Alcifer Comments

Page 1 of 20

  • Oh my god Secret of Mana is being 3D remastered for PS4, PC and Vita

  • Alcifer 25/08/2017

    @Bertie pre-orders are up on PSN:
    Ł32.99 (ouch) with a release date of 15/02/18.
    $39.99 on US, releasing 31/03/18.

    PS4 only (no Vita option or cross-buy :( ).
    Update: Vita is available on the Japanese store, probably a region exclusive (same as Setsuna).
    Reply +3
  • Alcifer 25/08/2017

    @Kremlik the first Seiken Densetsu game has been remade twice, first as Sword of Mana on the GBA and then as Adventures of Mana on the Vita/Android/iPhone. It'll be interesting to see if it gets a PS4/Steam release too. Reply 0
  • Alcifer 25/08/2017

    The previous 3D Mana remake, Adventures of Mana, was exclusive to Vita and mobile phones, as was the similarly styled Rise of Mana.

    At the time they hinted at future remakes of SoM and SD3 (OMG Yes!) based on the same tech if that was well received.

    It might be worth mentioning that the Switch got a collection of the original GameBoy/SNES games (SD1-3) in Japan earlier this year.
    Reply +4
  • Horizon Zero Dawn: the making of PS4 Pro's best 4K game

  • Alcifer 03/08/2017

    What's a depth prime?
    Probably the same as a Z pre-pass, an optimisation to reduce overdraw.

    You render a depth only pass of all the opaque geometry in the scene after which the depth buffer contains values for all the visible pixels that frame (the depth buffer is primed for later passes).

    When you come to render the geometry pass, which is far more resource intensive, the GPU will only shade the pixels which will be visible and avoid shading those which would be overwritten later in the pass.
    Reply +1
  • Input lag tested: COD vs Battlefield - which is the most responsive FPS?

  • Alcifer 29/07/2017

    But, it also means that the input lag perceived by the player will actually be higher, since they also need to factor in display lag.
    Indeed, each player's display will add it's own delay based on the exact make, model and settings used. These could vary quite widely between individual setups though.

    For measuring the total latency, including that of the display, the older LED + camera method is more appropriate. It also means only the controller needs to be modified, so has the added benefit of not needing to figure out how to dirty the signal to the Switch's internal display.

    In this case a quick and dirty test would just be to use two Switch consoles, one docked and the other portable next to the TV, and a high speed camera to capture (near) simultaneous inputs to both. If the latency of the TV is known then you could determine a rough value for the internal screen based on any difference.
    Reply +2
  • Alcifer 29/07/2017

    @Malek86, these results shouldn't include display latency because both the console output and the test signal go through the display (unlike the LED + camera method).

    The new test rig operates between the controller and the output signal of the console, the measured time is the difference between the dirty output appearing on screen and the triggered action appearing. Both events will go through the same display processing and so should appear after the same delay, removing display lag from the equation. It's a very elegant system.
    Reply +14
  • Why PC is the best place to play Mass Effect Andromeda

  • Alcifer 01/04/2017

    For those looking to learn a little more about the Frostbite team's approach to checkerboard rendering, there's a complete GDC presentation you can check out here
    Interesting link. I noticed this one on Framegraphs.
    A graph of all rendering operations for the entire frame is a useful abstraction. The industry can move away from “immediate mode” DX11 style APIs to a higher level system that allows simpler code and efficient GPU utilization.
    I guess The Shoot and Buzz Junior Games (PS3) were ahead of their time ;)
    Reply 0
  • What happened to gaming's Waterworld?

  • Alcifer 31/08/2016

    @dadrester yeah, I'm still writing, not finishing much of it, but still writing ;)

    How are things with you?
    Reply +1
  • Alcifer 29/08/2016

    @aniki I saw a presentation in the Abertay Union building about Project MyWorld, it must have been around 2005/6.

    I remember walking out afterward and trying to figure out if we'd just seen the future of gaming or the most insane project ever. Sadly, even then I think the consensus was that it was an incredibly ambitious undertaking with no real idea of what it was for.
    Reply +3
  • Dangerous Golf review

  • Alcifer 05/06/2016

    While the Burnout crash mode influence is strong, having watched a few videos it actually reminds me a lot of The Shoot (PS move title).

    The material physics in both titles look a little fake, things like wood and stone lack weight and tend to explode. In the case of The Shoot the environments were meant to be built of stuff like polystyrene but the integrated Havok destruction was capable of much better.

    While it increases the potential destruction it actually makes it far less satisfying to watch, Burnout Paradise by comparison really nailed the feeling or weight and momentum when cars smashed and buckled.

    It's a case of less being more.
    Reply 0
  • That time I was blacklisted by Sega while editing a Sega magazine

  • Alcifer 21/05/2016

    I still remember when I started uni and the day the first student loan cheques came in. The first thing a couple of my flatmates did was cash theirs in at some dodgy place and pick up a Dreamcast and TV.

    Of course it didn't take long to realise that eating (and drinking!) was slightly more important than next-gen gaming. The TV went back to the shop and someone picked up a slightly used DC for a really good price ;)

    The DC was actually surprisingly popular among students. I think the thriving pirate scene had a lot to do with that, I can remember 3-4 shops within walking distance that specialised in boot CDs and pirated games.
    Reply +1
  • Digital Foundry's top ten picks for PlayStation Neo upgrades

  • Alcifer 30/04/2016

    @TallestJon96 the PS3's Cell processor was actually pretty scalable. They could have produced a version with additional SPUs and a larger version of the RSX, it just wasn't commercially sound to do so.

    The PS4 and XB1 aren't exactly standard PC hardware. There are no PC APUs (or discrete CPU/GPU combos) that use GDDR5 system RAM or eSRAM to boost bandwidth.

    This gen consoles are using low power x86 CPUs, in previous generations PowerPC was king (as it was in Macs) and before that Mips. The hardware has always used existing architectures but never really off the shelf (even Cell was designed for everything from TV's to super computers).
    Reply +3
  • Alcifer 30/04/2016

    @TallestJon96 what difference does x86 architecture make?
    Nintendo did the same thing with the ARM chip in the 3DS/N3DS and with the PowerPC 750 in the GC/Wii/Wii U. Same architecture just bigger and/or faster.
    Reply 0
  • Valve releases Portal-themed SteamVR Performance Test

  • Alcifer 23/02/2016

    Nearly 10 years old but my CPUs are VR ready! :D

    Would just need to replace the somewhat newer 270X...
    Reply 0
  • Rare and the rise and fall of Kinect

  • Alcifer 16/12/2015

    I think Microsoft's big mistake was shipping XBox One with a peripheral that added around 50% to the cost and telling everyone it was integral to the platform but not designing a controller to compliment it.

    Why build a platform around gestures and movement then give someone a traditional pad that actively discourages that? If they were serious they should have gone all in, like Nintendo did, and shipped the console with a dual nunchuk set up with the traditional pad as an extra.
    Reply +3
  • Digital Foundry vs Forza Motorsport 6

  • Alcifer 17/09/2015

    @Milouse refraction from screenspace water droplets and heathaze isn't new, it's surprisingly simple to implement. They do look nice though.

    There are plenty of games on PS3/360 that use it.
    Reply +8
  • Ashes of the Singularity: the first DX12 gaming benchmark tested

  • Alcifer 18/08/2015

    @mdeneuve, the cost of state management might be to do with the GPU being a pipeline with multiple stages and always trying to maintain maximum occupancy of every stage. As one batch filters through you want another to start with a minimal gap in between so the drivers have to work out exactly when and how to change state to achieve that.

    Low level APIs on consoles, like GCM on PS3, allow developers to manually build and maintain command lists but also exposed a lot of the quirks of the hardware that are key to performance. With the high level APIs these are handled by the driver invisibly but not necessarily efficiently, changing states in the wrong order or using specific calls could hammer performance.

    With DX12 and Vulkan on PC there seems to be a sort of middle ground where we still have a certain amount of hardware abstraction for many different architectures, so programmers won't necessarily have access to all the little hardware features that aren't standard. That might be what's going on with this benchmark and the Nvidia cards, they're issuing state calls in an order or frequency that the architecture isn't designed to expect (perhaps to do with resource binding where GCN has an advantage).

    DX12 and Vulkan offer some really interesting opportunities for graphics programming, not necessarily just on PC. I don't think they'll replace high level APIs like DX11 and OpenGL though, more likely co-existing with them as an alternative.
    Reply +2
  • Alcifer 17/08/2015

    @mdeneuve, the main reason for batching draw calls is that changes to GPU state have traditionally had quite a high overhead on the driver side. The GPU runs linearly through a command buffer effectively configuring itself for every draw call before executing it, this buffer requiresvarious checks to ensure it doesn't break things. So the ideal rendering method was to render thousands of triangles at a time with the same state.

    Low level rendering APIs potentially reduce the overhead of state management, allowing the developer to change state more often with less cost and so have many smaller batches. In theory that lets a game use many more materials or lights per object.

    Multi-threaded support for command buffer creation was already available in DX11 but was poorly implemented, the main thread needed to perform all the additional checking once the worker threads submitted their command buffers. On DX12 this will become the responsibility of the developer who has to manually handle synchronisation and ensure command buffers are dealt with in order (most GPUs will still only have a single thread of command execution).

    As this benchmark seems to indicate though, it potentially moves the state change bottleneck from the driver to the hardware. Different hardware will have different costs associated with setting up draw calls. Something the DX11 drivers could account for during the command buffer translation/validation phase.
    Reply +11
  • Less than two months from launch, Tony Hawk's Pro Skater 5 changes art style

  • Alcifer 06/08/2015

    That must have taken a graphics programmer all of 30 minutes to implement. Reply +9
  • The re-engineering of Ethan Carter

  • Alcifer 22/07/2015

    @frunk the linked Gamasutra article puts that sentence into a little more context.

    In UE4, lightmaps are meant to contain indirect (bounced) lighting only; setting light source mobility to "static" causes lightmaps affected by it to also include direct lighting, but it is an exceptional situation, not the default. Such lights are not even considered for dynamic lighting. This means that they don't have specular highlights. None at all. You can probably imagine how flat and uninteresting the game looked because of that.
    In UE3 the engine obviously still considered baked light sources as dynamic and calculated the specular term in real time (based on light and viewer position) before adding it to the baked direct and indirect diffuse terms.

    My understanding is that to fix this he created a separate forward lighting pass which calculated the specular highlight and added it over the top as an emissive layer (light given off by the material itself). Effectively recreating the model used in UE3.

    This addition of the specular term would break the energy conservation principle at the heart of PBR (light reflected from a material should never exceed light received).
    Reply +5
  • Sony sets up North West Studio to make Project Morpheus games

  • Alcifer 19/05/2015

    I remember submitting a pitch document for a concept before Move was released, that would work pretty well with Morpheus. The two devices seem to compliment one another really well. Oh well... Reply 0
  • PlayStation 4 lifetime sales pass 22.3m worldwide

  • Alcifer 30/04/2015

    Does this make Microsoft Sauron? And Phil Spencer the Witch King?
    Reply +39
  • Microsoft demos Windows 10 hologram support

  • Alcifer 30/04/2015

    That's one way to make monitor/TV/Phone manufacturers slightly nervous.

    Given a sufficiently high resolution and unobtrusive AR display there really is no need for the plethora of screens we currently use.
    Reply +13
  • Digital Foundry: Hands-On with Project Cars

  • Alcifer 21/04/2015

    @Dantonir Goodbye. Reply +2
  • A Command & Conquer: Generals mod these many years later?

  • Alcifer 26/03/2015

    Generals was easily my favourite C&C title. So many hours put into multiplayer and co-op against the AI :D

    The Humvee of doom (3 rocket launcher troops and 2 snipers in a Humvee, all fully upgraded) was stupidly powerful if you micromanaged it.
    Reply +9
  • PlayStation Plus' free November offerings revealed

  • Alcifer 29/10/2014

    @dadrester, I think I've been looking forward to playing it for about 4-5 years :) Reply +2
  • Alcifer 29/10/2014

    @dadrester :D Reply +4
  • Aliens versus Predator retrospective

  • Alcifer 05/10/2014

    But to do so would be dishonest to the 12-year-old me who first played AVP in 1999
    Now I feel old.

    I remember playing this at uni on the first PC I bought with my own money. The marine campaign was terrifying at 4am (further heightened by a couple of litres of Asda own brand Red Bull substitute).

    Sadly what AvP2 gained in polish it lost in atmosphere and character. I'm pretty sure I finished it but it never had the same feeling of isolation and threat.
    Reply +14
  • Digital Foundry: Hands-on with DriveClub

  • Alcifer 06/09/2014

    I'm amazed he made it through a whole interview on real time lighting without once saying "It's mathematically correct." :D

    Looks really stunning, it really shows how much a good lighting model can bring in terms of atmosphere. Still, it would have been interesting to know if they could have hit 60fps at 720p or using Guerrilla's interlaced 960x1080 method.
    Reply +2
  • First footage of From Software's Project Beast

  • Alcifer 30/05/2014

    @marcofdeath except Hynix weren't producing HBM in 2013, they're only supposed to start ramping up production in the second half of this year, so it doesn't look like either console is using it.

    Besides, Chipworks have done a teardownn of the XB1.
    Apparently it's using H5TQ4G63AFR DDR3 SDRAM.

    In that light, I'm really looking forward to Nintendo's next console :D
    Reply +3
  • Hardware Test: PlayStation 4

  • Alcifer 13/11/2013

    Looks like the PS4 has AMD's TrueAudio tech, the same as the new Radeon cards. Reply +6
  • Digital Foundry vs. Dead Rising 3

  • Alcifer 12/11/2013

    holy bedjezus... that looks like a boring game... where's the tension?
    In the dramatic pauses between frames ;)
    Reply +8
  • Eurogamer's guide to system swansongs

  • Alcifer 10/11/2013

    @dogmanstaruk, you usually see the system's preowned section in Game getting clogged up with 3-4 iterations of Fifa as the system dies. Like some terminal degenerative disease. Not really what I'd call a swansong. Reply +27
  • Face-Off Preview: Battlefield 4 next-gen vs. PC

  • Alcifer 30/10/2013

    COD confirmed to be 720p XB1 and 1080p on PS4.

    There may be a pattern appearing here...
    Reply +4
  • Alcifer 29/10/2013

    @Suarez07, there's a tweet about the XBone AO on the NeoGAF thread.

    You can actually see it in the last comparison shot quite well in the DF article. Look at the guy on the left's gun and you'll see it shadow the elevator doors on the PS4 and PC images. Also this comparison shows it pretty well here.
    Reply 0
  • Alcifer 29/10/2013

    @Suarez07, DICE have already commented that the XB1 version is not final anyway, it doesn't have AO enabled.

    For a proper comparison of the game I'd wait for DF to compare final builds anyway. Enabling AO, and possibly AA, will most likely reduce the framerate further. With any luck DF will get better captures to work with when they don't have limited access, I interpreted the article as them capturing each platform on separate days.

    As it stands the comparison confirms what most people have been saying for a while, that the PS4 is more powerful and the XB1 versions of games are lagging behind in development.
    Reply +4
  • Editor's blog: Battlefield 4 Face-off Preview Q&A

  • Alcifer 29/10/2013

    The big question is - if ambient occlusion affects light as it hits a surface based on the surface itself and adds a softening effect similar to ambient light, does this combined with the limited RGB of the XB1, mean that when AO is added to the X1 version, we'll see similar softening as is appearing in the PS4 version?
    While it is true that AO tends to give a smooth shading effect it should not impact the clarity of texture detail in the final image, just darken areas of the image.

    It's more likely that the PS4 version is applying an additional shader based AA or motion blur pass that blurs the image. In the case of character faces in particular this could be part of the sub-surface scattering applied to skin.

    Fixing the RGB output of the XB1 will certainly reduce the contrast in the images, the DF captures make many textures appear "sharper" than those in the PC exemplar.
    Reply +4
  • Watch Epic explain Unreal Engine 4's fancy pants visual effects

  • Alcifer 24/10/2013

    @dogmanstaruk, it's pretty ironic considering all the "PS4 can't do 1080p 60" comments after the initial XBox reveal.

    I'm beginning to suspect they designed a system that could render 360 games at 1080p and didn't expect developers to do anything beyond that.
    Reply +1
  • Introducing Games of the Generation

  • Alcifer 21/10/2013

    I think the current gen still has a year or two in it, at least on some platforms. Will certainly be keeping an eye on these lists to see if there is anything really awesome that I've missed though.

    Not sure what my top 5 would be but I think Nier, Uncharted and Assassin's Creed 2 are definitely up there.
    Reply 0
  • Kingdom Hearts 3 PS4, Xbox One gameplay video

  • Alcifer 15/10/2013

    @Rogueywon, the cutscenes for Re:Coded and 356/2 Days are included in the HD remakes on PS3, so there's at least an option to avoid playing those two.

    No idea what they're planning to do with 3D regarding the HD remakes. Maybe a digital Final Mix closer to KH3 release...
    Reply 0
  • Jedi Knight 2: Jedi Outcast retrospective

  • Alcifer 13/10/2013

    I put a ridiculous amount of time into this game, and Jedi Academy, mostly just reloading checkpoints to replay fights over and over again, it never got dull. I think only Bushido Blade came anywhere near these games for visceral melee combat.

    That Lucas Arts could make such a mess of the Force Unleashed when these games existed is probably proof enough that they had no idea what they were doing. For all its gritty and dark atmosphere, watching Starkiller try to beat a storm trooper to death with his Fisher Price Padawan's first lightsabre was just sad.
    Reply +5
  • Assassin's Creed 4 director wants to take the series to ancient Egypt

  • Alcifer 12/10/2013

    I'm no historian, but could someone enlighten me on this "recent evidence"?
    I'm no historian either, but I think he might be referring to the discovery of towns built to house the workers who built the pyramids. These apparently show that the pyramids were built by gangs of seasonal labourers, rather than slaves as previously thought.

    I don't think this is evidence that slaves weren't used in ancient Egypt, just that the pyramids weren't built by armies of slaves but instead large numbers of normal Egyptians.
    Reply 0
  • Drakengard 3 will be a digital-only title on PS3 next year

  • Alcifer 10/10/2013

    Looks like it will be available at retail in the US. Will be importing this one then.
    Reply 0
  • Digital Foundry: the complete Xbox One architects interview

  • Alcifer 08/10/2013

    @khanthony, the main issue is that newer techniques may not be picked up by third parties who will want their games to be playable on all major platforms.

    In the previous generation full deferred shading was largely limited to Sony first party titles because it was not suitable on the 360 which would not support a G-Buffer larger than 10MB.
    Reply +2
  • Alcifer 08/10/2013

    @Kretenn, there is a difference between making a subjective comparison of Forza 5's graphics to other next gen titles and looking critically at the techniques they're using. Aside from an increase in resolution and quality, they don't appear to be doing a great deal that they didn't already have implemented in some form on Forza 4. The results look good, but they don't say anything about the graphics capabilities of the XB1.

    I would take Goossen's figures regarding ROPs with a pinch of salt. There is only a direct relationship between bandwidth and fillrate if none of the pixels are being rejected (depth/alpha testing). 16 ROPs will be suitable for scenes with little overdraw, like those in Forza 5 or Ryse, which to the best of my knowledge, is running at 900p 30fps. The fact that the game with the highest rate of overdraw, Dead Rising 3, is using dynamic resolution suggests that they are already having problems with fillrate or shading.

    Considering for the latter part of the current generation PC graphics have been largely crippled by the capabilities of both consoles, it would be a shame if PC and PS4 are held back by the limitations of the XB1 due to its GPU and RAM set up.

    Even the architects sound like they were disappointed with some of the compromises they had to make, rejecting GDDR5 and then a potentially much larger eDRAM pool. They certainly made the best system they could within the constraints they had, but they were pretty unlucky in what they had available to work with.
    Reply +1
  • Alcifer 07/10/2013

    @Ninja_skill, the cloudlight examples are really interesting and the latency figures might suggest the calculations could also be done locally across multiple frames.

    I'm not sure I'd expect games to be using the cloud for these sort purposes any time soon though. It would require GPU based servers, like NVidia's GRID and probably consumes hundreds of times more computing power than Titanfall uses for its servers. Even at a reduced rate and sharing data across multiple users that doesn't sound viable.
    Reply +1
  • Alcifer 06/10/2013

    @khanthony, not to downplay Azure, which is a very interesting resource, but I do think Sony also has a strength in their online offering.

    At least for the present, PSN is a much more open platform than Live, allowing developers to use their own servers and allow cross platform play, which is why games like FF:XIV, Timesplitters and War Thunder are being released on PS4 but not XB1.

    As with the hardware architectures, these are two very different approaches so it should be interesting see how things pan out.
    Reply 0
  • Alcifer 06/10/2013

    @SF1, we've heard about hardware for encoding and decoding audio and video on PS4. So it does certainly have some fixed function hardware. Though most of what is confirmed has come from interviews early on after the reveal. There are also some nice details about their hUMA setup but they're pretty technical.

    I agree it would be very nice to have Sony publish an overview of the system architecture as a whole to be able to make more direct comparisons.
    Reply +4
  • Alcifer 06/10/2013

    @CMcK, the distinction I'd make between the Cell on the PS3 and the GPU on the PS4 is that there is nowhere near as much requirement to use it for any particular purpose.

    The Cell chip effectively gave the PS3 only 33% of the CPU of the 360, the rest of it being made up of SPUs which required very specific program architecture. In the case of the PS4 (based on current clockspeed and accounting for SHAPE) it has maybe 80% of the CPU power of the XBox as well as having 50% additional GPU compute.

    To balance the comparison of the two next gen consoles: I'd probably say that if the PS4 uses brute force, it also has more flexibility in how those resources are applied. The XB1 is perhaps more efficient in the resources it has, but it is much more rigid in how those can be used.

    A good comparison of these from launch titles would be Forza 5 and KZ:SF. Forza 5 makes excellent use of the XBox system capabilities, but it could probably be implemented with only minor changes on the PS4. KZ:SF pushes heavy graphics workloads for its lighting and rendering, making use of a huge G-buffer (5 floating point buffers), which could not be replicated on the XBox.
    Reply +3
  • Alcifer 06/10/2013

    Ultimately this generation of consoles are very similar and it will be fascinating to see which approach works well most consistently. The brute force approach Sony have taken or the efficient approach Microsoft have chosen.
    Not picking on you in particular, but why do people think the XBox is more efficient than the PS4, or that the PS4 approach is brute force?

    You could equally say, overclocking the whole system and tacking on multiple fixed function processors is brute force, and adding extra CUs and ACEs with flexible memory access is a more elegant/efficient solution.

    These are clearly two very different solutions to enhancing what is effectively the same base hardware, it just seems odd that they're characterised in this way.
    Reply +7