Reply to thread

Message

Hi all,

A few thoughts on how one might tackle this problem in a general way. First, I know absolutely nothing about 3D TVs or the different schemes available. However, if the job is to (as stated in this thread) duplicate the whole GUI and then to scale both left and right sides to be able to place them side by side then here is what I think...

From an MP1 architecture perspective you do not want any of this to be done by the skin engine. The skin engine should, ideally, be completely unaware of what's going on. Having this design/architectural constraint will cause lots of problems to simply go away (like the font problem mentioned in this thread). It is better to handle this type of work at a low level of the architecture - in the case of MP1 the best place is likely at the D3D manipulation layer. All rendering for MP1 is accomplished using the FontEngine.cpp library of functions. All of this code is written in C++. If you're really interested in pulling this off I suggest you become a D3D (and C++) expert at some level. I know enough (learned along the way) about D3D to have made some changes and additions to the FontEngine but tackling something like this is beyond my current skill. You'll likely have to understand how the D3D drawing space is managed and take advantage of hardware acceleration to be able to perform the scaling for you (otherwise performance of MP1 could be miserable).

Sorry I don't have more specifics and that I'm not an expert in D3D to guide you but this, I think, is where you want to focus your attention. Depending on how the video is rendered into the frame at the D3D layer you may also look at how the layer manager works as stated in this thread. The combination of knowing how these two things (at least) should move the idea and an implementation down the road.

Good luck. I'll follow this thread and respond if there is anything more I can add that may help...

<blockquote data-quote="ajp8164" data-source="post: 793460" data-attributes="member: 64375">Hi all,A few thoughts on how one might tackle this problem in a general way.&nbsp; First, I know absolutely nothing about 3D TVs or the different schemes available.&nbsp; However, if the job is to (as stated in this thread) duplicate the whole GUI and then to scale both left and right sides to be able to place them side by side then here is what I think...From an MP1 architecture perspective you do not want any of this to be done by the skin engine.&nbsp; The skin engine should, ideally, be completely unaware of what's going on.&nbsp; Having this design/architectural constraint will cause lots of problems to simply go away (like the font problem mentioned in this thread).&nbsp; It is better to handle this type of work at a low level of the architecture - in the case of MP1 the best place is likely at the D3D manipulation layer.&nbsp; All rendering for MP1 is accomplished using the FontEngine.cpp library of functions.&nbsp; All of this code is written in C++.&nbsp; If you're really interested in pulling this off I suggest you become a D3D (and C++) expert at some level.&nbsp; I know enough (learned along the way) about D3D to have made some changes and additions to the FontEngine but tackling something like this is beyond my current skill.&nbsp; You'll likely have to understand how the D3D drawing space is managed and take advantage of hardware acceleration to be able to perform the scaling for you (otherwise performance of MP1 could be miserable).Sorry I don't have more specifics and that I'm not an expert in D3D to guide you but this, I think, is where you want to focus your attention.&nbsp; Depending on how the video is rendered into the frame at the D3D layer you may also look at how the layer manager works as stated in this thread.&nbsp; The combination of knowing how these two things (at least) should move the idea and an implementation down the road.Good luck.&nbsp; I'll follow this thread and respond if there is anything more I can add that may help...</blockquote>

[QUOTE="ajp8164, post: 793460, member: 64375"] Hi all, A few thoughts on how one might tackle this problem in a general way. First, I know absolutely nothing about 3D TVs or the different schemes available. However, if the job is to (as stated in this thread) duplicate the whole GUI and then to scale both left and right sides to be able to place them side by side then here is what I think... From an MP1 architecture perspective you do not want any of this to be done by the skin engine. The skin engine should, ideally, be completely unaware of what's going on. Having this design/architectural constraint will cause lots of problems to simply go away (like the font problem mentioned in this thread). It is better to handle this type of work at a low level of the architecture - in the case of MP1 the best place is likely at the D3D manipulation layer. All rendering for MP1 is accomplished using the FontEngine.cpp library of functions. All of this code is written in C++. If you're really interested in pulling this off I suggest you become a D3D (and C++) expert at some level. I know enough (learned along the way) about D3D to have made some changes and additions to the FontEngine but tackling something like this is beyond my current skill. You'll likely have to understand how the D3D drawing space is managed and take advantage of hardware acceleration to be able to perform the scaling for you (otherwise performance of MP1 could be miserable). Sorry I don't have more specifics and that I'm not an expert in D3D to guide you but this, I think, is where you want to focus your attention. Depending on how the video is rendered into the frame at the D3D layer you may also look at how the layer manager works as stated in this thread. The combination of knowing how these two things (at least) should move the idea and an implementation down the road. Good luck. I'll follow this thread and respond if there is anything more I can add that may help... [/QUOTE]