SPIRAL CITY / VEHICLE SPEED TEST AND SPLINE POSITION TESTING

LightWaveGuru

Active member

BTW for these tests the scene is calculated with the day light kernel and ambient occlusion from octane so that the lights do not illuminate their surroundings.
this would increase the calculation times a lot and that is unnecessary for tests.
only for the final rendering after all bugs are removed from the scene, the path tracing kernel and GI will be used for the final rendering.

snip LWGuru
 

Sensei

TrueArt Support
Distances between cars are too uniform for me. There is no red lights and green lights for others. Waves of cars starting engine after green appeared.

See this:

Video from video-recorder installed in vehicle is analyzed by special program, and entire city is build, including other cars..
One could make such for cars only to get vehicle position, velocity, acceleration. Transfer data from real world to simulated world..

Notice how car in game engine is shaking, while driving. Streets are not perfectly flat. Car suspension also plays a role.
 
Hi Rene,
in deiner Stadt ist aber eine ganze Menge an Kleintransporter unterwegs. Ist wohl schon ein Blick in die Zukunft wenn keiner mehr aus dem Haus gehen will und alle nur mehr online einkaufen ;).

ciao
Thomas
 

LightWaveGuru

Active member
Hi Rene,
in deiner Stadt ist aber eine ganze Menge an Kleintransporter unterwegs. Ist wohl schon ein Blick in die Zukunft wenn keiner mehr aus dem Haus gehen will und alle nur mehr online einkaufen ;).

ciao
Thomas

Hallo Thomas,

ja so ist das wenn man Instanzierungstests entlang von Splines vornimmt. Dafür reichen 5 verschiedene Fahrzeuge (Proxys/Dummys). Für das Finale Rendering müssen es dann Erfahrungsgemäss ca. mindestens 15 pro Spur sein damit genug Variationsvielfalt über die Karosserien und die Farbkombinationen (Die Car-Shader sind auch instanziert mit verschiedenen Lack-Farben via Instance-ID über das Octane Shading-System) vorliegen. Diese Tests dienen ergo auch der zu erwartenen Speicherauslastung des VRAM damit man ungefähr kalkulieren kann wann der VRAM voll ist. Ich habe hier von 8-12 GB diverse Karten insofern muss alles auf die Karten abgestimmt sein mit dem kleinsten VRAM. Ergo 8 GB.
Mehr darf nicht adressiert werden. Derzeit sind noch ca. 1,5 GB frei. Insofern. Es werden dann nicht soviele Transporter für das Finale. LACH :)

Grüsse nach Wien!

snipLWGuru
 
Hallo Thomas,

ja so ist das wenn man Instanzierungstests entlang von Splines vornimmt. Dafür reichen 5 verschiedene Fahrzeuge (Proxys/Dummys). Für das Finale Rendering müssen es dann Erfahrungsgemäss ca. mindestens 15 pro Spur sein damit genug Variationsvielfalt über die Karosserien und die Farbkombinationen....

Bei den PKWs fällt mir eigentlich nur der mit den spitzen, dreieckigen Rücklichtern (zumindest in der Nacht) "negativ" (zu oft) auf.

ciao
Thomas
 

LightWaveGuru

Active member
Bei den PKWs fällt mir eigentlich nur der mit den spitzen, dreieckigen Rücklichtern (zumindest in der Nacht) "negativ" (zu oft) auf.

ciao
Thomas

Ich hätte da auch Boxen nehmen können als Proxys die die gleiche Polygonanzahl hosten in der Szene. Für Zuschauer dieser Tests die diese Hintergründe nicht kennen wirken diese sicher unverständlich. Ich erwarte also auch nicht das man auf Anhieb versteht was ich da zeige. Ich schreibe es zwar immer hin aber um das zu verstehen muss man wenigstens einige Hintergründe solcher BRUTFORCE Szenen kennen. Es geht darum den Speicher zu füllen um drei Dinge zu erfahren.

A) Rendern die Rechner am Rande der Vollauslastung stabil die Szene zuverlässig durch. (= Produktionszeitbasierende Fragestellung)

B) Wieviele Objekte mit einer bestimmten Geometrielast kann ich laden bevor der VRAM voll ist. (Workflow-Optimierung für grösstmögliche Randomisierung solcher Szenen mit vielen Instanzen.)

C) Wenn Fehler auftreten wo treten diese auf und gibt es dafür mir bereits bekannte workflows (um den bug oder sonstiges zu umgehen) oder muss ein neuer Workflow dafür entwickelt werden um das zu fixen.

Ich hoffe diese Ausführungen helfen da etwas weiter, falls nötig.

snip LWGuru
 
Last edited:

LightWaveGuru

Active member
Kan vi hålla oss till svenska tack?
:)
ag kunde ha tagit lådor som proxys som är värd för samma antal polygoner i scenen. För dem som inte känner till dessa bakgrunder verkar dessa tester obegripliga. Jag förväntar mig därför inte att du ska förstå vad jag visar på en gång. Jag skriver alltid ner det, men för att förstå det måste man känna till åtminstone en del av bakgrunden till sådana BRUTFORCE-scener. Det handlar om att fylla minnet med två saker.

A) Kan datorerna vid gränsen till full kapacitet återge scenen på ett tillförlitligt sätt? (= produktionstidsbaserad fråga)

B) Hur många objekt med en viss geometrilast kan jag ladda innan VRAM är fullt. (Optimering av arbetsflödet för största möjliga slumpmässiga fördelning av sådana scener med många instanser).

C) Om fel uppstår, var uppstår de och finns det arbetsflöden som jag redan känner till (för att kringgå felet eller något annat) eller måste jag utveckla ett nytt arbetsflöde för att åtgärda felet.

Jag hoppas att dessa förklaringar hjälper dig lite, om det behövs.

snip LWGuru
 

LightWaveGuru

Active member
in ENGLISH PLEASE!

I could have used boxes as proxies that host the same polygon count in the scene. For viewers of these tests who do not know these backgrounds, they certainly seem incomprehensible. So I don't expect them to understand what I'm showing right away. I always write it down but to understand it you have to know at least some backgrounds of such BRUTFORCE scenes. It is about filling the memory to know three things.

A) Do the computers at the edge of full capacity render the scene reliably. (= production time based question)

B) How many objects with a certain geometry load can I load before the VRAM is full. (Workflow optimization for greatest possible randomization of such scenes with many instances).

C) If errors occur where do they occur and are there workflows already known to me (to work around the bug or something else) or do I have to develop a new workflow to fix this.

I hope these explanations help a little bit, if necessary.

snip LWGuru
 

prometheus

REBORN
1. Optimize the discussion for most effective discussion, that means..do I really have to use german, just to provoke a crazy Swede to state something in swedish, then back again to crazy Germans having fun with translating everything again to swedish, then back to english.
but It there is time, there is always time for fun.

And this is a clown, works in both swedish and english...and you know what, even in German.

🤡
 
1. Optimize the discussion for most effective discussion, that means..do I really have to use german, just to provoke a crazy Swede to state something in swedish, then back again to crazy Germans having fun with translating everything again to swedish, then back to english.

I’m not German.
ciao
Thomas
 

prometheus

REBORN
I’m not German.
ciao
Thomas

You surname and links and deutsch fooled me, but Austrian and deutch...same ol same ol, isn´t it :p not sure how different the language is there, let´s say comparing the differens between Norway and Sweden, erikals and me you know.

So easy to confuse..though I can keep Arnold S on the right side at Austria, Christopher Waltz as well :)
 

prometheus

REBORN
 

LightWaveGuru

Active member
You surname and links and deutsch fooled me, but Austrian and deutch...same ol same ol, isn´t it :p not sure how different the language is there, let´s say comparing the differens between Norway and Sweden, erikals and me you know.

So easy to confuse..though I can keep Arnold S on the right side at Austria, Christopher Waltz as well :)
austrian german has about 6000 different words than standard high german. both understand each other as long as standard high german is spoken. so when thomas starts talking in viennese slang or in austrian german it's like speaking another language. it's the same as with swiss german. these are all very similar dialects.

snip LWGuru
 

prometheus

REBORN
austrian german has about 6000 different words than standard high german. both understand each other as long as standard high german is spoken. so when thomas starts talking in viennese slang or in austrian german it's like speaking another language. it's the same as with swiss german. these are all very similar dialects.

snip LWGuru

Interesting..

It´s of no surprise though, as for Swedish and Norwegian, I think it´s about 80% similarity, but I think Danish Writings and Norwegian is more similar and I think they have it easier to understand the writing, for Swedes though, it´s far easier to understand most Norwegian language than Danish.

Norway has two languages though I think mainly (new Norwegian, and bokmål) and minority languages like samiska, where Sweden in general only have one and samiska as well, maybe gotländska is a bit off though(Gutniska) in our little Island in the baltic sea, though it is very understandable.
But we do indeed have some in common with German, being in the same language family I think.

words in the German language.. I can recognize quite a lot of words to understand vaguely what it is about, it´s the grammar and other words connecting it all that I do not understand.
English was of course standard from grade 2-3 classes when we were schoolkids, and of course I lernt by tv shows as well, entering grade 7-9 and gymnasiums after that, we
could choose German as extra language, I didn´t.

The interest of learning it has become more increasing though over the years, could have used it on my former job where we had some of our biggest product suppliers from germany, but I let my sales collegue handle that since he was almost fluent in speaking Deutch.

Polish is like elfish..
And Elfish is in words we do not speak in here, I also know blenderish, but people do not understand when I talk that, and just call it rubberish.:D
 

LightWaveGuru

Active member
Distances between cars are too uniform for me. There is no red lights and green lights for others. Waves of cars starting engine after green appeared.

See this:

Video from video-recorder installed in vehicle is analyzed by special program, and entire city is build, including other cars..
One could make such for cars only to get vehicle position, velocity, acceleration. Transfer data from real world to simulated world..

Notice how car in game engine is shaking, while driving. Streets are not perfectly flat. Car suspension also plays a role.

Distances between cars are too uniform for me.
There is no red lights and green lights for others.
Waves of cars starting engine after green appeared.


Hi Sensei.

these items are all on my list.

I therefore refer again to the already explained point that at the moment it is not about subtleties but about the final scene fitting into the memory of the graphics cards.

The next step is to install the traffic light system. After that you can work on the distribution of the vehicles along the splines and how to animate them to create these wave-like movements.

However, I have something else to do at the moment.
I will continue with it as soon as I have time.
Probably end of next week.

snip LWGuru
 
Top Bottom