PDA

Archiv verlassen und diese Seite im Standarddesign anzeigen : Übersetzung von "Hat der "brilineare" Filter eine Zukunft?"


Aqualon
2003-10-28, 18:34:51
Nachdem laut Leonidas nichts dagegen spricht, können wir ja den Artikel Hat der "brilineare" Filter eine Zukunft? (http://www.3dcenter.org/artikel/2003/10-26_a.php) übersetzen.

Da der Artikel aus sehr viel Text besteht, dürfte sich ne 3-Teilung pro Seite (oder auch mehr?) wohl am besten machen.

Wer würde denn mitmachen?

Ich hab gerade Zeit und übernehme mal auf Seite 1 die Einleitung (bis zur Überschrift "Brilinear" - eine sinnvolle Leistungssteigerung?").

Aqua

CannedCaptain
2003-10-28, 20:08:42
ich probier mich mal am rest der seite 1

huha
2003-10-28, 20:55:50
Ich mach Seite 2.

-huha

huha
2003-10-28, 21:15:02
Erste Hälfte von Seite 2 ist fertig, wo soll ich das Zeug hinposten? Hier rein?

-huha

Aqualon
2003-10-28, 21:40:39
Original geschrieben von huha
Erste Hälfte von Seite 2 ist fertig, wo soll ich das Zeug hinposten? Hier rein?

Genau :)

Aqua

huha
2003-10-28, 21:52:47
Die komplette 2. Seite ist bald fertiggestellt :naughty:
Der Autor dieses Textes sollte vielleicht mal gelobt werden, ehrlich. So ne fordernde, spannende Übersetzung hatte ich lange nicht mehr =)


-huha

Aqualon
2003-10-28, 21:58:57
Seite 1 — Teil 1

Will "brilinear" filtering persist?

Introduction

In the meantime the expression "brilinear" has been established for Nvidia´s towards bilinear filtering shifted pseudo-trilinear filtering. Although this term is semantically nonsense, we use it in this column.

There is an alarming development for a while: Newer Cards and/or newer drivers lower the image quality without being asked to do so. On a GeForce4Ti 4200 a better texture image quality (via anisotropic filtering) can be produced with current drivers (compared to a GeForceFX 5900 Ultra). On the other side the following is correct: The 4x-FSAA-Quality of the GeForce256 is by default better as with GeForce3 (and also as with GeForce 5900 Ultra). Because here stands 4x ordered grid supersampling vs. 4x ordered grid multisampling. But one does not speak of a backward step thereby ...

Before we rush into comparisons, a clean definiton of the bilinear and trilinear filter should take place.

The same textures can be supplied in different resolutions, in each case scaled down by factor 2 (per axis). These "sub-textures" are called MIP maps. Example: A 512x512 base-texture has before-computed MIP maps in the quantities 256x256, 128x128, 64x64, 32x32 etc. Since the amount of data decreases for each new MIP map (one could also say "MIP texture"), the additional memory consumption in the graphics card´s RAM is tolerable: about a third of the memory used by the base-texture.

An uncompressed 32-Bit-texture needs nevertheless 1 MB RAM for the size 512x512, all its MIP maps together need additional 0.33 MB RAM. Bigger textures than 512x512 are encountered quite rarely, since one needs higher texture resolutions also only with very high screen resolutions. And we are in the middle of the topic already: Because if an far-off object is texturized, it consists of fewer pixels, therefore (dynamically selected and computed by the 3D-Hardware) a smaller texture resolution is taken automatically. GeForce cards recalculate the MIP stage in each triangle for each 2x2 pixel block, which is accurate enough.

Bilinear filtering means to mix 4 texel from the "most suitable" MIP map to one color value. In fact so that each of the four values gets a suitable "weight", which is described here in detail. Now during texturizing, when the texture is put over the polygon, no texture value may be "forgotten" in the long run. Accordingly the size of the MIP texture is selected.

If the next larger MIP texture is applied, texture shimmering (so called aliasing) would appear, since the texture would be scanned already too coarse meshed with four samples per color value. Here the first disadvantage occurs, if one filters from only one MIP map: The "Level of Detail" has to be determined in such a way, that in any case aliasing is avoided, cause no one would (of course) accept texture shimmering. Distributed over the polygon, the pixels get more or less texture details depending upon that.

If then one resorts to the next smaller MIP map, "MIP banding" occurs. This is an artifact, which results from the fact that textures with high and low resolutions adjoin in the picture. In motion, these "bandings" are moved like a bow wave before oneself. The LOD formula provides altogether the highest possible sharpness degree, which is evenly achievable with bilinear filter. In addition, that means that the first MIP Band will be found relatively far "in front" of the picture and that those artifacts accordingly annoy.

Also for trilinear filtering there are accurate definitions. Microsoft specifies for Direct3D: "The texture filter to use between mipmap levels is trilinear mipmap interpolation. The rasterizer linearly interpolates pixel color, using the texels of the two nearest mipmap textures.".

The OpenGL definition even indicates a mathematical formula, but we are content with a periphrasis: The trilinear filter "crossfades" the two most adequate MIP maps evenly and gently. So, every pixel got the same sharpness (exception: skewed represented textures. For this anisotropic filtering is necessary, which however should not be our topic for today.) The formula for the trilinear filtering is of course selected in such a way, that the textures are as sharp as possible, but texture shimmering just does not appear.

The advantage of trilinear filtering is not only the consistent level of texture details. By the way, the over-all sharpness is not higher, "only" better distributed. The great advantage immediately evident in motion lies in the fact that the MIP Banding disappears: By a smoothly "crossfade" of the MIP textures there are no more sudden changes in the detail degree. Thus the bow waves are practically eliminated.

Let’s have a look at the "brilinear" filter now.

---

Aqua

Edit 1: Titel des Artikels am Anfang ergänzt.
Edit 2: Titel des Artikels geändert, thx@aths
Edit 3: Anmerkungen von aths eingearbeitet.

CannedCaptain
2003-10-28, 22:02:08
Instead of the appearing of the MIP-band-artefact during the use of trilinear filtering the whole area is overdrawn with a so called “Tri-Band”. The zone where actually two MIP-stages are used were meant with that. There is just a very small streak where just one MIP-texture determines the colour of the pixel.

“Brilinear” filtering likewise interpolates textures suppressing MIP-banding. The tri-band is significantly minified though. Within broad ranges solely bilinear filtering is exercised. Please keep in mind that “brilinear” is just an artificial word without any official use. “Pseudo-trilinear” sounds more technical in the first instance but it describe the filter just as insufficiently. The effect shall be illustrated with images by imitating MIP-map-colourings with a paint application.

Bilinear filtering causes hard bands at the point where different detail resolutions forgather.

Trilinear filtering always fades smoothly between two different detail resolutions.

This “range of fade” (“tri-band”) is obviously smaller during “brilinear” filtering – broad segments are excessively filtered bilinear.

Rolling back some time: When the GeForce256 was released the GeForce3 and its Multisampling-Anti-Aliasing which went easy on resources were already in development. Later, when the GeForce3 was available the GeForceFX was already in development, too. The special logic of the filter which is needed for brilinear filtering had to be implemented into the hardware, i.e. brilinear filtering is no compromise solution but a planned device. Before evaluating this filter and its future we should take a look at another feature.

For Sure, we focus our attention on anti-aliasing. Supersampling was exchanged for multisampling. While using the same subpixel mask supersampling provides better quality because for each subpixel the texture is sampled again (respectively the pixelshader is passed through). Multisampling calculates per triangle per pixel just one single value for the colour that applies to all subpixels which are covered by the triangle in that pixel. Consequently there is no improvement of the texture quality as it was possible with supersampling. In addition multisampling does not eliminate alphatest-artefacts – supersampling does. One could imply in a sneering way that multisampling hardware was developed to “cheat” better performance in anti-aliasing.

In spite of that it is worth to take a closer look: It was not pure performance that was achieved with the GeForce3 series because it still drops about 40% during the use of 2x anti-aliasing. When that card was released it was disproportional fast compared to available CPUs though that this weakness was not noticed intensively. Supersampling causes a drop of 50% -- why sacrificing quality for a little performance boost?

The great attainment of the GeForce3 utters in the more efficient sample raster in 2x anti-aliasing. Beside the little gain of performance there was achieved an obvious hike of quality. The fact that multisampling does not increase the quality of textures can be unattended because the GeForce3 facilitates 8° anisotropy during the filtering of textures (GeForce256 and GeForce2 just about 2°).

Still remaining the advantage using alpha textures with supersampling: Well, alpha testing is a “bad method” anyway aggravating the cooperation with anisotropic filtering. As anisotropic filtering sways in future the extinction of alpha testing is apparent. Additionally mentioned the geometric force of modern graphic cards gets more and more powerful i.e. complex structures will be modelled and not realised with alpha textures. Then multisampling will act absolutely efficient and everybody will be satisfied.

Let us draw a conclusion: The replacement of supersampling by multisampling meant approximately to acquiesce disadvantages in quality. Using other quality features which act independently of anti-aliasing provide better quality and cause less performance drops. Required is a hardware support for these features that needs relevant space of the chip architecture. Rakishly reworded one could say: “A clever mind saves power”.

How does that work with “brilinear” filtering? Modern TMUs (units that handle texture sampling) supply per clock only one bilinear sample. Realising trilinear filtering either two TMUs or two clocks are required. Consequently trilinear filtering halves the fill rate theoretically. “Brilinear” filtering works in broad ranges excessively bilinear thus seemingly a lot of work can be saved. In reality it is not as efficient.

huha
2003-10-28, 22:02:42
So, fast fertig.
Es fehlen nur noch ein bis zwei Sätze ganz am Schluß, da hatte ich erstens keine Lust mehr und zweitens weiß ich nicht, wie man in den Raum stellen übersetzen könnte.

Jetzt geht's los =)

The maximum fill rate has always been a highly inaccurate specification to estimate the real performance. Rendering of 3D Graphics is a complex process, there’s virtually no game that comes close to using large amounts of the theoretical fill rate of a graphics card. The extra work caused by trilinear filtering is chip-internal and is mostly done with ease. By switching to bilinear filtering one does by no means achieve twice as high frame rates.

“Brilinear” is, in practice, only giving a marginal boost in performance. However, the main problem is that there is no reasonable alternative to compensate the disadvantages. Let’s remind ourselves of texture improvement by supersampling. Nowadays textures are improved by anisotropic filtering, being far more efficient than supersmpling. Whereas supersampling would use 16 samples per pixel, anisotropic filtering can achieve the same quality with a maximum of 4 samples.

There isn’t such an option in the “brilinear” filter. When working with higher resulutions, the basic problem of the “brilinear” filter is still existent: Sensitivity for “bow waves”. These are generally occuring if filtering is not fully trilinear. Though, visibility is another question.

To be a bit more precise: Generally this artifact does also exist with trilinear filters. But the “bow waves” are stretched to the maximum reasonable amount (These are caused by MIP-Mapping, upon trilinear filtering being based) so that they are virtually disappearing. We have prepared some synthetic screenshots here:

[BILDER]

Bilineare Filterung ist nicht mehr zeitgemäß.

Bilinear filtering is not state-of-the-art anymore.

Der trilineare Filter liefert viel bessere Übergänge.

Trilinear filtering delivers much better transitions.

Mit "brilinear" nimmt die Sichtbarkeit der Detailstufen-Wechsel zu. In Bewegung ist der Effekt deutlich stärker wahrnehmbar als auf einem Standbild.

When using “brilinear” filtering, visibility of detail levels increases. When moving, this effect is recognized more distinct than on a still picture.

What can be done against bilinear filtering is also applicable for “brilinear” filtering: anisotropic filtering. For this will be featured in a latter article, we will just make a short excursion. AF displaces MIP-Levels “away” (e.g. in the background), artifacts caused by MIP-banding are therefore less visible. On the other hand, AF is consuming a huge amount of performance. Now, one could argue that no one would play without AF today, but there are two remaining objections:

One is the fact that NVIDIA is also “optimizing” anisotropic filtering. The degree of anisotropy is reduced both Angle-dependant and in general. 8° AF should mean every pixel getting such amounts of anisotropic filtering being necessary to increase quality; up to 8 AF-Samples and finally then aliasing is avoided through blur. With NVIDIAs “optimizations” enabled, 8° anisotropic filtering is applied to very for small areas.

Now angle-optimization is not obligatory in most drivers. For a short time, the user is forced to apply only a maximum of 2° anisotropic filtering on the primary texture stage. In Max Payne, where certain basic textures are on stage 1, this is clearly visible: blur. The GeForce4 Ti series could be reconfigured by the registry or – more easily – by means of tools like RivaTuner or aTuner. This makes sense as an option (unfortunately, newer drivers do not support this anymore) but should always remain an “opt-in” choice.

aths, der vorige Absatz war wohl einer der schwierigsten im ganzen Text...

The second objection concerns the small GeForce FX graphic cards featuring only limited amounts of fill rate. High AF-levels are not to be run a priori, as a result limitations of brilinear filtering can not be concealed so easily.

For a fairly long time NVIDIA forces FX-users to cope with “brilinear” filtering in UT2003. According to our tests, the actual quality difference is marginal.
HIER FEHTL EIN SATZ!
The most important word in this sentence is “optional”.

enjoy! =)
-huha

Aqualon
2003-10-28, 22:03:18
Original geschrieben von huha
Der Autor dieses Textes sollte vielleicht mal gelobt werden, ehrlich. So ne fordernde, spannende Übersetzung hatte ich lange nicht mehr =)

Hast du schon mal einen Artikel von Leonidas übersetzt? Der schreibt noch verdrehtere Sätze wie aths. Hier konnte man ja wenigstens meistens einigermaßen wörtlich vorgehen ohne jeden Satz komplett umzubasteln.

Aber ich geb dir Recht, spannend ist so eine Übersetzung schon immer (und vorallem interessanter als irgendwelche klassischen Texte *g*).

Aqua

Gohan
2003-10-28, 22:05:31
Bilineare Filterung ist nicht mehr zeitgemäß.

Bilinear filtering isn't longer up to date , klingt für mich... hm natürlicher :)

huha
2003-10-28, 22:06:29
Leonidas macht nur verdrehte Sätze, das kann ich auch ;) Nö, war schon spannend und ich bin jetzt richtig erfrischt, nachdem ich den ganzen Tag mehr oder weniger gar nichts gemacht hab...

Das fängt ja toll an... gleich n Fehler im ersten Satz... o.O
-huha

Aqualon
2003-10-28, 22:08:27
Original geschrieben von Gohan
Bilinear filtering isn't longer up to date , klingt für mich... hm natürlicher :)

Da würd ich dann aber "Bilinear filtering is no longer up to date" sagen.

Aqua

Edit: "In den Raum stellen" würde ich mit "to moot (http://dict.leo.org/?p=lURE.&search=to%20moot) = zur Debatte stellen" übersetzen. Das dürfte es sinngemäß gut treffen.

huha
2003-10-28, 22:11:12
Hups!

Da kommt ja noch was unten an der 2. Seite - das mach ich aber heute nicht mehr... :o

-huha

Aqualon
2003-10-28, 22:15:18
Original geschrieben von huha
Da kommt ja noch was unten an der 2. Seite - das mach ich aber heute nicht mehr... :o

Hat mich schon gewundert, wie flink du gewesen bist ;)

Falls du möchtest kannst ja den 2. Teil "Zwang zur Leistung?" für heute Nacht freigeben, falls jemand nicht schlafen kann und unbedingt was übersetzen möchte :D

Aqua

huha
2003-10-28, 22:24:13
Also gut... der zweite Teil ist für heute Nacht freigegeben... aber nur für heute nacht ;)

-huha

aths
2003-10-28, 22:56:24
Statt "fade" würde ich "blending" nehmen.

Original geschrieben von Aqualon
Hast du schon mal einen Artikel von Leonidas übersetzt? Der schreibt noch verdrehtere Sätze wie aths.Hehe, als aths. Leonidas hatte mal gemeint, dass ich gefälligst längere Sätze schreiben soll. Seitdem schreibe ich längere Sätze.

Eine Übersetzung kann ruhig sinngemäß sein. Werden aus zwei Sätzen drei gemacht, oder werden Sätze verschmolzen, ist das kein Problem. Wo ich etwas meine Bedenken habe, ist, ob der Ton rüberkommt. Einige Dinge stehen zwischen den Zeilen, oder werden nur sanft angedeutet.

Der Titel z. B. ist so gemeint: "Wird sich der brilineare Filter durchsetzen?"

StevenB
2003-10-28, 23:09:47
Eure Texte sind gut!

Ich lese mal, vieleicht fällt mir was auf.

mfg

Steven

edit:

This makes sense as an option (unfortunately, newer drivers do not support this anymore) but should always remain an “opt-in” choice.

makes hört sich immer kommisch an. Mir fällt aber gerade nich ein anderes wort für makes ein. Makes, ist praktisch machen. Mh, klingt komisch.

Aqualon
2003-10-28, 23:30:04
Original geschrieben von aths
Hehe, als aths.

Ok, das ist meine alte Deutschschwäche. Der Unterschied zwischen wie und als ist mir nie wirklich klar geworden und das mache ich öfters falsch.

Original geschrieben von aths
Statt "fade" würde ich "blending" nehmen.

Ok, werd ich morgen beim Korrekturlesen meines Textes beachten.

Original geschrieben von aths
Der Titel z. B. ist so gemeint: "Wird sich der brilineare Filter durchsetzen?"
Vorschlag 1:
Is there a future for the "brilinear" filter?

Vorschlag 2:
Will there be acceptance for the "brilinear" filter in the future?

Vorschlag 3:
Will the "brilinear" filter become accepted in the future?
bzw.
Is the "brilinear" filter becoming accepted in the future?

Bei Vorschlag Drei bin ich mir mit der Grammatik nicht sicher. Wobei die 2. Variante mit "is becoming" IMHO ziemlich falsch klingt.

Aqua

aths
2003-10-28, 23:58:54
Original geschrieben von Aqualon
Ok, das ist meine alte Deutschschwäche. Der Unterschied zwischen wie und als ist mir nie wirklich klar geworden und das mache ich öfters falsch."Leonidas schreibt längere Sätze als aths." "Leonidas schreibt gleich lange Sätze wie aths". Wenn Gleichheit, "wie". Wenn Unterschiede, "als". Oder anders gesagt, sobald es eine Steigerungsform gibt ("lang" -> "länger"), dann "als".

Ich merks mir so: "Das ist so gut, wie..." "Das ist besser, als..."


Das Wort "future" muss im Titel gar nicht umbedingt vorkommen :) Er könnte z. B. lauten: "Will "brilinear" filtering persist?" Man kann sich ruhig vom Wort lösen :) wenn das rüberkommt, was gemeint war.

Groß will ich euch in die Übersetzung aber nicht reinreden, da ich das unabhängig vom Artikel sehe. Ich les morgen oder so aber trotzdem mal drüber (englisch hören oder lesen fällt mir deutlich leichter, als es zu schreiben.)

aths
2003-10-29, 00:11:20
Original geschrieben von huha
So, fast fertig.
Es fehlen nur noch ein bis zwei Sätze ganz am Schluß, da hatte ich erstens keine Lust mehr und zweitens weiß ich nicht, wie man in den Raum stellen übersetzen könnte."behaupten".

In dieser Kolumne erlauben wir uns einen weniger technisch Ansatz und stellen mal in den Raum, dass eine Option, die nur sehr wenig Bildqualität für ein spürbares Leistungsplus opfert, auf jeden Fall wünschenswert ist.

In this column, we admit to have a less technical point of view and claim every option increasing performance for a very small amount of image quality, is preferable per se.

(Der Satzbau ist sicher fürchterlich falsch, und die Übersetzung holprig.)

Aqualon
2003-10-29, 00:11:49
Original geschrieben von aths
"Will "brilinear" filtering persist?"
Das klingt wirklich gut. Knapp und prägnant. Werd ich gleich mal übernehmen, bevor mir noch ein paar Sprachkonstrukte zum davonrennen einfallen ;)

Original geschrieben von aths
Ich les morgen oder so aber trotzdem mal drüber (englisch hören oder lesen fällt mir deutlich leichter, als es zu schreiben.)
Ist eine gute Idee, dann siehst du ja, wo wir vielleicht eine Aussage falsch interpretiert haben und kannst uns darauf hinweisen.

Aqua

P.S. Danke für die Erklärung, klingt ja eigentlich ganz einfach :)

huha
2003-10-29, 08:05:44
So, ich hab jetzt nichts zu tun, (2 Stunden später Schule =)) und fang schonmal die Übersetzung des 2. Teils der 2.Seite an.

-huha

huha
2003-10-29, 08:48:05
So, fast fertig.

Hier:

Constraint to performance?

NVIDIA advertises the FX series with “Cinematic Computing” and “Engineered with passion for perfection” and does not, however, allow use of true trilinear filtering in any Direct3D applications. We consider that ridiculous. NVIDIA defines how the client has to use its “Cinematic Computing” hardware: Performance-optimized. The user is deprived of the option to get the best quality possible.

For UT2003 requiring plenty of power it could be absolutely senseful to lower the filter quality. MIP banding is virtually eliminated by “brilinear” in UT2003. But why should this, according to NVIDIA, be the only choice to play UT2003? Why is this paradigm extended on all Direct3D applications and games?

ATi marketed 16° AF aggressively in times of the original Radeon and Radeon 8500. If anisotropic filtering was turned on, trilinear filtering could not be used due to a hardware limitation (up to Radeon 9200) so the user had to stick with bilinear anisotropy. This fact and the extreme angle dependence of Radeon AF was justified as a reasonable limitation: Important areas would get anisotropic filtering and trilinear AF would consume far too much performance for such a small quality increase.

One cannot let that one unanswered. GeForce3 users often used only bilinear AF, but in less demanding games, they could enjoy trilinear AF. The relatively faint angle-dependence of anisotropic filtering on GeForce3 series graphics cards was also contributing to good texture quality. At ATi no one every thought of taking the user trilinear filtering (without AF). Correctly trilinear filtered textures are possible on every Radeon – this is the minimum possible what a decent 3D graphics card has to offer.

Let’s summarize our concrete provisos against NVIDIAs new filter: Quality gained by trilinear filtering is partially destroyed. There is no resource-saving option that could correct this shortage. Furthermore, we doubt the sense of these “optimizations”: Fast graphics cards gain frames one does not need anyway (and sacrifice quality one would probably want). The entry-level cards do not become fast enough to display modern games at high resolutions fluently.

In the future, shader power will be of a great importance. Just because the GeForce FX does not have very fast shaders there is enough time to use high-quality texture filters at least. After all, this is the only advantage of NVIDIAs graphics cards: Shader power and anti-aliasing quality are clearly dominated by the competitors. No one has to offer such high-quality anisotropic filtering à la GeForce 3 (and successors) but NVIDIA does not do anything with this advantage. Of what use is AF that is based on the “brilinear” filter and takes over its disadvantages?

Offering the user options is greatly appreciated by us per se, even if the options do not make much sense. “Brilinear” could be the crucuial plus on the way to playability in certain situations when needing fairly visible quality decreases. Therefore we see possibilities for this option. In the longer run we assume that high-level trilinear anisotropic filtering – or even better techniques (there are better techniques than trilinear filtering) will prevail. In our opinion it is somehow paradox to develop sophisticated pixelshaders but to continously decrease texture filter quality.



-huha

huha
2003-10-29, 14:25:46
So, fertig - manches war wirklich hart und da mir besonders die englischen Idiome fehlen, mußte ich wohl oder übel manchen Satz auslassen. Nun gut, ich hoffe, daß ihr schön viele Fehler findet und meine Übersetzung in grund und Boden kritisiert ;)


Fehlender (Halb-)Satz
The GeForce 4 Ti is being slowed down artificially at the moment by the drivers (AF-Stage optimizations do not work anymore and apart from some benchmarks, Early-Z-Occlusion has been deactivated). The FX series are, however, pushed – by means of forced worse image quality, which the user can not deactivate, even if he wished to. It is somehow paradox that NVIDIA advertises the new drivers with the release highlight “enhanced image quality for both anisotropic filtering and anti-aliasing”.


To be constrained to accept missing true trilinear filtering is unacceptable in the year 2003.
Fehlender Satz
To deprive the user trilinear filtering seems to be an act of desperation from NVIDIA. The introduction of “brilinear” filtering, planned a long time ago was surely not thought of as the maximum quality. If NVIDIA is not going to correct this very quickly one cannot see the products as a decent piece of hardware for a computer gamer anymore.


Gegen Ende wurde es so schwierig, daß ich einfach Umschreibungen wählen mußte, ich hoffe aber, daß der letzte Satz richtig rüberkommt

-huha

aths
2003-10-29, 15:13:04
Einige Anmerkungen, direkt in den Quote gesetzt.

Original geschrieben von Aqualon
Seite 1 — Teil 1

Will "brilinear" filtering persist?

Introduction

In the meantime the malapropism "brilinear" has been established for Nvidia´s towards bilinear filtering shifted pseudo-trilinear filtering. "malapropism" versteht kein Mensch, zumindest keiner, der Englisch als Fremdsprache in webtauglichem Maße kann. Man könnte hier z.B. "expression" nehmen, das ist weniger wortgetreu, wird aber sofort verstanden. Although this term is semantically nonsense, we want to use it in this column.
Wir tun's ja auch, hehe... Das "want" klingt für mich hier eher so, als ob man wollte, aber nicht könnte. "We use it in this column" würde ausreichen.

There is an alarming development for a while: Newer Cards and/or newer drivers lower the image quality without being asked to do so. On a GeForce4Ti 4200 one can produce better texture image quality (via anisotropic filtering) with current drivers, as with a GeForceFX 5900 Ultra."Can [whatelse] better, than ..." wäre imo weniger holprig.
Now one could say,Das "nun könnte man sagen" ist eine füllende Passage, die könnte in der Übersetzung komplett entfallen, zumindest das "now".
on the other side the following is correct: The 4x-FSAA-Quality of the GeForce256 is by default better as with GeForce3 (and also as with GeForce 5900 Ultra). Because here stands 4x ordered grid supersampling vs. 4x ordered grid multisampling. But one does not speak of a backward step thereby ...

Before we rush into comparisons, a clean definiton of the bilinear and trilinear filter should take place.

The same textures can be supplied in different resolutions, in each case scaled down by factor 2 (per axis). These "sub-textures" are called MIP maps. Example: A 512x512 base-texture has before-computed MIP maps in the quantities 256x256, 128x128, 64x64, 32x32 etc. Since the amount of data decreases for each new MIP map (one could also say "MIP texture"), the additional memory consumption in the graphics card´s RAM is tolerable: about a third of the memory used by the base-texture.

An uncompressed 32-Bit-texture needs nevertheless 1 MB RAM for the size 512x512, all its MIP maps together need once more 0.33 MB RAM."noch mal" wird im Sinne von "zusätzlich" gebraucht, wäre imo mit "additional" besser übersetzt. Bigger textures than 512x512 are encountered quite rarely, since one needs higher texture resolutions also only with very high screen resolutions. And we are in the middle of the topic already: Because if an far-off object is texturized, it consists of fewer pixels, therefore (dynamically selected and computed by the 3D-Hardware) a smaller texture resolution is taken automatically. GeForce cards recalculate the MIP stage in each triangle for each 2x2 pixel block, which is exactly enough.

Bilinear filtering means to mix 4 texel from the "most suitable" MIP map to one color value. In fact so that each of the four values gets a suitable "weight", which is described here in detail. Now during texturizing, when the texture is put over the polygon, no texture value may be "overlooked" in the long run. [color=#009000]"übersehen" wäre imo im Kontext mit "forgotten" besser übersetzt.Accordingly the size of the MIP texture is selected.

If one took the next larger MIP texture, "If one took"? "Would the next lager MIP texture be applied, ...", "If the next larger..." eher was in der Art wäre imo flüssiger. it would already come to texture shimmering (so called aliasing), since one would then scan the texture with four samples per color value already too coarse meshed.Wieso wieder "one"? Here the first disadvantage reveals itself,Sie deckt sich nicht selbst auf :) "Here, the first disadvantage appears / comes to manner / occurs / ... if one filters from only one MIP map: The "Level of Detail" has to be determined in such a way, that in any case aliasing is avoided, cause nobody "no one", hier würde ich was mit one nehmen :)would (rightly)"of course" accept texture shimmering. Spread over the polygon, the pixels get more or less texture details depending upon that. Da wird nichts verteilt im Sinne von spread :) Mir fällt jetzt auch kein so richtig passendes Wort ein, "distributed" käme imo schon mal näher. Was in der Art...

If then one resorts to the next smaller MIP map, "MIP banding" occurs. This is an artifact, which results from the fact that textures with high and low resolutions collide in the picture. Sie stoßen zusammen, aber kollidieren nicht :) Sie grenzen aneinander. "adjoin" trifft's dann eher. In motion one moves these virtual "bandings" like a bow wave before oneself."one" moves?, "In motion, these "bandings" ..." Im Englischen hat man meist kurze Sätze. The LOD formula provides altogether for the highest possible sharpness degree,Die Formel liefert, aber nicht "für". Das "for" ist dort imo überflüssig. which is evenly attainable with bilinear filter. "achievable" wäre imo noch eine Idee näher dran. In addition, that means that the first MIP volume will be found relatively far "in front" of the picture and that those artifacts accordingly annoy."Band" bitte nicht mit "volume" übersetzen. "the first MIP-band" geht schon so.

Also for trilinear filtering there are accurate definitions. Microsoft specifies for Direct3D: "The texture filter to use between mipmap levels is trilinear mipmap interpolation. The rasterizer linearly interpolates pixel color, using the texels of the two nearest mipmap textures.".

The OpenGL definition even indicates a mathematical formula, but we are content with a periphrasis: The trilinear filter crossfades the two most adequate MIP maps evenly and gently. "crossfades" dann bitte in Anführungszeichen, imo ist die Bedeutung so nur im Audio-Bereich gebräuchlich.Thus it is reached that the "sharpness degree" is equal for each pixel (exception: distorted represented textures. Wozu komplizierte Konstrukte mit thus? "So, every pixel got the same "sharpnes". Verzerrte Texturen sind imo mit distorted suboptimal übersetzt. "skewed" oder was in die Richtung käme der Sinnbedeutung näher.For this anisotropic filtering is necessary, which however should not be our topic for today.) The formula for the trilinear filtering is naturally selected in such a way, that the textures are as sharp as possible, but texture shimmering just does not appear."natürlich" wurde von mir im Sinne von "of course" gebraucht :)

The usefulness of trilinear filtering does not only lie in the even distribution of the texture details (the sharpness degree is altogether not better by the way, but "only" better distributed). "lie" heißt liegen, "darin liegt ..." sollte man imo nicht unbedingt wörtlich übersetzen. "The advantage of trilinear filtering is not only the consistent level of texture details. By the way, the over-all sharpness is not higher, "only" better shared." oder was in der Art...The giant advantage immediately evident in motion lies in the fact that the MIP Banding disappears:Der Vorteil ist groß im Sinne von großartig, also "great". By a gentle crossfade of the MIP textures there are no more sudden changes in the detail degree. "smoothly" träfe es besser, als "gentle".Thus the bow waves are practically eliminated.

Let´s have a look at the "brilinear" filter now.
Leo wird heulen, aber ich bestehe auf ein Apostroph. Unicode dafür ist ’, also ’
---

Aqua

Edit 1: Titel des Artikels am Anfang ergänzt.
Edit 2: Titel des Artikels geändert, thx@aths

aths
2003-10-29, 15:45:02
Original geschrieben von huha
So, fast fertig.

Hier:

Constraint to performance?Habe im Moment keine Zeit, über den ganzen Text zu sehen. Mit "Zwang" meine ich meistens "force". Die Kapitelüberschrift "Zwang zur Leistung?" ist so gemeint: [Wird man] zur Leistung gezwungen? "Force to choose high framerate?" etc. ginge eher in die Richtung.

Aqualon
2003-10-29, 16:16:32
Original geschrieben von aths
Einige Anmerkungen, direkt in den Quote gesetzt.

Danke für die Anmerkungen. Hab sie größtenteils mit eingearbeitet.

Verteilt habe ich jetzt durchgehend mit distributed übersetzt, da ich nichts besseres gefunden habe ("shared" heisst eher etwas wie "gemeinsam benutzt" und passt IMO nicht wirklich).

Aqua

CannedCaptain
2003-10-30, 01:17:53
Ich bin von Freitag bis Sonntag nicht im Lande, sodass ich einen von Euch bitte, den Text zusammenzufügen und Probe zu lesen. Es wäre schade, wenn wir zu viel Zeit verlierten.

Madmanniman würdest Du meinen Text auf Inhalt und Grammatim prüfen?

Danke im Vorraus und ein schönes Wochenende!

CannedCaptain
2003-10-30, 01:27:01
Was mir noch aufgefallen ist: Guckt mal bitte, ob die Schreibweisen der Grafikartennamen immer gleich sind (Geforce3 / GeForce3 / GeForce 3 usw.)und ob ähnliche Bezeichnungen nicht von jedem Übersetzer krass anders bezeichnet worden sind.

aths
2003-10-30, 13:33:47
Ich schreibe GeForce2, GeForce3, GeForce4, GeForceFX, allerdings die vierstelligen Zahlen (ohne FX) dann mit Leerzeichen. Statt GeForce 5900 hätte es allerdings der Konsistenz wegen eher GeForceFX 5900 heißen sollen. Nvidia trennt allerdings auch das FX mit einem Leerzeichen ab, werde das wohl im nächsten Artikel beachten.

Aqualon
2003-10-30, 14:16:59
Bitte folgendes Posting beachten: http://www.forum-3dcenter.org/vbulletin/showthread.php?s=&postid=1319419#post1319419

Original geschrieben von CannedCaptain
"Brilinear" - A wise performance boost? fehlende Überschrift von mir ergänzt

Instead of the appearing of the MIP band artefact during the use of trilinear filtering the whole area is overdrawn with a so called "Tri-Band".das zweifache of the klingt IMO nicht gut. "Instead of the appearing MIP band artefact..." The zone where actually two MIP stages are used were"was" nicht "were" meant with that. There is just a very small streak "band" statt "streak" passt da besserwhere just one MIP texture determines the colour of the pixel.

"Brilinear" filtering likewise interpolates textures suppressing "to suppress" MIP banding. The tri-band is significantly minified da fände ich reduced besser though. Within broad ranges solely bilinear filtering is exercised "used" trifft es da wohl eher. Please keep in mind that "brilinear" is just an artificial word without any official use. "Pseudo-trilinear" sounds more technical in the first instance but it describes the filter just as insufficiently muss "insufficient" heissen. The effect shall be illustrated with images by imitating MIP map colourings colored MIP maps with a paint application.

Bilinear filtering causes hard bands at the point where different detail resolutions forgather da würde ich "adjoin" schreiben.

Trilinear filtering always fades smoothly between two different detail resolutions.

This "range of fade" ("tri-band") is obviously smaller during "brilinear" filtering – broad segments are excessively filtered bilinear.

Rolling back some time: When the GeForce256 was released the GeForce3 and its Multisampling-Anti-Aliasing sollen wir das durchgehend klein oder groß schreiben? which went easy on resources were already in development. Later, when the GeForce3 was available the GeForce FX Leerzeichen zwischen GeForce und FX was already in development, too das "too" ist hier IMO überflüssig. The special logic of the filter which is needed for brilinear filtering had to be implemented into the hardware, i.e. brilinear filtering is no compromise solution but a planned device. Before evaluating this filter and its future we should take a look at another feature.

For Sure "sure", we focus our attention on anti-aliasing. Supersampling was exchanged for multisampling. While using the same subpixel mask supersampling provides better quality because for each subpixel the texture is sampled again (respectively the pixelshader is passed through). Multisampling calculates per triangle per pixel just one single value for the colour that applies to all subpixels which are covered by the triangle in that pixel. Consequently there is no improvement of the texture quality as it was possible with supersampling. In addition multisampling does not eliminate alphatest-artefacts – supersampling does. One could imply in a sneering way that multisampling hardware was developed to "cheat" better performance in anti-aliasing.

In spite of that it is worth to take a closer look: It was not pure performance that was achieved with the GeForce3 series because it still drops about 40% during the use of 2x anti-aliasing. When that card was released it was disproportional fast compared to available CPUs though that "that" ist hier IMO überflüssig this weakness was not noticed intensively. Supersampling causes a drop of 50% -- why sacrificing quality for a little performance boost?

The great attainment of the GeForce3 utters in the more efficient sample raster in 2x anti-aliasing. Beside the little gain of performance there was achieved an obvious hike of quality "there was an... achieved". The fact that multisampling does not increase the quality of textures can be unattended because the GeForce3 facilitates 8° anisotropy during the filtering of textures (GeForce256 and GeForce2 just about 2°).

Still remaining the advantage using alpha textures with supersampling There still remains the advantage of supersampling when using alpha textures: Well, alpha testing is a "bad method" anyway aggravating the cooperation with anisotropic filtering. As anisotropic filtering sways in future the extinction of alpha testing is apparent. Additionally mentioned the geometric force of modern graphic cards gets more and more powerful i.e. complex structures will be modelled and not realised with alpha textures. Then multisampling will act absolutely efficient and everybody will be satisfied.

Let us draw a conclusion: The replacement of supersampling by multisampling meant approximately to acquiesce "accept" ist hier wohl verständlicher disadvantages in quality. Using other quality features which act independently of anti-aliasing provide better quality and cause less performance drops. Required is a hardware support for these features that needs relevant space of the chip architecture. Rakishly reworded one could say: "A clever mind saves power".

How does that work with "brilinear" filtering? Modern TMUs (units that handle texture sampling) supply per clock only one bilinear sample "supply only ... sample per clock. Realising "To realise" trilinear filtering either two TMUs or two clocks are required. Consequently trilinear filtering halves the fill rate theoretically. "Brilinear" filtering works in broad ranges excessively bilinear thus seemingly a lot of work can be saved. In reality it is not as efficient. Was meinst du mit dem Satz? Ich hätte da "It looks very different in reality however." gesagt

Hab meine Anmerkungen in grün dahinter geschrieben. Überarbeite bitte deinen Text wenn möglich bis heute Abend, ansonsten mache ich das.

Aqua

Edit 1: Könntest du bitte die normalen Hochkommas ("") verwenden, wie es aths auch in seinem Text gemacht hat?
Edit 2: Hab noch ein paar Anmerkungen gemacht

Aqualon
2003-10-30, 14:20:23
Original geschrieben von CannedCaptain
Ich bin von Freitag bis Sonntag nicht im Lande, sodass ich einen von Euch bitte, den Text zusammenzufügen und Probe zu lesen. Es wäre schade, wenn wir zu viel Zeit verlierten.


Ich könnte es heute noch fertigstellen, ansonsten müsste es huha übernehmen.

Aqua

EDIT: Noch eine allgemeine Frage: Wie sollen wir die ganzen MIP-irgendwas Begriffe schreiben? Zum Beispiel: MIP-Texture, MIP Texture, mip-texture, mip texture oder MIP texture?

EDIT 2: Habe mich jetzt für die Schreibweise MIP groß und das zusätzliche Wort klein ohne Bindestrich entschieden

Aqualon
2003-10-30, 14:47:58
Bitte folgendes Posting beachten: http://www.forum-3dcenter.org/vbulletin/showthread.php?s=&postid=1319419#post1319419


Ich habe die Übersetzungen von huha in diesem Posting zusammengefasst und die fehlenden Sätze (soweit sie mir aufgefallen sind) ergänzt.

Seite 2 übersetzt von huha

---
A statement of profit and loss

The maximum fill rate has always been a highly inaccurate specification to estimate the real performance. Rendering of 3D Graphics is a complex process, there’s virtually no game that comes close to using large amounts of the theoretical fill rate of a graphics card. The extra work caused by trilinear filtering is chip-internal and is mostly done with ease. By switching to bilinear filtering one does by no means achieve twice as high frame rates.

"Brilinear" is, in practice, only giving a marginal boost in performance. However, the main problem is that there is no reasonable alternative to compensate the disadvantages. Let’s remind ourselves of texture improvement by supersampling. Nowadays textures are improved by anisotropic filtering, being far more efficient than supersmpling. Whereas supersampling would use 16 samples per pixel, anisotropic filtering can achieve the same quality with a maximum of 4 samples.

There isn’t such an option in the "brilinear" filter. When working with higher resulutions, the basic problem of the "brilinear" filter is still existent: Sensitivity for "bow waves". These are generally occuring if filtering is not fully trilinear. Though, visibility is another question.

To be a bit more precise: Generally this artifact does also exist with trilinear filters. But the "bow waves" are stretched to the maximum reasonable amount (These are caused by MIP mapping, upon trilinear filtering being based hier hätte ich "upon which trilinear filtering is based") so that they are virtually disappearing. We have prepared some synthetic screenshots here:

Bilinear filtering is not state-of-the-art anymore....is no longer up to date

Trilinear filtering delivers much better transitions.

When using "brilinear" filtering, the visibility of detail level changes increases. When moving, this effect is recognized more distinct than on a still picture hier würde ich das recognized weglassen.

What can be done against bilinear filtering is also applicable for "brilinear" filtering: anisotropic filtering. For this will be featured in a latter "later" nicht "latter" article, we will just make a short excursion. AF displaces MIP levels "away" (e.g. in the background), artifacts caused by MIP banding are therefore less visible. On the other hand, AF is consuming a huge amount of performance. Now, one could argue that no one would play without AF today, but there are two remaining objections:

One is the fact that Nvidia is also "optimizing" anisotropic filtering. The degree of anisotropy is reduced both Angle-dependant kleinschreiben "angle-dependant" and in general. 8° AF should mean every pixel getting such amounts of anisotropic filtering being necessary to increase quality; up to 8 AF-Samples and finally then aliasing is avoided through blur. With Nvidias "optimizations" enabled, 8° anisotropic filtering is applied to very for small areas. Thus the MIP volumes are pushed less far to the back, than otherwise possible. Satz ergänzt.

Now angle-optimization is not obligatory in most drivers. For a short time, the user is forced to apply only a maximum of 2° anisotropic filtering on the primary texture stage "on" ist hier falsch, sinngemäß muss es "except for" heissen. In Max Payne, where certain basic textures are on stage 1, this is clearly visible: blur. The GeForce4 Ti series could be reconfigured by the registry or – more easily – by means of tools like RivaTuner or aTuner. This makes sense as an option (unfortunately, newer drivers do not support this anymore) but should always remain an "opt-in" choice.

The second objection concerns the small GeForce FX graphic cards featuring only limited amounts of fill rate. High AF-levels are not to be run a priori, as a result limitations of brilinear filtering can not be concealed so easily.

For a fairly long time Nvidia forces FX-users to cope with "brilinear" filtering in UT2003. According to our tests, the actual quality difference is marginal. In this column, we admit to have a less technical point of view and claim every option, that sacrifices a very small amount of image quality for a noticeable performance increase, to be preferable per se Satz ergänzt.. The most important word in this sentence is "optional".


Constraint to performance? Nach aths Einwand, würde ich das mit "Forced to high frame rates?" übersetzen.

Nvidia advertises the FX series with "Cinematic Computing" and "Engineered with passion for perfection" and does not, however, allow use of true trilinear filtering in any Direct3D applications. We consider that ridiculous. Nvidia defines how the client has to use its "Cinematic Computing" hardware: Performance-optimized. The user is deprived of the option to get the best quality possible.

For UT2003 requiring plenty of power it could be absolutely senseful to lower the filter quality. MIP banding is virtually eliminated by "brilinear" in UT2003. But why should this, according to Nvidia, be the only choice to play UT2003? Why is this paradigm extended on all Direct3D applications and games?

ATI marketed 16° AF aggressively in times of the original Radeon and Radeon 8500. If anisotropic filtering was turned on, trilinear filtering could not be used due to a hardware limitation (up to Radeon 9200) so the user had to stick with bilinear anisotropy. This fact and the extreme angle dependence of Radeon AF was justified as a reasonable limitation: Important areas would get anisotropic filtering and trilinear AF would consume far too much performance for such a small quality increase.

One cannot let that one unanswered. GeForce3 users often used only bilinear AF, but in less demanding games, they could enjoy trilinear AF. The relatively faint angle-dependence of anisotropic filtering on GeForce3 series graphics cards was also contributing to good texture quality. At ATI no one every thought of taking the user trilinear filtering (without AF). Correctly trilinear filtered textures are possible on every Radeon – this is the minimum possible what a decent 3D graphics card has to offer.

Let’s summarize our concrete provisos "reservations" würde hier IMO einfach verständlich sein against Nvidias new filter: Quality gained by trilinear filtering is partially destroyed. There is no resource-saving option that could correct this shortage. Furthermore, we doubt the sense of these "optimizations": Fast graphics cards gain frames one does not need anyway (and sacrifice quality one would probably want). The entry-level cards do not become fast enough "...are not fast enough..." to display modern games at high resolutions fluently.

In the future, shader power will be of a great importance "The up to now irrelevant "shader power" will gain strong significance in the future" bringt den Sinn wohl eher rüber. Just because the GeForce FX does not have very fast shaders there is enough time to use high-quality texture filters at least. After all, this is the only advantage of Nvidias graphics cards: Shader power and anti-aliasing quality are clearly dominated by the competitors. No one has to offer such high-quality anisotropic filtering à la GeForce 3 (and successors) but Nvidia does not do anything with this advantage. Of what use is AF that is based on the "brilinear" filter and takes over its disadvantages?

Offering the user options is greatly appreciated by us per se, even if the options do not make much sense in detail. "Brilinear" could be the crucuial plus on the way to playability in certain situations when needing fairly visible quality decreases. Der Satz gibt den Sinn nicht wieder, besser: "Brilinear" could be the crucial plus in certain situations to gain playability by accepting hardly visible quality losses.. Therefore we see possibilities for this option. In the longer run we assume that high-level trilinear anisotropic filtering – or even better techniques (there are better techniques than trilinear filtering) will prevail. In our opinion it is somehow paradox to develop sophisticated pixelshaders but to continously decrease texture filter quality.

Thus a flat aftertaste remains Satz ergänzt.: The GeForce 4 Ti is being slowed down artificially at the moment by the drivers (AF-Stage optimizations do not work anymore and apart from some benchmarks, Early-Z-Occlusion has been deactivated). The FX series are, however, pushed – by means of forced worse image quality, which the user can not deactivate, even if he wished to. It is somehow paradox that Nvidia advertises the new drivers with the release highlight "enhanced image quality for both anisotropic filtering and anti-aliasing".

To be constrained to accept missing true trilinear filtering is unacceptable in the year 2003. To avoid nitpicking, we see the divers aggravations of AF quality "only" disapproving (besides also ATI takes part in these useless games) To deprive the user of trilinear filtering seems to be an act of desperation from Nvidia. The introduction of "brilinear" filtering, planned a long time ago was surely not thought of as the maximum quality. If Nvidia is not going to correct this very quickly one cannot see the products as a decent "worthy" erscheint mir hier besser piece of hardware for a computer gamer anymore.
---

Korrekturen sind vorgenommen

Aqua

Gohan
2003-10-30, 15:42:37
Wir bereits gesagt, im ersten Satz beie Seite 2:

The maximum fill rate has always been a highly inaccurate specification to estimate the real performance.

Wir war das? Zeitangaben stehen immer am Anfag oder am Ende des Satzes ;)

Aqualon
2003-10-30, 16:01:33
Original geschrieben von Gohan
The maximum fill rate has always been a highly inaccurate specification to estimate the real performance.

Wir war das? Zeitangaben stehen immer am Anfag oder am Ende des Satzes ;)

Ist always eine Zeitangabe?

Wenn ich das always vor das has been verschiebe klingt der Satz für mich ziemlich holprig:

"The maximum fill rate always has been a highly inaccurate specification to estimate the real performance."

Und ans Ende passt das always definitiv nicht.

Aqua

Gohan
2003-10-30, 16:10:59
Original geschrieben von Aqualon
Ist always eine Zeitangabe?

Wenn ich das always vor das has been verschiebe klingt der Satz für mich ziemlich holprig:

"The maximum fill rate always has been a highly inaccurate specification to estimate the real performance."

Und ans Ende passt das always definitiv nicht.

Aqua

Ah sorry, ich revediere meine Aussage, ich muss mal langsam wieder aus dem japanischen Denken rauskommen ^^ Always passt da hin :-(

Aqualon
2003-10-30, 17:25:24
Zur Veröffentlichung freigegebene Übersetzung des Artikels.

---

Will "brilinear" filtering persist?

Introduction

In the meantime the expression "brilinear" has been established for Nvidia's towards bilinear filtering shifted pseudo-trilinear filtering. Although this term is semantically nonsense, we use it in this column.

There is an alarming development for a while: Newer Cards and/or newer drivers lower the image quality without being asked to do so. On a GeForce4Ti 4200 a better texture image quality (via anisotropic filtering) can be produced with current drivers (compared to a GeForceFX 5900 Ultra). On the other side the following is correct: The 4x-FSAA-Quality of the GeForce256 is by default better as with GeForce3 (and also as with GeForce 5900 Ultra). Because here stands 4x ordered grid supersampling vs. 4x ordered grid multisampling. But here it's not called "a step back".

Before we rush into comparisons, a clean definiton of the bilinear and trilinear filter should take place.

The same textures can be supplied in different resolutions, in each case scaled down by factor 2 (per axis). These "sub-textures" are called MIP maps. Example: A 512x512 base-texture has before-computed MIP maps in the quantities 256x256, 128x128, 64x64, 32x32 etc. Since the amount of data decreases for each new MIP map (you could also say "MIP texture"), the additional memory consumption in the graphics card's RAM is tolerable: about a third of the memory used by the base-texture.

An uncompressed 32-Bit-texture needs nevertheless 1 MB RAM for the size 512x512, all its MIP maps together need additional 0.33 MB RAM. Bigger textures than 512x512 are encountered quite rarely, since you need higher texture resolutions also only with very high screen resolutions. And we are in the middle of the topic already: Because if an far-off object is texturized, it consists of fewer pixels, therefore (dynamically selected and computed by the 3D-Hardware) a smaller texture resolution is taken automatically. GeForce cards recalculate the MIP stage in each triangle for each 2x2 pixel block, which is accurate enough.

Bilinear filtering means to mix 4 texel from the "most suitable" MIP map to one color value. In fact so that each of the four values gets a suitable "weight", which is described here in detail. Now during texturizing, when the texture is put over the polygon, no texture value may be "forgotten" in the long run. Accordingly the size of the MIP texture is selected.

If the next larger MIP texture is applied, texture shimmering (so called aliasing) would appear, since the texture would be scanned already too coarse meshed with four samples per color value. Here the first disadvantage occurs, if you filter from only one MIP map: The "Level of Detail" has to be determined in such a way, that in any case aliasing is avoided, cause no one would (of course) accept texture shimmering. Distributed over the polygon, the pixels get more or less texture details depending upon that.

If then you resort to the next smaller MIP map, "MIP banding" occurs. This is an artifact, which results from the fact that textures with high and low resolutions adjoin in the picture. In motion, these "bandings" are moved like a bow wave in front of you. The LOD formula provides altogether the highest possible sharpness degree, which is evenly achievable with bilinear filter. In addition, that means that the first MIP Band will be found relatively far "in front" of the picture and that those artifacts accordingly annoy.

Also for trilinear filtering there are accurate definitions. Microsoft specifies for Direct3D: "The texture filter to use between mipmap levels is trilinear mipmap interpolation. The rasterizer linearly interpolates pixel color, using the texels of the two nearest mipmap textures.".

The OpenGL definition even indicates a mathematical formula, but we are content with a periphrasis: The trilinear filter "crossfades" the two most adequate MIP maps evenly and gently. So, every pixel got the same sharpness (exception: skewed represented textures. For this anisotropic filtering is necessary, which however should not be our topic for today.) The formula for the trilinear filtering is of course selected in such a way, that the textures are as sharp as possible, but texture shimmering just does not appear.

The advantage of trilinear filtering is not only the consistent level of texture details. By the way, the over-all sharpness is not higher, "only" better distributed. The great advantage immediately evident in motion lies in the fact that the MIP Banding disappears: By a smoothly "crossfade" of the MIP textures there are no more sudden changes in the detail degree. Thus the bow waves are practically eliminated.

Let’s have a look at the "brilinear" filter now.


"Brilinear" - A wise performance boost?

Instead of the appearing MIP band artefact during the use of trilinear filtering the whole area is overdrawn with a so called "Tri-Band". The zone where actually two MIP stages are used was meant with that. There is just a very small band where just one MIP texture determines the colour of the pixel.

"Brilinear" filtering likewise interpolates textures to suppress MIP banding. The tri-band is significantly reduced though. Within broad ranges solely bilinear filtering is used. Please keep in mind that "brilinear" is just an artificial word without any official use. "Pseudo-trilinear" sounds more technical in the first instance but it describes the filter just as insufficient. The effect shall be illustrated with images by imitating colored MIP maps with a paint application.

Bilinear filtering causes hard bands at the point where different detail resolutions adjoin.

Trilinear filtering always blends smoothly between two different detail resolutions.

This "range of blending" ("tri-band") is obviously smaller during "brilinear" filtering – broad segments are excessively filtered bilinear.

Rolling back some time: When the GeForce256 was released the GeForce3 and its multisampling anti-aliasing which went easy on resources were already in development. Later, when the GeForce3 was available the GeForce FX was already in development. The special logic of the filter which is needed for brilinear filtering had to be implemented into the hardware, i.e. brilinear filtering is no compromise solution but a planned device. Before evaluating this filter and its future we should take a look at another feature.

For sure, we focus our attention on anti-aliasing. Supersampling was exchanged for multisampling. While using the same subpixel mask supersampling provides better quality because for each subpixel the texture is sampled again (respectively the pixelshader is passed through). Multisampling calculates per triangle per pixel just one single value for the colour that applies to all subpixels which are covered by the triangle in that pixel. Consequently there is no improvement of the texture quality as it was possible with supersampling. In addition multisampling does not eliminate alphatest-artefacts – supersampling does. You could imply in a sneering way that multisampling hardware was developed to "cheat" better performance in anti-aliasing.

In spite of that it is worth to take a closer look: It was not pure performance that was achieved with the GeForce3 series because it still drops about 40% during the use of 2x anti-aliasing. When that card was released it was disproportional fast compared to available CPUs wherefore this weakness was not noticed intensively. Supersampling causes a drop of 50% -- why sacrificing quality for a little performance boost?

The great attainment of the GeForce3 utters in the more efficient sample raster in 2x anti-aliasing. Beside the little gain of performance there was an obvious hike of quality achieved. The fact that multisampling does not increase the quality of textures can be unattended because the GeForce3 facilitates 8° anisotropy during the filtering of textures (GeForce256 and GeForce2 just about 2°).

There still remains the advantage of supersampling when using alpha textures: Well, alpha testing is a "bad method" anyway aggravating the cooperation with anisotropic filtering. As anisotropic filtering sways in future the extinction of alpha testing is apparent. Additionally mentioned the geometric force of modern graphic cards gets more and more powerful i.e. complex structures will be modelled and not realised with alpha textures. Then multisampling will act absolutely efficient and everybody will be satisfied.

Let us draw a conclusion: The replacement of supersampling by multisampling meant approximately to accept disadvantages in quality. Using other quality features which act independently of anti-aliasing provides better quality and causes less performance drops. Required is a hardware support for these features that needs relevant space of the chip architecture. Rakishly reworded you could say: "A clever mind saves power".

How does that work with "brilinear" filtering? Modern TMUs (units that handle texture sampling) supply only one bilinear sample per clock. To realise trilinear filtering either two TMUs or two clocks are required. Consequently trilinear filtering halves the fill rate theoretically. "Brilinear" filtering works in broad ranges excessively bilinear thus seemingly a lot of work can be saved. It looks very different in reality however.








A statement of profit and loss

The maximum fill rate has always been a highly inaccurate specification to estimate the real performance. Rendering of 3D Graphics is a complex process, there’s virtually no game that comes close to using large amounts of the theoretical fill rate of a graphics card. The extra work caused by trilinear filtering is chip-internal and is mostly done with ease. Switching to bilinear filtering however doesn't mean for sure that you get twice as high frame rates.

"Brilinear" is, in practice, only giving a marginal boost in performance. However, the main problem is that there is no reasonable alternative to compensate the disadvantages. Let’s remind ourselves of texture improvement by supersampling. Nowadays textures are improved by anisotropic filtering, being far more efficient than supersmpling. Whereas supersampling would use 16 samples per pixel, anisotropic filtering can achieve the same quality with a maximum of 4 samples.

There isn’t such an option in the "brilinear" filter. When working with higher resulutions, the basic problem of the "brilinear" filter is still existent: Sensitivity for "bow waves". These are generally occuring if filtering is not fully trilinear. Though, visibility is another question.

To be a bit more precise: Generally this artifact does also exist with trilinear filters. But the "bow waves" are stretched to the maximum reasonable amount (These are caused by MIP mapping, upon which trilinear filtering is based) so that they are virtually disappearing. We have prepared some synthetic screenshots here:

Bilinear filtering is no longer up to date.

Trilinear filtering delivers much better transitions.

When using "brilinear" filtering, the visibility of detail level changes increases. When moving, this effect is more distinct than on a still picture.

What can be done against bilinear filtering is also applicable for "brilinear" filtering: anisotropic filtering. There will be another upcoming article addressing that issue, so we will just make a short excursion. AF displaces MIP levels "away" (e.g. in the background), artifacts caused by MIP banding are therefore less visible. On the other hand, AF is consuming a huge amount of performance. Now, you could argue that no one would play without AF today, but there are two remaining objections:

The first is the fact that Nvidia is also "optimizing" anisotropic filtering. The degree of anisotropy is reduced both angle-dependent and in general. 8° AF should mean every pixel getting such amounts of anisotropic filtering being necessary to increase quality; up to 8 AF-Samples and finally then aliasing is avoided through blur. With Nvidias "optimizations" enabled, 8° anisotropic filtering is applied to very for small areas. Thus the MIP volumes are pushed less far to the back, than otherwise possible.

Now angle-optimization is not obligatory in most drivers. For a short time, the user is forced to apply only a maximum of 2° anisotropic filtering except for the primary texture stage. In Max Payne, where certain basic textures are on stage 1, this is clearly visible: blur. The GeForce4 Ti series could be reconfigured by the registry or – more easily – by means of tools like RivaTuner or aTuner. This makes sense as an option (unfortunately, newer drivers do not support this anymore) but should always remain an "opt-in" choice.

The second objection concerns the small GeForce FX graphic cards featuring only limited amounts of fill rate. High AF-levels are not to be run a priori, as a result limitations of brilinear filtering can not be concealed so easily.

For a fairly long time Nvidia forces FX-users to cope with "brilinear" filtering in UT2003. According to our tests, the actual quality difference is marginal. In this column, we admit to have a less technical point of view and claim every option, that sacrifices a very small amount of image quality for a noticeable performance increase, to be preferable per se. The most important word in this sentence is "optional".


Forced to high frame rates?

Nvidia advertises the FX series with "Cinematic Computing" and "Engineered with passion for perfection" and does not, however, allow use of true trilinear filtering in any Direct3D applications. We consider that ridiculous. Nvidia defines how the client has to use its "Cinematic Computing" hardware: Performance-optimized. The user is deprived of the option to get the best quality possible.

For UT2003 requiring plenty of power it could be absolutely senseful to lower the filter quality. MIP banding is virtually eliminated by "brilinear" in UT2003. But why should this, according to Nvidia, be the only choice to play UT2003? Why is this paradigm extended on all Direct3D applications and games?

ATI marketed 16° AF aggressively in times of the original Radeon and Radeon 8500. If anisotropic filtering was turned on, trilinear filtering could not be used due to a hardware limitation (up to Radeon 9200) so the user had to stick with bilinear anisotropy. This fact and the extreme angle dependence of Radeon AF was justified as a reasonable limitation: Important areas would get anisotropic filtering and trilinear AF would consume far too much performance for such a small quality increase.

You cannot let that one unanswered. GeForce3 users often used only bilinear AF, but in less demanding games, they could enjoy trilinear AF. The relatively faint angle-dependence of anisotropic filtering on GeForce3 series graphics cards was also contributing to good texture quality. At ATI no one ever thought of taking the user trilinear filtering (without AF). Correctly trilinear filtered textures are possible on every Radeon – this is the minimum a decent 3D graphics card has to offer.

Let’s summarize our concrete reservations against Nvidias new filter: Quality gained by trilinear filtering is partially destroyed. There is no resource-saving option that could correct this shortage. Furthermore, we doubt the sense of these "optimizations": Fast graphics cards gain frames you don't need anyway (and sacrifice quality you would probably want). The entry-level cards are not fast enough to display modern games at high resolutions fluently.

The up to now irrelevant "shader power" will gain strong significance in the future. Just because the GeForce FX does not have very fast shaders there is enough time to use high-quality texture filters at least. After all, this is the only advantage of Nvidias graphics cards: Shader power and anti-aliasing quality are clearly dominated by the competitors. No one has to offer such high-quality anisotropic filtering à la GeForce 3 (and successors) but Nvidia does not do anything with this advantage. Of what use is AF that is based on the "brilinear" filter and takes over its disadvantages?

Offering the user options is greatly appreciated by us per se, even if the options do not make much sense in detail. "Brilinear" could be the crucial plus in certain situations to gain playability by accepting hardly visible quality losses. Therefore we see possibilities for this option. In the longer run we assume that high-level trilinear anisotropic filtering – or even better techniques (there are better techniques than trilinear filtering) will prevail. In our opinion it is somehow paradox to develop sophisticated pixelshaders but to continously decrease texture filter quality.

Thus a flat aftertaste remains: The GeForce 4 Ti is being slowed down artificially at the moment by the drivers (AF-Stage optimizations do not work anymore and apart from some benchmarks, Early-Z-Occlusion has been deactivated). The FX series are, however, pushed – by means of forced worse image quality, which the user can not deactivate, even if he wished to. It is somehow paradox that Nvidia advertises the new drivers with the release highlight "enhanced image quality for both anisotropic filtering and anti-aliasing".

To be constrained to accept missing true trilinear filtering is unacceptable in the year 2003. To avoid nitpicking, we see the divers aggravations of AF quality "only" disapproving (besides also ATI takes part in these useless games) To deprive the user of trilinear filtering seems to be an act of desperation from Nvidia. The introduction of "brilinear" filtering, planned a long time ago was surely not thought of as the maximum quality. If Nvidia is not going to correct this very quickly you cannot see the products as a worthy piece of hardware for a computer gamer anymore.

---

Die Übersetzung ist hiermit vorläufig abgeschlossen. Vielen Dank an alle die bei der Übersetzung mitgeholfen haben :up:

Ich habe Leonidas den Artikel in der oben stehenden Form zur Veröffentlichung gegeben. Weitere Fehlerkorrekturen können natürlich weiter gepostet werden. Diese müssen aber dann von Leonidas eingepflegt werden.

Aqua

mapel110
2003-10-30, 21:11:42
2. absatz
But one does not speak of a backward step thereby ...

but here it's not called "a step back".
////////////////
7.absatz
Here the first disadvantage occurs, if one filters from only one MIP map:

...if you filter only from one MIP map:
////////////////
8.absatz
In motion, these "bandings" are moved like a bow wave before oneself.

in front of you.
////////////////
A statement of profit and loss

The maximum fill rate has always been a highly inaccurate specificationto estimate the real performance. Rendering of 3D Graphics is a complex process, there’s virtually no game that comes close to using large amounts of the theoretical fill rate of a graphics card. The extra work caused by trilinear filtering is chip-internal and is mostly done with ease. By switching to bilinear filtering one does by no means achieve twice as high frame rates.

hm, da würde ich den letzten satz anders machen, weiss nur nicht wie. "one" hört sich jedenfalls doof an. kommt übrigens recht häufig im text vor.
//////////////////////
Bilinear filtering is no longer up to date.

Trilinear filtering delivers much better transitions.

When using "brilinear" filtering, the visibility of detail level changes increases.
When moving, this effect is more distinct than on a still picture.

What can be done against bilinear filtering is also applicable for "brilinear" filtering:anisotropic filtering. For this will be featured in a later article

For that theme there will be another upcoming article
/////////////////////
Forced to high frame rates?
4.absatz

One cannot let that one unanswered.

you cannot let this unanswered.
//////////////////
hoffe, davon gefällt irgendwas besser :)

Aqualon
2003-10-30, 21:42:16
Original geschrieben von mapel110
2. absatz
But one does not speak of a backward step thereby ...

but here it's not called "a step back".
Übernommen.


7.absatz
Here the first disadvantage occurs, if one filters from only one MIP map:

...if you filter only from one MIP map:
Übernommen.


In motion, these "bandings" are moved like a bow wave before oneself.

in front of you.
Übernommen.


By switching to bilinear filtering one does by no means achieve twice as high frame rates.

hm, da würde ich den letzten satz anders machen, weiss nur nicht wie. "one" hört sich jedenfalls doof an. kommt übrigens recht häufig im text vor.
Mit dem one geb ich dir Recht, das kommt wirklich sehr häufig vor. You ist da wohl doch persönlicher und spricht den Leser eher an.

Hab es jetzt folgendermaßen übersetzt:

"Switching to bilinear filtering however doesn´t mean for sure that you get twice as high frame rates."


For this will be featured in a later article

For that theme there will be another upcoming article
Hab ich jetzt mit "There will be another upcoming article addressing that issue" übersetzt.


One cannot let that one unanswered.

you cannot let this unanswered.
Übernommen.

hoffe, davon gefällt irgendwas besser :) Danke, hat mir gut gefallen =)

Hab auch noch ein paar andere Kleinigkeiten beim korrigieren gefunden und ausgebessert.

Aqua

Anmerkungen, die nach diesem Posting folgen, sind nicht mehr in http://www.forum-3dcenter.org/vbulletin/showthread.php?s=&postid=1319419#post1319419 eingearbeit.

Canned im Exil
2003-10-31, 12:07:05
Ich froi mich!

Leonidas
2003-11-01, 14:15:05
Superb, ist online:
http://www.3dcenter.org/artikel/2003/10-26_a_english.php

Leonidas
2003-11-02, 15:35:33
Wider aller Erwartungen läuft der Artikel ganz gut im englischsprachigen Raume.

Exxtreme
2003-11-02, 15:44:02
Original geschrieben von Leonidas
Wider aller Erwartungen läuft der Artikel ganz gut im englischsprachigen Raume.
Ich habe ja auch Propaganda im R3D-Forum gemacht. ;)

zeckensack
2003-11-02, 16:15:57
http://www.beyond3d.com/forum/viewtopic.php?t=8799 :naughty:

Aqualon
2003-11-02, 20:41:34
Original geschrieben von zeckensack@B3D
Oh my, what an interesting translation

War das jetzt positiv oder negativ gemeint? Ein bisschen Feedback, was an der Übersetzung gut oder schlecht ist und was man beim nächsten Mal besser machen könnte, wäre nicht schlecht :)

Aqua

mapel110
2003-11-02, 21:54:28
der engl artikel ist auch auf warp2search verlinkt worden.
http://www.warp2search.net/modules.php?name=News&file=article&sid=15166

zeckensack
2003-11-03, 20:13:42
Original geschrieben von Aqualon
War das jetzt positiv oder negativ gemeint? Ein bisschen Feedback, was an der Übersetzung gut oder schlecht ist und was man beim nächsten Mal besser machen könnte, wäre nicht schlecht :)

Aqua Es holpert ein wenig. Leo's Texte sind eben eine Herausforderung :)
Für eine komplette Korrekturlesung ist's jetzt wohl zu spät. Beim nächsten Mal bin ich aber hoffentlich aktiv dabei.

Aqualon
2003-11-03, 21:27:27
Original geschrieben von zeckensack
Es holpert ein wenig. Leo's Texte sind eben eine Herausforderung :)

Der Artikel ist von aths und nicht von Leonidas, aber das macht die Sache trotzdem nicht wirklich einfacher ;)

Aber solange der Text inhaltlich korrekt ist und trotz aller Holprigkeit verständlich bleibt, kann man ihn ja so lassen.

Original geschrieben von zeckensack
Für eine komplette Korrekturlesung ist's jetzt wohl zu spät.
Wenn du Zeit hast oder dir langweilig ist, kannst ja die Korrekturen trotzdem posten. Ob Leonidas sie noch mit einbaut, bleibt ihm überlassen.

Original geschrieben von zeckensack
Beim nächsten Mal bin ich aber hoffentlich aktiv dabei.
Wäre toll, tatkräftige Hilfe ist immer willkommen =)

Und dass zumindest mein Englisch stellenweise noch stark verbesserungswürdig ist, geb ich gerne zu. Aber dafür kann man ja üben :)

Aqua

nggalai
2003-11-04, 21:18:37
Hola,

zuerstmal: gute Arbeit! :) im Thread hier (http://www.forum-3dcenter.org/vbulletin/showthread.php?s=&threadid=104704) hab' ich mich nach eurer Organisation in Sachen "Endabnahme" und "Lektorat" informiert. Ich habe in dem Thread auch was von "kapitalen Böcken" geschrieben, aber bevor ihr mir an die Gurgel springt, es geht mir nicht um Rechtschreibung--eher darum, dass Satzbau und Textorganisation für Leser ohne Deutschkenntnisse häufig eher schwer bis nicht zu verstehen sind. Das folgende soll also nicht als "Korrektur" verstanden werden, sondern vielmehr als Denkanstoss, wie ihr bei zukünftigen Artikelübersetzungen 'ran gehen könnt.

Kurz gesagt--am besten versucht ihr nicht, den Text zu "übersetzen", sondern macht eine Übertragung des Inhalts. Redewendungen und deutsche Textorganisation lassen sich nur selten 1:1 ins Englische übersetzen, ohne dass der Leser anfängt zu lachen oder "Nicaragua???" zu schreien. Versucht daher, den Sinn eines Satzes / Paragraphen rüberzubringen, und klammert euch nicht zusehr an das Original-Wording (besonders dann nicht, wenn ihr was von Leo übersetzen sollt ;D). Texte fürs Internet und 3DCenter zu übertragen ist was anderes, als an der Schule während einer Prüfung eine Übersetzung abgeben zu müssen.

Soviel zur Vorrede, hier als Input einfach mal die ersten paar Absätze, wie ich sie übertragen würde resp. als Lektor anpassen täte. Die sind auch nicht 100% sauber (natürlich ned, hab zuviel Bier intus), aber sollten euch eine Idee geben, wie Ihr Übersetzungsarbeiten fürs 3DCenter in Zukunft angehen könnt.


Introduction

During the course of translation, the expression "brilinear" has established itself as a viable replacement for aths's original "pseudo-trilinear" nomenclature. Although the term itself is pure linguist nonsense, it will be used throughout this article to describe Nvidia's "optimised" trilinear / bilinear filtering.

An alarming development has been observed for some time now: New graphics boards and / or drivers are released, lowering the theoretically possible image quality without any user intervention. What's worse, these "optimisations" cannot always be disabled. Simply put: With current drivers, a GeForce 4Ti 4200 will deliver higher texture quality (using anisotropic filtering) than Nvidia's most expensive offering, the GeForce FX 5900 Ultra. Some may interject that much the same can be said of GeForce 256's and GeForce 3's anti-aliasing quality: The GeForce 256's 4x FSAA is, by default, superior to what is available without third-party tools on a GeForce 3. The former supports 4x ordered grid supersampling, the latter 4x ordered grid multisampling. But this design decision usually has not been called a step back ...

Before we rush into comparisons, it is necessary to agree on clean definitons of both bilinear and trilinear filtering.

One single texture can be supplied at different resolutions, in each case scaled down by a factor of 2 (per axis). These "sub-textures" are called MIP maps. Example: A 512x512 base-texture can include precomputed MIP maps at 256x256, 128x128, 64x64, 32x32, etc. pixels. Since the amount of data decreases for each new MIP map (you could also say "MIP texture"), the larger video memory footprint is tolerable: about an additional third of the memory already used by the base-texture.

Nevertheless, an uncompressed 32 bits texture needs 1 MB at 512x512, all its MIP maps together claim 0.33 MB of video memory. Textures larger than 512x512 are encountered quite rarely as you usually need them only at very high screen resolutions. Which takes right into the topic: Because far-off objects consist of fewer pixels onscreen, a smaller texture (dynamically selected by the VGA) can be used for texturing said object. GeForce GPUs recalculate each triangle's MIP stage for each 2x2 pixels block, which is accurate enough.

93,
-Sascha.rb