 --> 00:19:36,360 Yeah. And so you're like, OK, it's 379 00:19:36,360 --> 00:19:37,720 possible. I'll leave that to go. 380 00:19:37,960 --> 00:19:40,080 But once you notice it, you can't stop looking at it, you 381 00:19:40,080 --> 00:19:40,880 know? What's that? 382 00:19:40,880 --> 00:19:42,760 This thing's flickering. Why is my screen breaking? 383 00:19:42,760 --> 00:19:44,640 It literally looks like the screen's gonna turn off in a 384 00:19:44,640 --> 00:19:46,400 minute because Joe. I mean, it's like, Oh no, it's 385 00:19:46,400 --> 00:19:48,000 expected deadline. Yeah, yeah. 386 00:19:48,480 --> 00:19:50,680 So it's like, Oh no. So there's that. 387 00:19:50,840 --> 00:19:53,680 There's then like weird warping and stuff. 388 00:19:54,200 --> 00:19:57,400 So you'll find if there's a straight line and your your 389 00:19:57,400 --> 00:19:59,480 guy's got like a pistol or something and you're pulling in 390 00:19:59,480 --> 00:20:01,600 and out. I showed this on Resident Evil 4 391 00:20:01,640 --> 00:20:05,880 with loss of scaling. So where there's like a wall 392 00:20:05,880 --> 00:20:09,520 edge is trying to artificially like fake what's around it 393 00:20:09,560 --> 00:20:12,400 because it's so clear cut like in Chris, what's happening? 394 00:20:12,920 --> 00:20:15,240 You can see it trying to. It literally looks like you've 395 00:20:15,240 --> 00:20:18,240 got one of those like, balls, see through balls and you look 396 00:20:18,240 --> 00:20:19,360 for it. Yeah, yeah. 397 00:20:19,360 --> 00:20:21,520 And it warps in. It looks like you're pushing one 398 00:20:21,520 --> 00:20:23,960 of these spheres, like backwards and forwards. 399 00:20:23,960 --> 00:20:26,360 It's really weird. Your example in Rebirth Final 400 00:20:26,680 --> 00:20:29,160 Fantasy 7 Rebirth going up the stairs. 401 00:20:29,160 --> 00:20:31,160 Yeah, that's horrendous. Which is another example of like 402 00:20:31,160 --> 00:20:34,280 the defined straight lines, straight lines of steps that 403 00:20:34,280 --> 00:20:35,840 you're walking up. Yeah, straight lines are a real. 404 00:20:35,840 --> 00:20:38,000 Problem and they're moving. Yes, because it's moving. 405 00:20:38,000 --> 00:20:40,920 It's trying to predict where that straight line should be in 406 00:20:40,920 --> 00:20:42,480 the artificial frames in between. 407 00:20:42,480 --> 00:20:44,640 Yeah. And because of that, it it's 408 00:20:44,640 --> 00:20:46,920 going to get it wrong. Yeah, Yeah, because there's such 409 00:20:46,920 --> 00:20:49,000 straight lines, but the camera's moving and the character's 410 00:20:49,000 --> 00:20:51,920 moving over the top of it. So you get first you get this 411 00:20:51,920 --> 00:20:55,480 weird was a ghosting effect around where the character 412 00:20:55,480 --> 00:20:56,320 would. Be. 413 00:20:56,320 --> 00:20:59,080 Yeah, yeah. And you also get the lines of 414 00:20:59,080 --> 00:21:03,880 change position or they've bent. That's doesn't just make me feel 415 00:21:03,880 --> 00:21:06,680 a bit sick. Like I say, it pulls me out of 416 00:21:06,680 --> 00:21:08,360 the game. Yes, now I'm worried about what 417 00:21:08,360 --> 00:21:10,000 technical thing is going wrong here. 418 00:21:10,000 --> 00:21:13,160 Yes, Yeah, yeah, yeah. So for a tool to pull you out of 419 00:21:13,160 --> 00:21:15,040 the game means the experience has been ruined. 420 00:21:15,040 --> 00:21:17,040 Because now my head's in. How do I fix that problem? 421 00:21:17,040 --> 00:21:19,080 Yeah, yeah, something's broken. Yeah, let me fix it. 422 00:21:19,080 --> 00:21:20,320 You're not enjoying the game anymore. 423 00:21:20,320 --> 00:21:22,560 You're not in that world. Whatever, you know. 424 00:21:22,560 --> 00:21:25,520 So that's the thing I found. And this is the, the other 425 00:21:25,520 --> 00:21:29,440 slight, the difference between the the other two types of frame 426 00:21:29,440 --> 00:21:33,000 Gen. where you've got frame Gen. that's been implemented by the 427 00:21:33,000 --> 00:21:35,920 developers. So that has been this works, 428 00:21:35,920 --> 00:21:39,960 we've built this into the game. And then software or driver 429 00:21:39,960 --> 00:21:43,040 based frame Gen. that you're slapping over yourself like AFMF 430 00:21:43,240 --> 00:21:47,640 or lossless scaling. So AMD's FSR 3 frame Gen. is 431 00:21:47,640 --> 00:21:51,040 what will be pushed in from the developer as well as DLSS frame 432 00:21:51,040 --> 00:21:53,000 Gen. But I'm just talking about FSR 3 433 00:21:53,000 --> 00:21:57,680 here for a second that so let's take spider Man Miles Morales 434 00:21:57,680 --> 00:21:59,560 because it's something I played the past couple of months or 435 00:21:59,560 --> 00:22:02,800 whatever that has been put in there by the devs. 436 00:22:03,000 --> 00:22:03,720 So they've gone. Yep. 437 00:22:04,120 --> 00:22:07,960 And what that means is it will apply frame Gen. to the game but 438 00:22:07,960 --> 00:22:12,320 not the overlay. So the HUD is all all the 439 00:22:12,600 --> 00:22:14,760 graphics and stuff are not being framed Gen. 440 00:22:14,760 --> 00:22:16,960 D. So that's good because it means 441 00:22:16,960 --> 00:22:20,400 none of that gets artifacting. But the game will still have 442 00:22:20,400 --> 00:22:23,600 this tearing or you might have ghosting or like weird warping 443 00:22:23,600 --> 00:22:26,560 like bits like here and then or people like smudging into the 444 00:22:26,560 --> 00:22:27,720 wall. Do you know what I mean? 445 00:22:27,720 --> 00:22:30,760 Like you get like all these weird problems, which that is 446 00:22:30,760 --> 00:22:33,440 the best use case of software based. 447 00:22:33,720 --> 00:22:38,560 If a game hasn't got it put in like Rebirth Final Fantasy 7 and 448 00:22:38,560 --> 00:22:43,080 you force it over the top with lossless or AFMF then it can't 449 00:22:43,080 --> 00:22:46,880 differentiate what the game is and what the overlay is, so it 450 00:22:46,880 --> 00:22:48,280 just frames everything because it's. 451 00:22:48,280 --> 00:22:50,560 Just saying the whole picture. It's saying the entire thing. 452 00:22:50,560 --> 00:22:56,720 So like with Resident Evil 4 with the crosshair, Leon's gun, 453 00:22:57,000 --> 00:22:59,280 I was moving that around load and it was like there was like 454 00:22:59,280 --> 00:23:01,520 loads of them because it was like, whoa, what's that little 455 00:23:01,520 --> 00:23:03,880 lion doing? You know, So your crosshairs 456 00:23:03,880 --> 00:23:05,360 like completely all over the place. 457 00:23:05,360 --> 00:23:07,800 Obviously you can tune it a bit depending on your settings, and 458 00:23:07,840 --> 00:23:10,320 I'll show you how to do that in that guide, but it's still 459 00:23:10,320 --> 00:23:12,960 there. So I think truthfully it's a 460 00:23:12,960 --> 00:23:17,800 case of right now, I like the idea, OK. 461 00:23:17,880 --> 00:23:21,240 I like the idea of a tool that can increase performance without 462 00:23:21,240 --> 00:23:23,680 having like hardware like limitations. 463 00:23:23,720 --> 00:23:25,760 That's kind of cool, having to upgrade your gear. 464 00:23:25,760 --> 00:23:27,280 Yes. And giving you longevity to 465 00:23:27,280 --> 00:23:30,640 older devices and especially with more and more PC handhelds, 466 00:23:30,640 --> 00:23:32,920 which we'd love yes, they're always going to be slightly 467 00:23:32,920 --> 00:23:34,880 underpowered compared to a desktop, absolutely. 468 00:23:35,680 --> 00:23:37,880 So being able to get squeeze more performance out of those 469 00:23:38,080 --> 00:23:43,720 awesome, but it's really, it's really imperfect. 470 00:23:43,720 --> 00:23:47,600 Yes, yeah, yeah. And I want to say because it's 471 00:23:47,600 --> 00:23:49,200 early days, but it's not that new. 472 00:23:49,200 --> 00:23:51,400 No, it's been around. And it has been rolled out. 473 00:23:51,400 --> 00:23:53,440 So it's kind of inexcusable in some, in a lot of. 474 00:23:53,440 --> 00:23:57,960 Cases, well I think we're on that cusp like the precipice now 475 00:23:57,960 --> 00:24:02,320 of where we go from here. So NVIDIA seem to be whilst 476 00:24:02,320 --> 00:24:05,280 people are like hating on them for taking the fake frames 477 00:24:05,280 --> 00:24:10,960 route, they are trying to make games better, you know, by using 478 00:24:11,120 --> 00:24:14,120 AI and stuff you know. So whilst we might not be there 479 00:24:14,120 --> 00:24:17,920 now, maybe this is the start of it and it already looks like the 480 00:24:17,920 --> 00:24:20,680 50 series from what I've seen examples wise. 481 00:24:20,680 --> 00:24:23,600 And people like Digital Foundry showing it is definitely better 482 00:24:23,600 --> 00:24:27,360 than it was. We know AMD have turned around, 483 00:24:27,360 --> 00:24:29,520 held their hands up and gone. You know what, we're not doing 484 00:24:29,520 --> 00:24:31,400 software based anymore. They've said that. 485 00:24:31,400 --> 00:24:32,680 So they have. Yeah, yeah, yeah. 486 00:24:32,680 --> 00:24:36,360 So when RDNA 4 comes, This is why I'm really hyped for when we 487 00:24:36,360 --> 00:24:40,080 get an RDNA four hand. 3Z. 3 extreme and stuff like that. 488 00:24:40,400 --> 00:24:42,800 That's then going to start hardware encoding. 489 00:24:43,320 --> 00:24:47,960 So we will then get non software based frame Gen. and upscaling 490 00:24:47,960 --> 00:24:50,960 and that that is like AM DS looked at, you know, their 491 00:24:50,960 --> 00:24:54,520 cousin NVIDIA and gone. All right, we should copy that. 492 00:24:54,520 --> 00:24:58,040 When you consider the prospect of that from the developer side, 493 00:24:58,040 --> 00:25:00,440 it's a lot easier for them to account for it when they're 494 00:25:00,880 --> 00:25:05,080 developing the game and develop that with that in mind, which 495 00:25:05,080 --> 00:25:08,960 we'll see more and more of. So certainly an exciting idea, 496 00:25:09,560 --> 00:25:14,200 but I guess our kind of like reluctance to it is that it's 497 00:25:14,200 --> 00:25:17,960 being used too much as a a crotch to achieve big numbers. 498 00:25:17,960 --> 00:25:21,400 Yeah, a crotch even. I like it being used as a 499 00:25:21,400 --> 00:25:22,840 crotch. You can use it as your crotch, I 500 00:25:22,880 --> 00:25:24,440 don't mind. Frame your crotch. 501 00:25:24,440 --> 00:25:30,560 Yeah, that's really Yeah. I've got 100 FPS down here. 502 00:25:30,560 --> 00:25:34,680 No way. So we break out at the club. 503 00:25:34,680 --> 00:25:39,400 Yeah, my crotch is 240 FPS motion. 504 00:25:39,400 --> 00:25:44,760 Blows off. Yeah, no, I agree with you 505 00:25:44,760 --> 00:25:46,120 though. Like I feel like it's been 506 00:25:46,120 --> 00:25:49,080 relied on too much, too much. To achieve big numbers. 507 00:25:49,080 --> 00:25:51,200 Yes, that I'll. Go back to the original point, 508 00:25:51,200 --> 00:25:53,600 don't feel like the big numbers, they still feel like the base 509 00:25:53,600 --> 00:25:58,280 rate, have massive delay introduced and have huge 510 00:25:58,280 --> 00:26:01,240 artifacting problems. But yes, these things are being 511 00:26:01,360 --> 00:26:04,640 ironed out, but they're not yet. There yet we're not there yet. 512 00:26:04,760 --> 00:26:07,320 No, not at all. And I, and I feel like for me 513 00:26:07,320 --> 00:26:10,680 personally, like whilst you can use it if you're absolutely 514 00:26:10,680 --> 00:26:14,120 struggling, I would only use it if like the device I'm using 515 00:26:14,120 --> 00:26:15,920 just is really having a hard time. 516 00:26:16,200 --> 00:26:19,000 But again, even then that's not a great idea because we've 517 00:26:19,000 --> 00:26:21,840 already said you want at least 45 FPS. 518 00:26:22,440 --> 00:26:24,240 Look at this the implementation on the Steam Deck. 519 00:26:24,240 --> 00:26:25,280 Do you watch any of the videos of? 520 00:26:25,280 --> 00:26:27,120 It yeah, that's. Horrendous and horrendous. 521 00:26:27,160 --> 00:26:29,800 And, and I like, I'm not hating on the Steam Deck at all. 522 00:26:29,800 --> 00:26:33,360 I love the Steam Deck. But I mean those frame Gen. you 523 00:26:33,360 --> 00:26:36,040 know, there's basically for you that don't know if a game is 524 00:26:36,040 --> 00:26:40,800 DirectX 12 and supports DLSS frame Gen. you can do a hack 525 00:26:40,800 --> 00:26:45,160 that injects FSR 3 frame Gen. instead of DLSS frame Gen. 526 00:26:45,240 --> 00:26:48,320 So in the game you would say DLSS frame Gen. but actually 527 00:26:48,320 --> 00:26:53,120 you're like faking in FSR 3 frame Gen. on the Steam Deck. 528 00:26:53,120 --> 00:26:56,040 So it does work but you have to meet those criteria. 529 00:26:56,040 --> 00:27:00,200 It's direct X12 and DLSS frame Gen. and it it looks horrendous 530 00:27:00,520 --> 00:27:01,920 like. And the latency. 531 00:27:01,920 --> 00:27:05,120 Yeah, the latency is terrible, but the image quality is even 532 00:27:05,120 --> 00:27:07,680 worse than, you know, native on it. 533 00:27:07,680 --> 00:27:09,040 So I. Totally agree. 534 00:27:09,040 --> 00:27:10,640 And the artifacting and stuff, yeah, it's. 535 00:27:10,840 --> 00:27:13,400 It's like. Amplified banding and tearing. 536 00:27:13,440 --> 00:27:15,440 Awful. I think it's more noticeable in 537 00:27:15,440 --> 00:27:19,720 the Steam Deck because a it's 800 P to start with, it's 538 00:27:19,720 --> 00:27:24,200 already struggling to output at that point with modern games, so 539 00:27:24,200 --> 00:27:27,680 people are probably blowing the resolution to probably about 540 540 00:27:27,680 --> 00:27:30,160 P, So it really doesn't look good anyway. 541 00:27:30,480 --> 00:27:33,120 And then you're applying frame Gen. on top of that to try and 542 00:27:33,120 --> 00:27:36,440 get a decent FPS. So you've got like really low 543 00:27:36,440 --> 00:27:39,280 resolution with loads of artifacts with very low 544 00:27:39,280 --> 00:27:41,440 textures. It just makes it look like a 545 00:27:41,440 --> 00:27:44,560 water colour painting, you know it's a big mush on the screen. 546 00:27:44,560 --> 00:27:48,840 It's great that these tools can exist to increase hardware life, 547 00:27:48,840 --> 00:27:50,800 but they are not magic ones. No, no, no. 548 00:27:50,800 --> 00:27:52,080 I think that's the biggest statement. 549 00:27:52,080 --> 00:27:54,640 For me, I think what you said earlier is the thing that resin 550 00:27:55,200 --> 00:27:58,480 like resonated with me was it pulls me out of the experience 551 00:27:58,920 --> 00:28:01,560 and every game I've tried it on, there's not one game that I've 552 00:28:01,560 --> 00:28:04,560 tried it on where I've gone sweet, I'll leave it on every 553 00:28:04,560 --> 00:28:07,440 time I've gone let me try and tweak this And I spent like 2 554 00:28:07,440 --> 00:28:10,880 hours trying to get frame Gen. to be good and I'm like and that 555 00:28:11,120 --> 00:28:14,360 and ultimately I always resort to just disabling it. 556 00:28:14,640 --> 00:28:17,880 I'd much rather turn my settings down, whatever to get a stable 557 00:28:17,880 --> 00:28:20,760 FPS that I'm happy with and just play at that. 558 00:28:21,120 --> 00:28:24,240 And that's I can't do it because though any artifacting I just 559 00:28:24,240 --> 00:28:25,720 go. It's annoying. 560 00:28:25,760 --> 00:28:27,280 Yeah, I feel like something's broken. 561 00:28:27,840 --> 00:28:31,040 That's literally how I feel. So I'm just like, Nah, no frame 562 00:28:31,040 --> 00:28:32,880 job. I'd love to know what you guys 563 00:28:32,880 --> 00:28:35,160 think. Do you agree with us or have we 564 00:28:35,160 --> 00:28:38,920 triggered you? A huge fan of Frame Gen. 565 00:28:39,280 --> 00:28:41,840 You're a paid subscriber? I don't know. 566 00:28:41,960 --> 00:28:44,680 Offramegen.com. Yeah, Frame gen.com. 567 00:28:44,680 --> 00:28:47,280 You, you know, you love it. That's no doubt. 568 00:28:47,280 --> 00:28:49,520 Have you the founder? Wow. 569 00:28:50,000 --> 00:28:51,680 Did you invent Frame Gen. Oh. 570 00:28:51,680 --> 00:28:52,960 No, I was just thinking of the website. 571 00:28:53,400 --> 00:28:55,160 Is that a website? Yeah, probably. 572 00:28:55,200 --> 00:29:00,600 Probably, yeah. OK, so do you ever like, go? 573 00:29:00,600 --> 00:29:03,360 There's a stationary store that I heard you like to go to. 574 00:29:03,520 --> 00:29:06,720 Ohyeahpenisland.com. That's where you go to increase 575 00:29:06,720 --> 00:29:08,280 your crotch. Isn't it? 576 00:29:09,040 --> 00:29:10,200 Oh damn it, you knew what I was all about. 577 00:29:11,680 --> 00:29:14,800 Stupid jokes are also included for free, so make sure you 578 00:29:14,800 --> 00:29:17,440 subscribe. Yeah, like before you leave. 579 00:29:17,520 --> 00:29:19,040 Oh I love that it's the new sayings. 580 00:29:19,040 --> 00:29:21,920 Like before you leave. Oh yeah, and. 581 00:29:22,200 --> 00:29:24,000 How come I remember? Yeah, that too. 582 00:29:24,040 --> 00:29:26,440 And you can talk to us over on Discord and tell us how much 583 00:29:26,440 --> 00:29:27,720 we're wrong. Yeah, that's. 584 00:29:27,720 --> 00:29:29,800 Over there and stupid jokes are also there. 585 00:29:29,800 --> 00:29:30,960 Yeah, they are there. Yeah. 586 00:29:31,200 --> 00:29:32,600 So until the next time, bye.