Pushing AFPD to the limit
09-11-2009, 07:26 AM
#1
Member
Thread Starter
Join Date: Jun 2008
Location: Coronel SuarezPcia Buenos Aires , ARGENTINA
Posts: 93
Likes: 0
Received 0 Likes
on
0 Posts
Pushing AFPD to the limit
Fine detail in an AFPD model has a cost. The count of polygons (faces in Metasequoia) and vertices increases rapidly. If you want wheels that look round or curves that are not angular then you must pay this price. Every program has limits on computer memory and the number of components it can handle. So what are the limits in AFPD?
Here is a test model, shown below, which allowed the number of polygons to be increased until the program, AFPD, broke. The model is based on a single component, one engine pod from an A10 model, also shown below. The model has 45 copies plus a fraction of this component. Each engine pod has 6406 vertices and 10149 polygons.
The program, AFPD, has two parts as far as the user is concerned. The first part is where the model is loaded; you can see the revolving model. This also checks the graphics components of the model defined by the TMG file. If the model does not display correctly at this stage, there is an error in the TMG file. The second part is where the simulation occurs; where you fly your model in the selected scenery. If the model disintegrates or bounces like a ball at this stage, there is an error in the TMD file which sets all the physical parameters of the model.
So where did AFPD break? The limit is in the first part, the loading of the model. The second stage, the simulation itself, is much more generous.
The test model below has:
Vertices 326317
Polygons (Faces) 495215
An increase of 4 polygons over this figure for a single model causes the first part of AFPD to crash.
The first part of AFPD will accept twice this quantity of polygons using Aircraft I and Aircraft II. but a single model is limited to this count of polygons.
The interesting thing is the ability of the second part of AFPD to simulate both Aircraft I and Aircraft II with this huge model. In the case of the computer being used, without any apparent reduction in performance or display frame rate. This computer has an Intel 2 Quad cpu with 3.25 Gb of RAM and a Nvidia GeForce 9800 GTX graphics card.
To put all this in perspective, AFPD was intended for the training of RC pilots. A typical model for this purpose might be the HobbiStar60 model, a single engined high wing aircraft with a fixed tricycle undercarriage with a wingspan of 1.79m. This model has 47769 vertices and 57603 polygons. Roughly 10% of the limit.
The conclusion must be that the limits of AFPD are very generous for its intended purpose.
Of course, when I come along with my super-detailed life size models, things get a little different. I estimate that the model currently being built will require of the order of 400,000 polygons. So I should be alright, just.
Please, IPACS, give me a little more. Give me up to 750,000 polygons. Then I shall be happy; at least for a time.
All the best
Mike
Here is a test model, shown below, which allowed the number of polygons to be increased until the program, AFPD, broke. The model is based on a single component, one engine pod from an A10 model, also shown below. The model has 45 copies plus a fraction of this component. Each engine pod has 6406 vertices and 10149 polygons.
The program, AFPD, has two parts as far as the user is concerned. The first part is where the model is loaded; you can see the revolving model. This also checks the graphics components of the model defined by the TMG file. If the model does not display correctly at this stage, there is an error in the TMG file. The second part is where the simulation occurs; where you fly your model in the selected scenery. If the model disintegrates or bounces like a ball at this stage, there is an error in the TMD file which sets all the physical parameters of the model.
So where did AFPD break? The limit is in the first part, the loading of the model. The second stage, the simulation itself, is much more generous.
The test model below has:
Vertices 326317
Polygons (Faces) 495215
An increase of 4 polygons over this figure for a single model causes the first part of AFPD to crash.
The first part of AFPD will accept twice this quantity of polygons using Aircraft I and Aircraft II. but a single model is limited to this count of polygons.
The interesting thing is the ability of the second part of AFPD to simulate both Aircraft I and Aircraft II with this huge model. In the case of the computer being used, without any apparent reduction in performance or display frame rate. This computer has an Intel 2 Quad cpu with 3.25 Gb of RAM and a Nvidia GeForce 9800 GTX graphics card.
To put all this in perspective, AFPD was intended for the training of RC pilots. A typical model for this purpose might be the HobbiStar60 model, a single engined high wing aircraft with a fixed tricycle undercarriage with a wingspan of 1.79m. This model has 47769 vertices and 57603 polygons. Roughly 10% of the limit.
The conclusion must be that the limits of AFPD are very generous for its intended purpose.
Of course, when I come along with my super-detailed life size models, things get a little different. I estimate that the model currently being built will require of the order of 400,000 polygons. So I should be alright, just.
Please, IPACS, give me a little more. Give me up to 750,000 polygons. Then I shall be happy; at least for a time.
All the best
Mike
