Digital Foundry - Hellblade Switch Analysis: Just How Close Is It To PS4?

~Fake

Member
Dec 20, 2018
1,157
996
260
Rio de Janeiro

Even we had to do a double-take with this one! Just how did developer GLOC manage to make Hellblade on Switch look so close to the PS4 version of the game? @dark10x and Rich - with an assist from Tom - get to grips with the port and how it manages to deliver its sometimes remarkable visuals.

Summary:
- PS4 vs Nintendo Switch
- While dock, the game use dynamic resolution at 720p (1024x576~1280x720)
- While portable, the game use dynamic resolution at 720p, but more agressive (711x400~1280x720)
- Most of the cutscenes are full motion video
- The video used in cutscenes run at a higher resolution than the game inself, but still based on NS
- Texture are in lower resolution, texture filtering looks more agressive
- LOD reduced
- Lightning very different in comparison with others version. They maybe find another solution for global illumination on NS
- Even after those compromises, still works and quite impressive
- Perfomance time
- The game runs at 30 fps
- Both dock and portable are not great
- Some dips, 20~30 fps
- Frame pacing issue
- Some audio bugs (14:06)
 
Last edited:

DynamiteCop!

Member
Mar 3, 2018
1,153
1,894
270
And then reality

  • Dynamic resolution
  • Maximum resolution of 720p, bottom end 576p when docked, 400p in handheld. Docked tends to generally run lower than 720p.
  • Uses TAA to clean up jagged edges but further blurs the already extremely low resolution
  • 30 FPS which can drop into the low 20's with horrible frame pacing (think FromSoftware bad)
  • Low resolution textures
  • Reduced foliage
  • Reduced environmental assets
  • Reduced environmental geometry
  • Reduced reflections
  • Reduced shadows
  • Removed ambient occlusion
  • Possibly removed global illumination model present on every other platform (evidence of this)
  • Reduced LoD on distant objects and draw in
  • FMV's in place of real time cutscenes
  • Possible audio glitches causing cutting in and out
 

stranno

Member
Dec 7, 2016
686
495
285
León, Spain
The good old real-time to pre-rendered cutscenes trick, that never worked. At least this time they work with some bitrate, decent video format and minimal artifacts. Nothing like the Wii version of The Amazing Spider-Man (with amazing macro-blocks) or Mafia on Playstation 2.
 
  • Like
Reactions: mckmas8808

ethomaz

Member
Mar 19, 2013
22,613
1,545
520
36
Brazil
Where is the close?

It is a way bigger difference than a PS4 to XB1 difference 🤔
 
Last edited:

Afro Republican

GAF>INTERNET>GAF, BITCHES
Aug 24, 2016
3,205
1,544
615
So it can't even touch base Xbox One on Switch let alone PS4.

I predict significantly less console ports going to Switch next year.
 
  • Like
Reactions: mckmas8808

TLZ

Member
Oct 20, 2015
5,210
1,925
385
Sydney
They should call all these ports Demasters.

Playing them handheld is cool though.
 

Armorian

Member
Jan 17, 2018
528
341
210
So it can't even touch base Xbox One on Switch let alone PS4.

I predict significantly less console ports going to Switch next year.
Switch is 0.384 TF in docked mode and Xbox One S is ~1.3TF so over 3 times more powerful, not to mention CPU, RAM size and BW differences.

I dont know why people even expect Switch ports to be close to console versions, many of them run at absymal resolutions/framerates (or both).
 

DynamiteCop!

Member
Mar 3, 2018
1,153
1,894
270
Switch is 0.384 TF in docked mode and Xbox One S is ~1.3TF so over 3 times more powerful, not to mention CPU, RAM size and BW differences.

I dont know why people even expect Switch ports to be close to console versions, many of them run at absymal resolutions/framerates (or both).
Probably because it exists in the same market, it's a console and it costs more money than both others which dust it..
 
Last edited:
  • Like
Reactions: scydrex

ethomaz

Member
Mar 19, 2013
22,613
1,545
520
36
Brazil
Switch is 0.384 TF in docked mode and Xbox One S is ~1.3TF so over 3 times more powerful, not to mention CPU, RAM size and BW differences.

I dont know why people even expect Switch ports to be close to console versions, many of them run at absymal resolutions/framerates (or both).
Because the title falsely led you to believe that...
 
  • Like
Reactions: mckmas8808

EHuntingon

Member
Jul 30, 2013
59
58
310
I’ll probably pick this up over the weekend. I’m sure it’d look much better on my PS4 Pro, but I really want to support Switch ports.
 

MrTickles

Gold Member
Feb 22, 2018
2,105
2,636
440
Switch is 0.384 TF in docked mode and Xbox One S is ~1.3TF so over 3 times more powerful, not to mention CPU, RAM size and BW differences.

I dont know why people even expect Switch ports to be close to console versions, many of them run at absymal resolutions/framerates (or both).
Switch is 0.6tf in docked mode, <0.3 in portable.
 

MrTickles

Gold Member
Feb 22, 2018
2,105
2,636
440
Last edited:

Omeggos

Member
Jan 12, 2018
2,858
2,137
270
Probably because it exists in the same market, it's a console and it costs more money than both others which dust it..
Except that it doesn’t? The ps4 still currently stands at a msrp of $299 which is equal to the switch. The switch may be weaker but it also has the gimmick of being a handheld fighting in the home console space which is its biggest selling point (alongside heavy hitter exclusives like Pokémon, smash and Mario kart.)

Plus power has been proven before to be part of the equation, not the factor itself.
 

Armorian

Member
Jan 17, 2018
528
341
210
1TF in FP16 (and that's with higher than Switch clock), by that metric Pro is 8.4TF :p

TX1's performance is well-known, so with those clock speeds, that puts Switch's compute power at 393-471 / 157-196 GFLOPs base/boost docked/undocked.

For comparison, the Wii U benchmarked close to 172 GFLOPs, but with specific custom features that helped it do a few things a little beyond that specification. Xbox One pulls 1300 GFLOPs and PlayStation 4 pulls 1800 GFLOPs.
Code:
https://www.reddit.com/r/NintendoSwitch/comments/6bwmos/nintendo_switch_soc_compute_power/
 
Last edited:

DynamiteCop!

Member
Mar 3, 2018
1,153
1,894
270
Except that it doesn’t? The ps4 still currently stands at a msrp of $299 which is equal to the switch. The switch may be weaker but it also has the gimmick of being a handheld fighting in the home console space which is its biggest selling point (alongside heavy hitter exclusives like Pokémon, smash and Mario kart.)

Plus power has been proven before to be part of the equation, not the factor itself.
Have you seen how a base xboner runs AAA games these days? 720p is common.

The WiiU was 355 gigaflops. The Switch is considerably faster than that console. Might want to tell Gamespot they have their figures wrong: https://www.gamespot.com/gallery/console-gpu-power-compared-ranking-systems-by-flop/2900-1334/10/

I think we both went way off track. The Switch is up to 1 TF docked, 0.6tf undocked. Would explain why even in portable mode it can manage ue4 games at 540p.

No.

 

MrTickles

Gold Member
Feb 22, 2018
2,105
2,636
440
'Specific custom features that help it do things beyond that specification' makes the whole teraderp bullshit meaningless then. The numbers by themselves indicate less than they would seem to at first glance because visuals and performance don't directly scale with them. On top of that the Tegra architecture differs completely to the AMD architectures in current and in previous consoles.
 
Last edited:

SonGoku

Member
Aug 16, 2018
1,309
1,006
220
Have you seen how a base xboner runs AAA games these days? 720p is common.

The WiiU was 355 gigaflops. The Switch is considerably faster than that console. Might want to tell Gamespot they have their figures wrong: https://www.gamespot.com/gallery/console-gpu-power-compared-ranking-systems-by-flop/2900-1334/10/

I think we both went way off track. The Switch is up to 1 TF docked, 0.6tf undocked. Would explain why even in portable mode it can manage ue4 games at 540p.

No, the wiiu was 160 gigglypuffs
GS is a terrible source to use, switch docked is in the 300s gflops
 
Last edited:

Omeggos

Member
Jan 12, 2018
2,858
2,137
270
Because it doesn't remotely come close to 1TF, that's a stock Tegra X1 running FP16, this is underclocked and peaks at 393.2 Gigaflops docked...

That's how it's relevant.
I wasn’t arguing specs. I was arguing price point. I’m well aware of the switch’s weaker system but on top of the fact it’s a nvidia build (which console history shows they’re fairly expensive in general) the system needs to include joycons and the tv adapter, built in screen and battery, etc) the joycons and dock alone cost around $70-80 each which also adds to the price.

Hence my point, there’s more to it’s pricing than just its hardware speed.
 
Last edited:

MrTickles

Gold Member
Feb 22, 2018
2,105
2,636
440
He right tho. Homebrew allows you to overclock Switch to 1.4ghz. A constant, stable 1.2ghz with stock cooler in docked mode is easy. BAM 1TF. Performance improvements in ue4 games abound, we talking no more framedrops in games like doom.


So we both right.
 
Last edited:

DynamiteCop!

Member
Mar 3, 2018
1,153
1,894
270
He right tho. Homebrew allows you to overclock Switch to 1.4ghz. A constant, stable 1.2ghz with stock cooler in docked mode is easy. BAM 1TF. Performance improvements in ue4 games abound, we talking no more framedrops in games like doom.


So we both right.
No we're not both right because not only can a now gross minority of Switch's do this, probably 0.0001% of people with them actually will or want to try this. That's not the Switch, that's not how it runs, you were wrong and that's the end all be all end of it.

157.2 Gigaflops handheld, 393.2 Gigaflops docked
 

ethomaz

Member
Mar 19, 2013
22,613
1,545
520
36
Brazil
i.e. 99.9999999% of Switch's in circulation lol, don't give me this back handed jailbroken comment like it means anything.
Wut?

Switch boost clock in some games going up to 921Mhz docked.

That happens with 100% of Switch in the market.
 
Last edited:

Kagey K

Member
Dec 18, 2013
2,128
1,711
430
Who cares how it is to the PS4? How is it to Xbox?

I’m guessing better then then the stock Xbox but much, much worse then it is to the X.

XB1S might help make the Switch look better for these comparisons.

Clickbait.
 

DynamiteCop!

Member
Mar 3, 2018
1,153
1,894
270
Who cares how it is to the PS4? How is it to Xbox?

I’m guessing better then then the stock Xbox but much, much worse then it is to the X.

XB1S might help make the Switch look better for these comparisons.

Clickbait.
The Xbox One version is near identical to the PS4 version with the exception of resolution...

 

JaffeLion

Member
Nov 28, 2016
814
245
210
why ps4? xbox one X is way better than any other version. they should compare it to xbox one x
 

Kagey K

Member
Dec 18, 2013
2,128
1,711
430
The Xbox One version is near identical to the PS4 version with the exception of resolution...

Which is why I said it would be a better comparison for the switch.


But you go ahead and keep getting ready for that console war.
 

DynamiteCop!

Member
Mar 3, 2018
1,153
1,894
270
Which is why I said it would be a better comparison for the switch.


But you go ahead and keep getting ready for that console war.
It wouldn't be a better comparison because the result would still be the same in terms of graphics and lack of cutbacks, the PS4 and Xbox One version are graphically in parity.

Inform yourself.
 

Kagey K

Member
Dec 18, 2013
2,128
1,711
430
It wouldn't be a better comparison because the result would still be the same in terms of graphics and lack of cutbacks, the PS4 and Xbox One version are graphically in parity.

Inform yourself.
Receipts?

I can only find pro vs X comparisons.
 

llien

Gold Member
Feb 1, 2017
4,416
1,702
475
Is nVidia reality distortion field still a thing?

whatdiyouexpect.gif
 

Virex

Member
Jan 26, 2018
1,140
1,858
320
South Africa
Including a shitty game like Hellblade. Right?
Hellblade seemed to review well on PS4. So it's probably not a shitty game. But the Switch version looks ridiculously inferior. And you'll hear the justification like "but...but..I can play it on the go." That seems to be the excuse for shitty ports on Switch in general. I have a Switch but don't have many games. I only got it for Zelda, Xenoblade, Mario and Octopath. But my Switch has never left the dock though
 

Kagey K

Member
Dec 18, 2013
2,128
1,711
430
Hellblade seemed to review well on PS4. So it's probably not a shitty game. But the Switch version looks ridiculously inferior. And you'll hear the justification like "but...but..I can play it on the go." That seems to be the excuse for shitty ports on Switch in general. I have a Switch but don't have many games. I only got it for Zelda, Xenoblade, Mario and Octopath. But my Switch has never left the dock though
It was a superior game as by all the game of the year noms and whatever. The second it was announced for Xbox all the sudden people hated it,and now that’s is for the Switch it may as well not have happened.
 
Jul 29, 2013
803
667
470
The Switch is basically a 360/PS3 in portable mode, and half a Xbox One in docked mode. It's great they're bringing all these games over.