Support NeoGAF

[megreotsugua]

The Sun is static in this game. The only time it changes is when you change it in debug mode.

But aren't they using Lumen here? That's still a significant power hog as opposed to baked lighting.

I believe it's not the nanite (details) that's expensive but the lumen (lighting). UE5 is impressive because devs will no longer create lods and bake lighting. But Insomniac can make the same amount of details the hard way and bake lighting the hard way while also saving on performance.

The good thing about UE5 is that it saves a lot of dev time. But it doesn't mean that AAA devs can't surpass it by exerting more development time. Especially if the game won't have night and day cycle.

 

[kikkis]

But aren't they using Lumen here? That's still a significant power hog as opposed to baked lighting.

I believe it's not the nanite (details) that's expensive but the lumen (lighting). UE5 is impressive because devs will no longer create lods and bake lighting. But Insomniac can make the same amount of details the hard way and bake lighting the hard way while also saving on performance.

The good thing about UE5 is that it saves a lot of dev time. But it doesn't mean that AAA devs can't surpass it by exerting more development time. Especially if the game won't have night and day cycle.

Software rasterazation of nanite works around ineffeciency of hw rast on small triangles, so you can't make more detail with traditional methods because its so slow on hw. Modern baking pipeliness aren't that cumbersome to use and can have little interactivity like cod mw. So I doubt lumen is going to be used that much since fully dynamic environments aren't common.

 

[OsirisBlack]

Just getting to try this out, pretty impressive stuff. Turning off the matrix filer makes the game look very very good

 

[Rubim]

will this tflops talk never end? really??

30 tflops rtx 3080 is barely pushing %40-50 more fps than series x in halo infinite/forza horizon 5. its a useless and bloated metric that nvidia boosted to get a hold in "muh tflops wars". it cannot even realistically push actual 2x performance over series x, despite having "2.5x" tflops.

NVIDIA GeForce GTX 780 Specs

NVIDIA GK110, 902 MHz, 2304 Cores, 192 TMUs, 48 ROPs, 3072 MB GDDR5, 1502 MHz, 384 bit

www.techpowerup.com

look at this sweet 4.2 tflops kepler monster, its fire and lit, it must match ps4 pro right?

oh... it barely matches the 1.8 tflops ps4... okay...



i guess...

how about horizon zero dawn at medium settings (original ps4 preset). can the mighty 4.2 tflops super mega beast gtx 780 can handle that?



whoops. this is literally worse performance than 1.8 tflops ps4. nice, i guess. stuff happens huh?

im pretty sure matrix awakens demo if released right now would run horrible on ampere GPUs. their async implementation is still wonky/kind of fake. they don't even have a proper hardware scheduler and instead relies on CPU cycles to do the task.

ampere is kepler all over again. stop using the bloated tflops numbers as a argument to downplay 4/12 tflopos that consoles range. 1.8 tflops ps4 survived the entire gen at respectable resolutions. new consoles have even much stronger CPU that would scale the playing field

Just a correction, you can't get the same settings on RDR2 PC vs CONSOLES, Consoles has settings lower than low.

 

[Rubim]

will this tflops talk never end? really??

30 tflops rtx 3080 is barely pushing %40-50 more fps than series x in halo infinite/forza horizon 5. its a useless and bloated metric that nvidia boosted to get a hold in "muh tflops wars". it cannot even realistically push actual 2x performance over series x, despite having "2.5x" tflops.

NVIDIA GeForce GTX 780 Specs

NVIDIA GK110, 902 MHz, 2304 Cores, 192 TMUs, 48 ROPs, 3072 MB GDDR5, 1502 MHz, 384 bit

www.techpowerup.com

look at this sweet 4.2 tflops kepler monster, its fire and lit, it must match ps4 pro right?

oh... it barely matches the 1.8 tflops ps4... okay...



i guess...

how about horizon zero dawn at medium settings (original ps4 preset). can the mighty 4.2 tflops super mega beast gtx 780 can handle that?



whoops. this is literally worse performance than 1.8 tflops ps4. nice, i guess. stuff happens huh?

im pretty sure matrix awakens demo if released right now would run horrible on ampere GPUs. their async implementation is still wonky/kind of fake. they don't even have a proper hardware scheduler and instead relies on CPU cycles to do the task.

ampere is kepler all over again. stop using the bloated tflops numbers as a argument to downplay 4/12 tflopos that consoles range. 1.8 tflops ps4 survived the entire gen at respectable resolutions. new consoles have even much stronger CPU that would scale the playing field

Im retarded and i can't edit my post but.

Comparing TFLOPS from different architetures is dumb. If you're checking for Tflops then you do it with NVIDIA vs NVIDIA, not with AMD vs NVIDIA.

Consoles were comparable to a Radeon 7850, maybe a downclocked 7870. It's really hard to make a comparision based on specs alone of different architetures.

 

[Bogroll]

The people that bought an XSX and PS5 do. The people that bought an XSS because they could not get an XSX caring or not caring does not make much of a difference to them, the others … well there is quite a lot to say about that but there is a topic in and of itself (would they be much better served by a cheaper digital only XSX? Cheerful if they do not see how downscaling from 4K to 1080p and having better effects and textures makes the graphics look better? Cheerful they get One S versions of BC titles and not the One X? Etc…) … still not sure how this sounds anything less than a corporate “machine is selling, it is good because it is selling” way to go around the point.

Xbox Series S was a strategic pincher movement thought to hurt a single expensive PS5 SKU, but while it has some pros it also has some cons.

You don't know that and do you really think it was wow moment for those in the know or even any of us when the digital PS5 announced. It wasn't like the reveal of the wheel.
I'd guess it was more of a GP machine, cheap and a alternative to any PS5 for those who are not so bothered about graphics etc.

 

[yamaci17]

Im retarded and i can't edit my post but.

Comparing TFLOPS from different architetures is dumb. If you're checking for Tflops then you do it with NVIDIA vs NVIDIA, not with AMD vs NVIDIA.

Consoles were comparable to a Radeon 7850, maybe a downclocked 7870. It's really hard to make a comparision based on specs alone of different architetures.

bro, my entire point was that we should not compare tflops / architectures. my post was targeted at people who keeps using bloated Ampere numbers as a way to downplay consoles. can you please read the posts before me? there was a person that is saying there are 20+ tflops GPUs in the market and 4 tflops is funny compared to them. so... you're missing my point

"Who would have thought it would look this good on a 4 TF, yes 4 TF machine … while PCs are going to 24 TF soon .."


as i've said, its about the optmization. kepler was a beast. in games too. it all went downfall once the maxwell is released. i vividly remember this topic and the outrage that followed it.

The Witcher 3 - Kepler GPU

The Witcher 3 - Kepler GPU gtx 960=gtx 780 gtx 970>gtx 780ti what's going on here ? how is this possible ? nvidia is deliberately downgrading kepler gpu's performance in favor of maxwell using driver or poor optimization game under 700 card by cdprojektred ? I read the forums, and I'm in shock...

forums.cdprojektred.com

this was the last game that kepler got "special" driver codepath optimizations. after that, it was mostly a wreck.

i'm not saying ps4 runs the rdr 2 at high settings, but it looks gorgeous. gtx 780 clip however looks atrocious.



ps4 settings are a mixed bag. yes, there are stuff that are lower than low, but then again, there are lots of stuff that are at medium/high/ultra. like shadows, ssao, tesselation, particles, volumetric resolution, geometry lod (ultra), textures (ultra). game simply looks beautiful on a ps4. i just tried to show it as an example to prove that TFLOPS are not be-all end-all.

in the end, 4 tflops of rdna 2 on an optimized space might output the performance of a 8-10 tflops ampere gpu. who knows? why downplay 4 tflops? its clear that 1.8 tflops of old GCN can provide exceptional performance in games like rdr 2, horizon zero dawn, god of war and many more. even doom eternal runs at 1080p 60 fps on that thing. 4 tflops of an advanced, newer architecture should give enough juice to series s for its lifetime and its target resolution. at least I hope so.

 

[Lethal01]

The Sun is static in this game. The only time it changes is when you change it in debug mode.

It's not treated as static, the light is calculated as dynamically as if it were moving slowly. The lighting of this demo isn't baked.

 

[megreotsugua]

Software rasterazation of nanite works around ineffeciency of hw rast on small triangles, so you can't make more detail with traditional methods because its so slow on hw.

I really doubt Sony's first party will use UE5 because of the 5% cut. I just hope they have something up their sleeve than the traditional method. I'd like spider-man 2 to look this good.

 

[RoadHazard]

It's not treated as static, the light is calculated as dynamically as if it were moving slowly. The lighting of this demo isn't baked.

It takes a significant amount of time for it to calculate the new lighting when changing from day to night though (I mean, not LONG, but a very noticeable amount of time), it's not like it's fully recalculated every frame. Which makes me wonder how it would look if the sun was continually moving, or if you wanted dynamic changes in response to destruction etc.

 

[DenchDeckard]

When we start looking at price to performance, the series s does not fare well since it offers 1/4 the performance more than half the price. Pretty much every game that runs at native 4k on the XSX runs at 1080p on the Series s with lower quality settings. Thats 1/4th the resolution. Guardians, Halo Infinite, Forza Motorsport. Even in those rare occasions when we get 1440p vs 4k versions, the settings are dialed down.

If we are looking at the price of performance ratio, a console offering 1/4th performance should cost 1/4th less, right? That should put the XSS at $125 when compared against what the XSX is offering. $100 when compared against the PS5.

this is a bad comparison, bro. You cant just choose one thing to compare like the resolution and stick to that. thats Like saying a 1660 super should be one quarter of the price of a 3080 and we know it isn’t. The series s is an entire console with the same cpu and more or less same speed ssd etc.

 

[Panajev2001a]

You don't know that and do you really think it was wow moment for those in the know or even any of us when the digital PS5 announced. It wasn't like the reveal of the wheel.
I'd guess it was more of a GP machine, cheap and a alternative to any PS5 for those who are not so bothered about graphics etc.

They might have known earlier than we did, but they also finalised their plans before we know they did.

Yes, I do think MS did not expect the $399 PS5 and they setup a brilliant strategy in that scenario, you are quoting other reasons for it to exist (that I disagree with if you compare it to say a $349-399 512 GB of SSD digital only XSX, it would have provided more benefits to users and devs… if you really do not care much about graphics you stay one more year or so on an XOX, PS4, etc…)

 

[SlimySnake]

will this tflops talk never end? really??

30 tflops rtx 3080 is barely pushing %40-50 more fps than series x in halo infinite/forza horizon 5. its a useless and bloated metric that nvidia boosted to get a hold in "muh tflops wars". it cannot even realistically push actual 2x performance over series x, despite having "2.5x" tflops.

NVIDIA GeForce GTX 780 Specs

NVIDIA GK110, 902 MHz, 2304 Cores, 192 TMUs, 48 ROPs, 3072 MB GDDR5, 1502 MHz, 384 bit

www.techpowerup.com

look at this sweet 4.2 tflops kepler monster, its fire and lit, it must match ps4 pro right?

oh... it barely matches the 1.8 tflops ps4... okay...



i guess...

how about horizon zero dawn at medium settings (original ps4 preset). can the mighty 4.2 tflops super mega beast gtx 780 can handle that?



whoops. this is literally worse performance than 1.8 tflops ps4. nice, i guess. stuff happens huh?

im pretty sure matrix awakens demo if released right now would run horrible on ampere GPUs. their async implementation is still wonky/kind of fake. they don't even have a proper hardware scheduler and instead relies on CPU cycles to do the task.

ampere is kepler all over again. stop using the bloated tflops numbers as a argument to downplay 4/12 tflopos that consoles range. 1.8 tflops ps4 survived the entire gen at respectable resolutions. new consoles have even much stronger CPU that would scale the playing field

Lmao. Why are you comparing nvidia tflops to AMD tflops? Everyone who knows anything about pc gaming knows never to do that. They were only comparable for about a year when rdna 1.0 cards and rtx 20 series cards had the same tflops and performance and only for two cards, the 2060 and 5700, and the 2070 and 5700xt. Since then nvidia inflated their tflops count but you can always compare rdna 2.0 cards on PC very easily.

Here is a forza benchmark. Notice how performance scales roughly in line with the tflops gap.

5500 xt - 5.2 tflops- 36 fps
6600 xt - 10.6 tflops - 67 fps
6300 xt - 13.2 tflops - 86 fps
6800 xt - 20.7 tflops - 117 fps
6900 xt - 23 tflops - 126 fps

The 13tf 6700xt and 5tf 5500 xt is a great comparison because they are very comparable to the xsx and xsx. 8 tflops difference between the two. the tflops differnce between them is 2.5x and the performance difference is 2.3x.

The 10.6 tflops 6600xt and 20.7 tflops have a 1.95x tflops difference and a 1.75x performance difference.

The 10.6tf 6800xt and the 13.2 tflops 6700xt have a 1.25x tflops difference and 1.28x performance difference.

Tflops and performance scale linearly when you are comparing the same family of GPUs. That’s always been true. Its how GPUs are priced. It’s how they are benchmarked. If anything the higher end cards end up performing slightly worse due to various issues from making out bandwidth or pushing CU counts too high. The xsx and xss have a 3x tflops differential and a 4x performance differential. The xss should be offering 1/3rd the performance of xsx but it only offers 1/4th and that too with graphics settings paired back.

 

[SlimySnake]

this is a bad comparison, bro. You cant just choose one thing to compare like the resolution and stick to that. thats Like saying a 1660 super should be one quarter of the price of a 3080 and we know it isn’t. The series s is an entire console with the same cpu and more or less same speed ssd etc.

But you are the one who brought up price to performance.

it doesn’t deserve all its hate. If you look at price vs performance Imo.

 

[DenchDeckard]

Lmao. Why are you comparing nvidia tflops to AMD tflops? Everyone who knows anything about pc gaming knows never to do that. They were only comparable for about a year when rdna 1.0 cards and rtx 20 series cards had the same tflops and performance and only for two cards, the 2060 and 5700, and the 2070 and 5700xt. Since then nvidia inflated their tflops count but you can always compare rdna 2.0 cards on PC very easily.

Here is a forza benchmark. Notice how performance scales roughly in line with the tflops gap.

5500 xt - 5.2 tflops- 36 fps
6600 xt - 10.6 tflops - 67 fps
6300 xt - 13.2 tflops - 86 fps
6800 xt - 20.7 tflops - 117 fps
6900 xt - 23 tflops - 126 fps

The 13tf 6700xt and 5tf 5500 xt is a great comparison because they are very comparable to the xsx and xsx. 8 tflops difference between the two. the tflops differnce between them is 2.5x and the performance difference is 2.3x.

The 10.6 tflops 6600xt and 20.7 tflops have a 1.95x tflops difference and a 1.75x performance difference.

The 10.6tf 6800xt and the 13.2 tflops 6700xt have a 1.25x tflops difference and 1.28x performance difference.

Tflops and performance scale linearly when you are comparing the same family of GPUs. That’s always been true. Its how GPUs are priced. It’s how they are benchmarked. If anything the higher end cards end up performing slightly worse due to various issues from making out bandwidth or pushing CU counts too high. The xsx and xss have a 3x tflops differential and a 4x performance differential. The xss should be offering 1/3rd the performance of xsx but it only offers 1/4th and that too with graphics settings paired back.

the series s isn’t offering 1/4 on every game. On some it is 1440p and the big brother consoles are using drs too so they aren’t native 4k either.

i Appreciate the info though, it is interesting to see how this stuff stacks up.

the thing I don’t get is series x is not showing it’s advantages on the matrix demo over ps5 and I would love to know why.

 

[DenchDeckard]

But you are the one who brought up price to performance.

yeah but then compare other things such as the ssd speed being more or less identical, the cpu being identical. It more or less offers similar performance overall but the resolution takes a hit and in some situations dialled back settings, which is completely acceptable imo.

If I specd a pc with the same cpu and mobo 8 gb vs 16gb memory, 512 gb nvme vs 1tb and 1660 super vs 2080 super you would probs have to do similar settings dialed back and drop resolution to get a comparable framerate and there would only be a few hundred pounds between the two.

I dunno why I’m comparing pc to console but it’s all I can do lol.

I personally don’t think the series s is too bad for its price Yet it’s constantly picked on. which I feel is unwarranted,but each to their own. It seems to be doing quite we’ll in sales so I guess there is a lot of people who share the sentiment.

 

[SlimySnake]

the series s isn’t offering 1/4 on every game. On some it is 1440p and the big brother consoles are using drs too so they aren’t native 4k either.

i Appreciate the info though, it is interesting to see how this stuff stacks up.

the thing I don’t get is series x is not showing it’s advantages on the matrix demo over ps5 and I would love to know why.

It’s very rare for it to offer 1440p modes. And even then there are downgrades in graphics settings. The vast majority of games including the most recent high profile ones are 1/4th.

I wouldn’t be too sure about series x not showing its advantage. The DF clip i saw had the ps5 dip to 20 fps far longer than the xsx during a car crash. Xsx drops for maybe 1-2 seconds, ps5 felt like 3-4.

They need to do better tests. Preferably without the same cars. And calculate the minimum resolution at the time of the dip to 20 fps.

 

[yamaci17]

Lmao. Why are you comparing nvidia tflops to AMD tflops? Everyone who knows anything about pc gaming knows never to do that. They were only comparable for about a year when rdna 1.0 cards and rtx 20 series cards had the same tflops and performance and only for two cards, the 2060 and 5700, and the 2070 and 5700xt. Since then nvidia inflated their tflops count but you can always compare rdna 2.0 cards on PC very easily.

Here is a forza benchmark. Notice how performance scales roughly in line with the tflops gap.

5500 xt - 5.2 tflops- 36 fps
6600 xt - 10.6 tflops - 67 fps
6300 xt - 13.2 tflops - 86 fps
6800 xt - 20.7 tflops - 117 fps
6900 xt - 23 tflops - 126 fps

The 13tf 6700xt and 5tf 5500 xt is a great comparison because they are very comparable to the xsx and xsx. 8 tflops difference between the two. the tflops differnce between them is 2.5x and the performance difference is 2.3x.

The 10.6 tflops 6600xt and 20.7 tflops have a 1.95x tflops difference and a 1.75x performance difference.

The 10.6tf 6800xt and the 13.2 tflops 6700xt have a 1.25x tflops difference and 1.28x performance difference.

Tflops and performance scale linearly when you are comparing the same family of GPUs. That’s always been true. Its how GPUs are priced. It’s how they are benchmarked. If anything the higher end cards end up performing slightly worse due to various issues from making out bandwidth or pushing CU counts too high. The xsx and xss have a 3x tflops differential and a 4x performance differential. The xss should be offering 1/3rd the performance of xsx but it only offers 1/4th and that too with graphics settings paired back.

i literally said they they should not be compared with my own examples

why are you repeating what I've said? how I not made myself clear enough? if so, there's a big misunderstanding here.

 

[Bernd Lauert]

I personally don’t think the series s is too bad for its price Yet it’s constantly picked on.

It's free real estate for console warriors. They can't say much about the XSX (and if they try, they might get banned), but the XSS is a free target, and they can always hide behind "well I'm just discussing performance in a performance thread", despite some posts clearly being console war bait.

(Not talking about Slimy here btw)

 

[Rubim]

bro, my entire point was that we should not compare tflops / architectures. my post was targeted at people who keeps using bloated Ampere numbers as a way to downplay consoles. can you please read the posts before me? there was a person that is saying there are 20+ tflops GPUs in the market and 4 tflops is funny compared to them. so... you're missing my point

"Who would have thought it would look this good on a 4 TF, yes 4 TF machine … while PCs are going to 24 TF soon .."


as i've said, its about the optmization. kepler was a beast. in games too. it all went downfall once the maxwell is released. i vividly remember this topic and the outrage that followed it.

The Witcher 3 - Kepler GPU

The Witcher 3 - Kepler GPU gtx 960=gtx 780 gtx 970>gtx 780ti what's going on here ? how is this possible ? nvidia is deliberately downgrading kepler gpu's performance in favor of maxwell using driver or poor optimization game under 700 card by cdprojektred ? I read the forums, and I'm in shock...

forums.cdprojektred.com

this was the last game that kepler got "special" driver codepath optimizations. after that, it was mostly a wreck.

i'm not saying ps4 runs the rdr 2 at high settings, but it looks gorgeous. gtx 780 clip however looks atrocious.



ps4 settings are a mixed bag. yes, there are stuff that are lower than low, but then again, there are lots of stuff that are at medium/high/ultra. like shadows, ssao, tesselation, particles, volumetric resolution, geometry lod (ultra), textures (ultra). game simply looks beautiful on a ps4. i just tried to show it as an example to prove that TFLOPS are not be-all end-all.

in the end, 4 tflops of rdna 2 on an optimized space might output the performance of a 8-10 tflops ampere gpu. who knows? why downplay 4 tflops? its clear that 1.8 tflops of old GCN can provide exceptional performance in games like rdr 2, horizon zero dawn, god of war and many more. even doom eternal runs at 1080p 60 fps on that thing. 4 tflops of an advanced, newer architecture should give enough juice to series s for its lifetime and its target resolution. at least I hope so.

Yeah i misread you, im sorry.

But i think we agree on the point that tflops is a weird metric to be talking about.

 

[rofif]

It takes a significant amount of time for it to calculate the new lighting when changing from day to night though (I mean, not LONG, but a very noticeable amount of time), it's not like it's fully recalculated every frame. Which makes me wonder how it would look if the sun was continually moving, or if you wanted dynamic changes in response to destruction etc.

of course. It is experimental after all.
if the time was real time - it would change more gradually

 

[SlimySnake]

i literally said they they should not be compared with my own examples

why are you repeating what I've said? how I not made myself clear enough? if so, there's a big misunderstanding here.

Ummm no. You are using the nvidia 30 series tflops fiasco as a barometer to dismiss ALL tflops comparisons. You can and should compare tflops within the same family of GPUs. You literally started your posts mocking tflops as a performance measuring tool even though we were using the xsx tflops to compare the xss tflops comparisons.

 

[Fafalada]

My point is that there is no delineation between a graphics engineer in gaming and a graphics engineer in film. They both have skillsets that can be used interchangeably.

That's a... Can of worms... While I always subscribed to software engineering being malleable across roles far more than most industries accept(sadly lots of folks hang on to this idea of hard skills that define roles), I would be really hesitant to call skillsets here "interchangeable".

But let me ask first what it is that you define as a "graphics engineer", as we may be talking different things.

 

[ReBurn]

.

 

[Bogroll]

They might have known earlier than we did, but they also finalised their plans before we know they did.

Yes, I do think MS did not expect the $399 PS5 and they setup a brilliant strategy in that scenario, you are quoting other reasons for it to exist (that I disagree with if you compare it to say a $349-399 512 GB of SSD digital only XSX, it would have provided more benefits to users and devs… if you really do not care much about graphics you stay one more year or so on an XOX, PS4, etc…)

It amazes me you think Xbox didn't expect a digital PS5. Do you not think Xbox considered a digital XSX and of course they will release one in the future. Gamepass has been around for 5years and this another way for Xbox to go. They know they can't go head to head with PS5.

 

[DenchDeckard]

It’s very rare for it to offer 1440p modes. And even then there are downgrades in graphics settings. The vast majority of games including the most recent high profile ones are 1/4th.

I wouldn’t be too sure about series x not showing its advantage. The DF clip i saw had the ps5 dip to 20 fps far longer than the xsx during a car crash. Xsx drops for maybe 1-2 seconds, ps5 felt like 3-4.

They need to do better tests. Preferably without the same cars. And calculate the minimum resolution at the time of the dip to 20 fps.

Yeah, fair play. I have no issue discussing with you on here.I know you’re one of the level headed peeps and you are always open to discuss. I don’t think there is any problem with questioning the series S and it will be interesting to see how far the cuts go when developing next gen games for it, but as a little box that’s perfect for a kids bedroom etc I think it offers a great foot in the door for this generation of games.

a 4tf gpu is pretty measly compared to the big boys but I have been quite impressed with what it’s delivered so far.

 

[rofif]

It’s very rare for it to offer 1440p modes. And even then there are downgrades in graphics settings. The vast majority of games including the most recent high profile ones are 1/4th.

I wouldn’t be too sure about series x not showing its advantage. The DF clip i saw had the ps5 dip to 20 fps far longer than the xsx during a car crash. Xsx drops for maybe 1-2 seconds, ps5 felt like 3-4.

They need to do better tests. Preferably without the same cars. And calculate the minimum resolution at the time of the dip to 20 fps.

I am not sure how can you measure fps drops in such a dynamic environment. So maybe good they didn't focus on it

 

[Riky]

You don't know that and do you really think it was wow moment for those in the know or even any of us when the digital PS5 announced. It wasn't like the reveal of the wheel.
I'd guess it was more of a GP machine, cheap and a alternative to any PS5 for those who are not so bothered about graphics etc.

He would probably be better off actually listening to what Jason Ronald said about why they made Series S, he explains it fully.

Die shrinks are more difficult and will not give the sort of cost savings that we have in the past, therefore the chances of the machines getting down to those mass market price ranges is very slim.
So they released a console that starts at that price range. It was no more complicated than that.

 

[ToTTenTranz]

I’d love to see a one X try and run this demo.

The One X with a standard HDD would probably be unable to run it without stuttering and lots of LOD popups.

 

[RoadHazard]

UE5 was developed and being refined by people from both industries. My point is that there is no delineation between a graphics engineer in gaming and a graphics engineer in film. They both have skillsets that can be used interchangeably. Getting more graphics engineers from film will be a positive contributing factor to the tech evolution. Back in the day, game developers would never use film software engineers for their game development for fear that they didn't know enough to be successful.

There is absolutely a big difference between real-time game rendering and offline movie CG rendering. Very different techniques are needed. With UE5 they're starting to converge, but I think that's about the only example of this so far? So I don't think it's accurate to say the skill sets needed are completely interchangeable.

 

[Panajev2001a]

It amazes me you think Xbox didn't expect a digital PS5. Do you not think Xbox considered a digital XSX and of course they will release one in the future. Gamepass has been around for 5years and this another way for Xbox to go. They know they can't go head to head with PS5.

I do not think they expected a $399 fully featured DE of PS5 at launch, no. Two models at $299 and $499 from them makes me think that they did not expect it.

Of course Riky they decided on how it would be feasible for them to offer a $299 model at launch without taking a bath in red ink (which again shows the “trillion dollar company using their checkbook to cover all expenses” was e-peen waving and not based in reality). So, sure after the fact rationalisations of how they went after that price point in a cost effective way, but that is not why they went after that price point IMHO.

 

[Riky]

I do not think they expected a $399 fully featured DE of PS5 at launch, no. Two models at $299 and $499 from them makes me think that they did not expect it.

Of course Riky they decided on how it would be feasible for them to offer a $299 model at launch without taking a bath in red ink (which again shows the “trillion dollar company using their checkbook to cover all expenses” was e-peen waving and not based in reality). So, sure after the fact rationalisations of how they went after that price point in a cost effective way, but that is not why they went after that price point IMHO.

Despite Jason Ronald going through the rationale of why before the prices were announced.
Sort of puts a hole in that theory.

 

[Bogroll]

I do not think they expected a $399 fully featured DE of PS5 at launch, no. Two models at $299 and $499 from them makes me think that they did not expect it.

Of course Riky they decided on how it would be feasible for them to offer a $299 model at launch without taking a bath in red ink (which again shows the “trillion dollar company using their checkbook to cover all expenses” was e-peen waving and not based in reality). So, sure after the fact rationalisations of how they went after that price point in a cost effective way, but that is not why they went after that price point IMHO.

You saying this because of the cheaper digital PS5 or do you think the SS should of released a third of the price of the SX as its roughly a third as powerful?

 

[VFXVeteran]

UE5 is developed by Epic, like the previous UE versions. They make games and game engines, not movies, even if now UE is starting to be used for movies and commercials.

A graphics engineer in gaming has pretty similar skillsets than one from movies, but it's pretty different because running stuff in real time in a single PC or console -and more in a pretty humble hardware- has nothing to do with a a movie scene pre-rendered in cloud farm.

If you're an vfx artist you should know that, because vfx in games and movies are pretty different because real time -specially during gameplay- is way more limited.

You are looking at render times. I'm looking at the actual working environment with knowledge of rendering.

My point is that on a technical level, the engineer knowledge base has merged. I know plenty of software engineers from film that work at Epic and other gaming companies as graphics engineers. Graphics will start looking better because of the knowledge they bring to the table on advanced rendering features/shaders.

Also not liking the tone over this so if you think I'm full of shit based on your first combative reply, then so be it.

 

[DaGwaphics]

They might have known earlier than we did, but they also finalised their plans before we know they did.

Yes, I do think MS did not expect the $399 PS5 and they setup a brilliant strategy in that scenario, you are quoting other reasons for it to exist (that I disagree with if you compare it to say a $349-399 512 GB of SSD digital only XSX, it would have provided more benefits to users and devs… if you really do not care much about graphics you stay one more year or so on an XOX, PS4, etc…)

The only issue with that theory is why would MS assume that the PS5 would be more than $399. That had been the successful price point for them. The price point they had stuck with on the Pro, which likely released somewhat close to when target specs were getting finalized on the Series systems. Sandwiching the $399 PS was probably the point from the beginning.

 

[Panajev2001a]

The only issue with that theory is why would MS assume that the PS5 would be more than $399. That had been the successful price point for them. The price point they had stuck with on the Pro, which likely released somewhat close to when target specs were getting finalized on the Series systems. Sandwiching the $399 PS was probably the point from the beginning.

Sure, but we’re not the rumors focusing on a more expensive than normal PS5 (targeting $449-499 range) because of the custom SSD, UHD Blu-Ray, etc…?

 

[VFXVeteran]

There is absolutely a big difference between real-time game rendering and offline movie CG rendering. Very different techniques are needed. With UE5 they're starting to converge, but I think that's about the only example of this so far? So I don't think it's accurate to say the skill sets needed are completely interchangeable.

I'm not saying that.

I'm saying the knowledge of a graphics engineer in film and games are starting to be interchangeable whereas it wasn't before. Engineers in gaming never had to have much knowledge about RT and how to implement true lighting techniques that are more physically plausible. Hardware is getting fast enough now that this is being adopted in games. For example, the hair shader in film is still significantly better and more complex than games. Eventually, we'll start seeing fully PBR 4-lobe specular distributions with self-shadowing/AO/and GI. Something film engineers dealt with for some time already.

 

[Panajev2001a]

Despite Jason Ronald going through the rationale of why before the prices were announced.
Sort of puts a hole in that theory.

Not really, it is not like they designed and manufactured the XSS without specifically targeting that price point sustainably. So, no.

 

[DaGwaphics]

Sure, but we’re not the rumors focusing on a more expensive than normal PS5 (targeting $449-499 range) because of the custom SSD, UHD Blu-Ray, etc…?

Those rumors were quite a bit later though. You'd think the Silicone for all three designs would have been deep into development by then.

 

[Panajev2001a]

You saying this because of the cheaper digital PS5 or do you think the SS should of released a third of the price of the SX as its roughly a third as powerful?

I am not commenting on what the price XSS should have been. My preference is for it not to be there and have a cheaper 512 GB SSD digital only XSX for reasons I mentioned 245 times , but that is not a comment on its price.

They did want to sandwich PS5 but it started strongly out of the gate and the digital model helped lower the price perception too.

 

[Panajev2001a]

Those rumors were quite a bit later though. You'd think the Silicone for all three designs would have been deep into development by then.

Perhaps, but even then, given the surprise of the dual SKU online and its pricing (and for the disc drive model being sold at cost and not at a loss not long after launch surprised people) , it is likely it was not understood even back then.

 

[DaGwaphics]

Perhaps, but even then, given the surprise of the dual SKU online and its pricing (and for the disc drive model being sold at cost and not at a loss not long after launch surprised people) , it is likely it was not understood even back then.

Could just be a case of both companies moving forward a little differently. Ryan said that AMD approached them about a weaker console sibling and they said no, so, sometimes it is AMD that offers up certain things to start with. I think AMD was the catalyst for the mid-gen refreshes too, unless I have that mixed up. I doubt that either side was surprised by what the other released. Hell, half of GAF had guessed the correct pricing across the board. The XSS skew and the digital PS5 had been basically known for months.

 

[ToTTenTranz]

I do not think they expected a $399 fully featured DE of PS5 at launch, no. Two models at $299 and $499 from them makes me think that they did not expect it.

I wonder if the PS5 DE was itself a reaction to Lockhart's rumors.
Of course Sony can potentially make a lot more money on distribution royalties from a DE than a normal PS5, but I doubt 2019 Sony was keen in selling a console at a (seemingly large) loss. Especially considering how they ended up comfortably at the top during the previous generation.

 

[Riky]

Not really, it is not like they designed and manufactured the XSS without specifically targeting that price point sustainably. So, no.

What you say means nothing to what I said. They created a console to start at a lower price point, a price point where the majority of console sales take place because they believe die shrinks are going to be more difficult, so of course they targeted a price point, just not for the reasons you think.

It's easy to see who was reacting, remove a disc drive or design a whole different console, which would take longer........

 

[SlimySnake]

I am not sure how can you measure fps drops in such a dynamic environment. So maybe good they didn't focus on it

It's no different than any open world environment tbh. There is just more of it. Alex literally picked a car from the main spawn and drove it in a straight line then crashed. Thats a good test. Even with no traffic ahead of him, the fps dropped and sat around 24 fps. Then hit 20 fps when he crashed. Same test can be done on the xbox series x with the same car. He inexplicably chose a slower car then fumbled around after not being to crash. This guys goes to extreme lengths to find advantages between consoles and PCs in cross gen games like Hitman, AC Vahalla, etc and here is the moment of truth and he does one throw away test and calls it a day.

John also mentioned how the framerate drops while you are flying. That would be an easier test since it removes having to drive through the traffic simulation.

a 4tf gpu is pretty measly compared to the big boys but I have been quite impressed with what it’s delivered so far.

I like the idea of a smaller cheaper box. If you have an era account, you can probably pull up a thread of mine from 2018 or 2019 where I said I wanted a cheaper $399 box for 1080p gaming and a more expensive $599 pro model for 4k. People revolted as they typically do on era lol, but I was never against a two console solution. I just think 4 tflops is too low. An 8 tflops $399 console that aimed for 1080p gaming wouldve been perfect.

 

[Bogroll]

I am not commenting on what the price XSS should have been. My preference is for it not to be there and have a cheaper 512 GB SSD digital only XSX for reasons I mentioned 245 times , but that is not a comment on its price.

They did want to sandwich PS5 but it started strongly out of the gate and the digital model helped lower the price perception too.

Well I can't comment on your 245 times as I don't read everything on here.

Why does it have to be compared to PS5 all the time I know its inevitable. Xbox have basically said they're doing their own thing (Much like NIntendo do their own thing)
MS have invested billions in new studios. The games are coming and some are here with Halo and FH5 If you want a new console to play Xbox games and multiplats plus have the option to have a great service GP, you can pay £450 or £250.
If most people want a PS5 to play PS5 exclusives and multiplats they're going to get one.
If the SS was £200 at launch and there wasn't a PS5DE all the people who want a PS5 wouldn't suddenly go I'm going to get a SS cause its cheaper. They would get a PS5 regardless, so I don't know why the $299 $499 tells you Xbox didn't expect a PS5 DE.

 

[adamsapple]

will this tflops talk never end? really??

30 tflops rtx 3080 is barely pushing %40-50 more fps than series x in halo infinite/forza horizon 5. its a useless and bloated metric that nvidia boosted to get a hold in "muh tflops wars". it cannot even realistically push actual 2x performance over series x, despite having "2.5x" tflops.

While the Tflop discussion DOES get tiresome, in this example I think we can put it down more to the lack of optimization.

Halo:Infinite just doesn't perform as well on PC as it could. The game is oddly demanding on system resources and there's a pretty sizable difference in how the game works on AMD cards vs similar or closest Nvidia cards.

 

[RoadHazard]

It's no different than any open world environment tbh. There is just more of it. Alex literally picked a car from the main spawn and drove it in a straight line then crashed. Thats a good test. Even with no traffic ahead of him, the fps dropped and sat around 24 fps. Then hit 20 fps when he crashed. Same test can be done on the xbox series x with the same car. He inexplicably chose a slower car then fumbled around after not being to crash. This guys goes to extreme lengths to find advantages between consoles and PCs in cross gen games like Hitman, AC Vahalla, etc and here is the moment of truth and he does one throw away test and calls it a day.

John also mentioned how the framerate drops while you are flying. That would be an easier test since it removes having to drive through the traffic simulation.


I like the idea of a smaller cheaper box. If you have an era account, you can probably pull up a thread of mine from 2018 or 2019 where I said I wanted a cheaper $399 box for 1080p gaming and a more expensive $599 pro model for 4k. People revolted as they typically do on era lol, but I was never against a two console solution. I just think 4 tflops is too low. An 8 tflops $399 console that aimed for 1080p gaming wouldve been perfect.

There already is a 10.3TF $399 console though...

 

[Kataploom]

It's no different than any open world environment tbh. There is just more of it. Alex literally picked a car from the main spawn and drove it in a straight line then crashed. Thats a good test. Even with no traffic ahead of him, the fps dropped and sat around 24 fps. Then hit 20 fps when he crashed. Same test can be done on the xbox series x with the same car. He inexplicably chose a slower car then fumbled around after not being to crash. This guys goes to extreme lengths to find advantages between consoles and PCs in cross gen games like Hitman, AC Vahalla, etc and here is the moment of truth and he does one throw away test and calls it a day.

John also mentioned how the framerate drops while you are flying. That would be an easier test since it removes having to drive through the traffic simulation.


I like the idea of a smaller cheaper box. If you have an era account, you can probably pull up a thread of mine from 2018 or 2019 where I said I wanted a cheaper $399 box for 1080p gaming and a more expensive $599 pro model for 4k. People revolted as they typically do on era lol, but I was never against a two console solution. I just think 4 tflops is too low. An 8 tflops $399 console that aimed for 1080p gaming wouldve been perfect.

The hardware inside XSS is way too good for the price tho, the only weak part is the GPU and that seems to be strong enough for gaming in lower resolutions and lower settings, they couldn't make it more powerful because they also have to sell the X and both systems have to be different enough to not be confused or cannibalize each other sales.

I think the balance between both systems is very good, see how good this demo looks on the S, graphics analyst can say whatever about lower settings, but in my eyes it was only lower resolution and I have very good eyes, let's say I didn't care about what was lacking since at first glance it was good enough for me, I guess most people getting the S will feel the same.

As long as base game keeps the same and graphics keep good enough relative to bigger consoles, everyone can be happy with their purchase.

 

[SlimySnake]

While the Tflop discussion DOES get tiresome, in this example I think we can put it down more to the lack of optimization.

Halo:Infinite just doesn't perform as well on PC as it could. The game is oddly demanding on system resources and there's a pretty sizable difference in how the game works on AMD cards vs similar or closest Nvidia cards.

I was just about to create a thread on that actually. My friend has a 580 which is basically the same 6 tflops cards thats in the X1X give or take 4 CUs. It cannot run Halo at 60 fps. he tops out at 30 fps in the open world. I refused to believe him and told him to enable DRS but it didnt help him. Hes playing on 1080p so maybe DRS doesnt work below 1080p? To me, if you set your minimum FPS at 60 fps, it should always stay there, no? If the game needs to drop to 720 or even 640p, it should drop the res and maintain 60 fps, but its not doing that for him. I then googled and found several youtube videos where it tops out at 35 fps, reddit posts complaining about the game's performance on the 580 and its all true.

It simply shouldnt happen. The jaguar CPUs are awful compared to PC CPUs. PCs have more system ram. 580 is an AMD card. There is no reason why the X1X can run it at 60 fps and a similar PC GPU cant. HAS to be optimization.