- I-DLSS ye-NVIDIA ivelele ngekhwalithi yayo ephezulu ngenxa yokusebenzisa ama-Tensor cores kanye nehadiwe ezinikele.
- I-FSR ye-AMD iyindlela eguquguqukayo kakhulu futhi efinyeleleka kalula njengoba iwumthombo ovulekile futhi iyahambisana cishe nanoma iyiphi i-GPU.
- I-XeSS ye-Intel inikeza ibhalansi yokuncintisana, okwenza kube ngcono ukusebenza kwamakhadi ayo e-Arc.
Uma usanda kungena emidlalweni ye-PC noma usanda kuthuthukisa ikhadi lakho lehluzo, kungenzeka ukuthi uhlangane namagama afingqiwe azwakala sengathi ayisiwula: i-DLSS, i-FSR, kanye ne-XeSS. Empeleni, sikhuluma ngobuchwepheshe bokukhulisa izinga lokudlala obuklanyelwe ukwenza izimpilo zethu zibe lula, okuvumela imidlalo yethu ukuthi isebenze kahle kakhulu ngaphandle kokuba isithombe sibe yingxubevange yamaphikseli.
Umqondo ulula kodwa uhlakaniphile: ikhadi lehluzo lenza umdlalo ube nesisombululo esiphansi ukuze ugweme ukusebenza kanzima, bese, usebenzisa ama-algorithms noma ubuhlakani bokwenziwa , ugcwalisa izikhala ukuze ubukeke njengesithombe esinesinqumo esiphezulu kusikrini sakho. Empeleni, kuyicebo ukuthola ozimele ngomzuzwana (i-FPS) ngaphandle kokuthenga i-GPU engu-€3,000 njalo eminyakeni emibili.
I-NVIDIA DLSS: Iphayona lokufunda okujulile
Uma sikhuluma nge-DLSS (Deep Learning Super Sampling), singena emkhakheni wezinsizakalo ze-NVIDIA ezisezingeni eliphezulu. Lobu buchwepheshe abuyona nje isofthiwe; budinga ihadiwe ethile ebizwa ngokuthi i-Tensor Cores , etholakala kuphela kumakhadi e-RTX (kusukela ochungechungeni luka-2000 kuqhubeke). Amakhadi amadala e-GTX awafakiwe ngoba awanawo amakhono okucubungula i-AI.
I-DLSS isebenza ngendlela eyinkimbinkimbi kakhulu. Isebenzisa amanethiwekhi e-neural aqeqeshiwe ukwakha kabusha izithombe zekhwalithi ephezulu ngokusekelwe ekuxazululweni okuphansi kanye nama-vector okunyakaza. Ithuthuke kakhulu kuzo zonke izinguqulo zayo:
- I-DLSS1.0: Umzamo wokuqala, nakuba wawunemibono emisha, ngezinye izikhathi wawushiya isithombe singabonakali kahle.
- I-DLSS2.0: Igxathu elikhulu. Lapha baqala ukusebenzisa ulwazi oluvela kuzinhlaka zangaphambilini ukuze bafeze ubukhali obucishe bufane nobomdabu.
- I-DLSS3.0: Yethula i-Frame Generation ne-NVIDIA Reflex edumile, okuvumela dala ozimele bokwenziwa ukwandisa ukugeleza, yize kusebenza kuphela kumakhadi ochungechunge lwe-RTX 40.
- I-DLSS3.5: Wengeze iRay Reconstruction, eyenza Ukulandelela imisebe kubonakala kuhlanzekile kakhulu futhi ngaphandle komsindo.
- I-DLSS4.0: Yakhelwe uchungechunge lwe-RTX 50, ithuthukisa ukwenziwa kwamafreyimu amaningi kulowo nalowo okhiqizwe.
Ngokuphathelene nezindlela zayo, i-NVIDIA inikeza izinketho ezisukela kumodi ye-Quality (lapho ukukhulisa kubushelelezi) kuya ku-Ultra Performance, efanelekile ukudlala imidlalo ku-4K ngenkathi idlala ku-720p. Kukhona ne -DLAA, engahlose ukwandisa i-FPS , kodwa isebenzisa uhlobo lokuhlakanipha lokulwa nokungahlanganisi ukuze imiphetho ibukeke iphelele.
I-AMD FSR: Indinganiso evulekile neguquguqukayo
I-FidelityFX Super Resolution (FSR) iyimpendulo ye-AMD futhi inefilosofi ehluke ngokuphelele: iwumthombo ovulekile . Lokhu kusho ukuthi ayidingi ikhadi le-AMD ukuze isebenze; ungayivumela kuma-NVIDIA noma ama-Intel GPU, okwenza kube inketho yentando yeningi kakhulu emakethe.
Ngokungafani nezinguqulo zangaphambilini ze-DLSS, i-FSR yokuqala ayizange isebenzise i-AI, kodwa kunalokho i -algorithm yokukala indawo eyayithola imiphetho. Lokhu kwenze kwaba lula ukuyisebenzisa, kodwa ikhwalithi ebonakalayo yashiya okuningi okwakudingeka ngezindlela zokuhlasela. Kodwa-ke, izinto zishintshile:
- I-FSR 1: Ukulinganisa okuyisisekelo, kuyahambisana ngisho namakhadi ehluzo amadala kakhulu (uchungechunge lwe-RX 400).
- I-FSR 2: Wethula ukukala isikhathi, esondela kakhulu ku- Ukuzinza kokubona kwe-NVIDIA.
- I-FSR 3: Yethula ukwenziwa kohlaka kusetshenziswa ama-Fluid Motion Frames (AFMF), okuvumela imidlalo emidala ukuthi izwakale ibushelelezi kakhulu, futhi yayihambisana ne- I-Radeon RX 5000 nangaphezulu.
- I-FSR 4: Izindaba ezinkulu ukuhlanganiswa kwe Intelligence yokufakelwa ukuthuthukisa ukwakhiwa kabusha kwesithombe, okufana nalokho okwenziwa yizimbangi zayo.
I-AMD iphinde inikeze izindlela ezisukela ku-Quality kuya ku-Ultra Performance, ngisho nemodi ye -Native AA ku-FSR 3, evumela ukwenziwa kohlaka ngaphandle kokwehlisa isixazululo somdlalo.
I-Intel XeSS: Umncintiswano omusha wenza umbono oqinile
I-Intel ifike sekwephuzile ephathini, kodwa ayikho lapha ukuzokwenza i-splash. I-XeSS (Xe Super Sampling) iyinhlanganisela ethokozisayo. Njenge-NVIDIA, isebenzisa amamodeli e-AI ukuze ikhulise i-scaling, kodwa i-Intel ibilokhu iguquguquka kakhulu ngokuhambisana.
Kunezindlela ezimbili zokusebenzisa i-XeSS: imodi ye-XMX , etholakala kuphela kwihluzo ze-Intel Arc , esebenzisa ama-core e-AI azinikele ngekhwalithi ephezulu, kanye nemodi ye-DP4a, esebenza kusetshenziswa imiyalelo ejwayelekile ye-GPU futhi ehambisana namakhadi e-AMD kanye ne-NVIDIA , yize isebenza kancane.
Izinguqulo zayo zithuthuke ngokushesha. Ngenkathi inguqulo 1.x igxile ekukhulisweni okumsulwa, izinguqulo 2.0 kanye no-2.1 zethule ukukhiqizwa kohlaka (XeSS-FG) kanye nokunciphisa ukubambezeleka (XeSS-LL), okwenza ifane nezinye izinhlobo ezimbili ezinkulu. Ngokuphathelene nezindlela, mhlawumbe iyona enemininingwane kakhulu, inikeza amasethingi ayisithupha ahlukene , kufaka phakathi i-Ultra Quality Plus kulabo abafuna incazelo eyengeziwe.
Ukuqhathanisa okuqondile: Yimuphi okufanele uwusebenzise kumdlalo wakho?
Uma usemenyu yezilungiselelo zomdlalo futhi zonke izinketho zitholakala, ungesabi. Kukhona ukuvumelana okuvamile okusekelwe ekusebenzeni kanye nekhwalithi yokubuka . Uma unekhadi le-NVIDIA RTX, vele uye ku- DLSS ; yiyona ndlela ephucuke kakhulu. Uma une-GTX noma i-Radeon, i-FSR ivame ukuba yindlela engcono kakhulu ngenxa yokuzinza kwayo kanye nokusekelwa kwayo.
Kubasebenzisi be-Intel Arc, ukubeka i-XeSS phambili kunengqondo , njengoba ihadiwe ilungiselelwe yona. Kodwa-ke, kunezingqinamba, njengoba kwezinye izihloko ezithile, ukusetshenziswa kwe-XeSS kungadlula lokho kwe-FSR 2 ngokubonakalayo, kusondela ngendlela eyingozi kwikhwalithi ye-NVIDIA.
Ngokuphathelene nokusekelwa, i-DLSS isahamba phambili ngoba nje ibilokhu isemakethe isikhathi eside. I-FSR ilandela eduze ngenxa yokuhlanganiswa kwayo okulula konjiniyela, kuyilapho i-XeSS ikhula ngokushesha kodwa isekhona ngezihloko ezimbalwa.
Ukukhetha kokugcina kuncike njalo kuhadiwe yakho nokuthi ubuchwepheshe busetshenziswe kanjani kumdlalo ngamunye othize. Ngenkathi i-NVIDIA igxile ekusebenzeni okuphezulu nokukhawulelwe , i-AMD ne-Intel zihlose ukunikeza ulwazi olubushelelezi kubantu abaningi ngangokunokwenzeka, kungaba ngama-algorithms avulekile noma ukwesekwa kwe-cross-platform, ukuqinisekisa ukuthi izinga lohlaka liyinto eza kuqala kunoma yikuphi ukucushwa.
