Jump to content

A Rainy day in Paris


LightinaBox

Recommended Posts

I took a look at the other models of the eiffel tower, and they do not have these metal flower-like in the center, there is only a square, which does not correspond to the eiffel tower in the real world

I'm trying to get as close to the tower as possible in the real world

about the geometry, I'm trying to reduce it, because when I start to clone the panels, the amount of objects can exceed 10 thousand

 

I can not just give a connect + delete and make everything a mesh only because I need to keep everything procedural, of course if I did, the model would be lighter to work with

but it would not be nice if I bought a model and it came all sewn

 

Thanks for the help, I'll see what I do here to fix this

Link to post
  • Replies 60
  • Created
  • Last Reply

 

7 hours ago, LightinaBox said:

Update:

No triangles :)

qud4.thumb.PNG.a6daa4d7768674d05dca991a3f0529b8.PNGQUD3.thumb.PNG.547ac953d3a99ea3016043253577f27e.PNGqud2.thumb.PNG.c9fb6011283a8f83958bb6ccf308f8fe.PNGQUD1.thumb.PNG.2a56747be79a7874dc277d494ec7ed10.PNG

 

Out of curiosity is there a reason why you feel like you need to keep it productural ? also being honest the topology is getting pretty bad here,  it's all well and good to have no n gons or triangles but not if this mesh is the result, not to be discouraging but i feel this maybe a little out of your modelling range skill wise, atleast for the moment, of course this is only my opinion. 

Link to post
Well this is exactly what I meant when I asked why he felt the need to make it all quads. The screenshots demonstrate the problem when people are given the "rule" to model all quads just on "principle", without understanding why or when using quads is necessary, or understanding different modelling techniques.
 
Although I hesitate to say this, because I do not consider myself an expert or very experienced modeller at all, (so if I am somehow wrong I will accept it), but I think that if he's not subdividing or deforming the mesh, but rather just using spline extrusions and booleans and then laboriously trying to cut all the triangulated flat surfaces into little quads, it just makes no sense and is a waste of time. It makes no difference to the quality of the result with the current method he's using for modelling and is actually creating a lot of unnecessary and messed up geometry in the process.

 

I never meant to say that it doesn't matter if you model with quads or tris or n-gons, but I was looking at the context of the work he was showing and the modelling approach he's using.
 
Either model those curvy and star-shaped bits properly using SDS techniques, or don't bother with the quads (or just model it as quads to begin with). I presume most of the other elements can be just straight box/poly modelled since they will be mostly simple rectangular shapes.
Link to post

i have to say i'm with king of snake on this one. all those star shapes and stuff would be most efficiently modelled by drawing out splines and extruding them, leaving the caps ngons. the model just gets unnecessarily heavy, slows down the viewport and slows down the rendering. i only see disadvantages here in cutting everything to quads, makes no sense at all. in cg the fastest way to get you there is always the right one, if it doesn't compromise the quality of the end result. and by end result i don't mean a wireframe render. if you wanna show off how good of a modeller you are that's another thing, but if you want to model it as quickly and efficient as possible and if you want to keep the poly count low then make those shapes ngons wherever you can.

Link to post
1 hour ago, King of Snake said:

Well this is exactly what I meant when I asked why he felt the need to make it "all quads" . The screenshots demonstrate the problem when people are given the "rule" to model all quads just on "principle", without understanding why or when using quads is necessary, or understanding different modelling techniques.

OP is selling the model, which means it cannot have any ngons, because a) different apps resolve ngons differently, so you cannot guarantee consistency across platforms and b) most 3D model sites simply won't allow it. I would add c) it's just not professional modelling to leave ngons in a commercial model, but that is just my opinion. But, and especially with a model as oft-produced and sold as this one, there will always be someone else who has done it properly, and they will be able to charge more cash for their superior models, and other modellers will respect their work, and they will build reputation.

 

It's also worth pointing out that while ngons might not crash the program any more, they do still prevent the polygon tools working predictably, which is surely yet another reason to avoid them ?

 

But I agree that what OP has done here for a quads solution is not the answer either. This has, as you pointed out, added a ton of geometry, and it's still horrible (and now unnecessarily dense and irregular), despite being quads. 

 

1 hour ago, King of Snake said:

Either model those curvy and star-shaped bits properly using SDS techniques, or don't bother with the quads (or just model it as quads to begin with).

This is good advice - most of the horribleness of the current topology has happened because OP tried to fix spline-based objects (compounded by non-optimal spline interpolation settings) rather than model them from polygons in the first place, which (even with some triangles) would have produced the best / most commercially appropriate results whether using subdivision surfaces, or with bevels, or just as they are. 

 

Quads-only would have given additional flexibility in developing the model into a high poly version if that was ever needed later either by OP or by people who had bought the model, and would look neater, more professional and more attractive to anyone browsing wireframes. Admittedly, some people really don't care about this when buying models, but let's put it this way; nobody is going to not buy your models because they were modelled nicely, but some people might not buy your models if they look like cr@p in the wireframe. So that's my arguments for all quads and decent edge flow, rather than it being for only modelling purist reasons.

 

CBR

Link to post
1 hour ago, King of Snake said:
Well this is exactly what I meant when I asked why he felt the need to make it all quads. The screenshots demonstrate the problem when people are given the "rule" to model all quads just on "principle", without understanding why or when using quads is necessary, or understanding different modelling techniques.
 
Although I hesitate to say this, because I do not consider myself an expert or very experienced modeller at all, (so if I am somehow wrong I will accept it), but I think that if he's not subdividing or deforming the mesh, but rather just using spline extrusions and booleans and then laboriously trying to cut all the triangulated flat surfaces into little quads, it just makes no sense and is a waste of time. It makes no difference to the quality of the result with the current method he's using for modelling and is actually creating a lot of unnecessary and messed up geometry in the process.

 

I never meant to say that it doesn't matter if you model with quads or tris or n-gons, but I was looking at the context of the work he was showing and the modelling approach he's using.
 
Either model those curvy and star-shaped bits properly using SDS techniques, or don't bother with the quads (or just model it as quads to begin with). I presume most of the other elements can be just straight box/poly modelled since they will be mostly simple rectangular shapes.

Well i think it demonstrates more a lack of knowledge/technique/skill, this is clearly shown by the way he's constructed  the model in the first place and by principle i mean by way of doing it properly, as everfresh pointed out fastest way is better, as long as it doesn't compromise quality, and if this was a commercial piece and under deadlines then fair enough, make it with drawn splines and extrudes, as it would be faster, if it's seen from a distance, yea even better i wouldn't even bother bevelling the edges, you could make an argument for that in a commercial environment, or it was a quick concept model to realise an idea have ngons and triangles all day long, as that is not the final model, and a fully quad based higher poly model will be created later on based on the original concept.

 

But when you're trying to make a profit from a model, that's in competition from other models that are either equal or of better quality, which he specifically stated as wanting the model to attain a level of detail surpassing the other models on the website then you're not gonna have the flexibility to achieve that with just extruded splines, i agree trying to make it quads afterwards is not the way to go, if that's the aim do that from the beginning, but using boole's the results are inconsistent and create equally horrendous topology. As cerbera pointed out many websites ngons are a strict no, as in they will not accept the model if it contains them, triangles not so much. really it all comes to down to personal preference, i've seen litterature sold and bought and distributed in the thousands by places like cgsociety that contain absolutely horrendous examples of topology, modelling and just over all laziness that would make Cerbera wet his big girl knickers! and that is from some renowned industry artists and these are being sold for like 60 -170 dollars and are currently sold out images below are from

these books 

http://www.ballisticpublishing.com/books/dartiste/character_modeling/

http://www.ballisticpublishing.com/books/essence/face/

 

https://gyazo.com/7b5b8ff8c311e0333a7488f49159e250

 

https://gyazo.com/d5e74c449eef1422fbb28e614edfa60e

 

Now although I've been assured theres is good content in these book's (i don't own them personally) the fact they contain stuff like that is unforgivable imo.

So it would seem a lot of people would agree with you that 100 percent quads are not necessary, nor is getting rid off all the ngons or triangles, but personally to me and Cerbera atleast, just because you don't need to do something to get the job done, doesn't necessarily mean you shouldn't and personally i'd rather be called out for being a quad Wh0re and overly obsessive with meshes, than put my name to anything resembling the examples i posted here.

Link to post

Here's a practical example that demonstrates the sort of topology you could have used on one of those floral plates. This version I made using regular poly modelling, symmetry, and being very minimal on the polygons, but it is still all quads (or will be when symmetries are combined and centerline edges dissolved).

It took 5 minutes, which is arguably faster than making the splines in the first place !

 

starquads.thumb.jpg.71c42a9feb9c71f171c2a2de9c8322ba.jpg

 

I appreciate that it might take someone with less experience more time to do it this way, but if you train yourself from the ground-up to model things in quads then you become as fast at doing that as lesser skilled people are at doing it the lazier spline-based ways.

 

CBR

 

 

Link to post
1 hour ago, VECTOR said:

Well i think it demonstrates more a lack of knowledge/technique/skill, this is clearly shown by the way he's constructed  the model in the first place and by principle i mean by way of doing it properly, as everfresh pointed out fastest way is better, as long as it doesn't compromise quality, and if this was a commercial piece and under deadlines then fair enough, make it with drawn splines and extrudes, as it would be faster, if it's seen from a distance, yea even better i wouldn't even bother bevelling the edges, you could make an argument for that in a commercial environment, or it was a quick concept model to realise an idea have ngons and triangles all day long, as that is not the final model, and a fully quad based higher poly model will be created later on based on the original concept.

 

But when you're trying to make a profit from a model, that's in competition from other models that are either equal or of better quality, which he specifically stated as wanting the model to attain a level of detail surpassing the other models on the website then you're not gonna have the flexibility to achieve that with just extruded splines, i agree trying to make it quads afterwards is not the way to go, if that's the aim do that from the beginning, but using boole's the results are inconsistent and create equally horrendous topology. As cerbera pointed out many websites ngons are a strict no, as in they will not accept the model if it contains them, triangles not so much. really it all comes to down to personal preference,

 

i've seen litterature sold and bought and distributed in the thousands by places like cgsociety that contain absolutely horrendous examples of topology, modelling and just over all laziness that would make Cerbera wet his big girl knickers! and that is from some renowned industry artists and these are being sold for like 60 -170 dollars and are currently sold out images below are from

these books 

 

Now although I've been assured theres is good content in these book's (i don't own them personally) the fact they contain stuff like that is unforgivable imo.

 

So it would seem a lot of people would agree with you that 100 percent quads are not necessary, nor is getting rid off all the ngons or triangles, but personally to me and Cerbera atleast, just because you don't need to do something to get the job done, doesn't necessarily mean you shouldn't and personally i'd rather be called out for being a quad Wh0re and overly obsessive with meshes, than put my name to anything resembling the examples i posted here.

 

Hey I'm not disagreeing with any of these points really. As I said I my comment was in the context of the techniques he was showing, his specific statement that the model would not be seen up close as well as his comment that "the most important thing" was to get rid of n-gons. To me it shows a mistaken priority and a lack of understanding of why people say you should model all quads or have good topology.
He only mentioned "maybe selling it on turbosquid" as an aside at first and the comment about making it more detailed than other models was made after my initial comment.
Of course I don't disagree that having good topology is something to strive for, quite the opposite! But as you say yourself just making something all quad is not the same as making a good model and I see quite often beginners (even more beginner than me ;)) that seem to have read somewhere that n-gons and triangles are bad but do not know why. That's why I'm trying to say "all quad" should not be the goal or principle in itself. It should be a natural element of using good technique.

 

Quote

It's also worth pointing out that while ngons might not crash the program any more, they do still prevent the polygon tools working predictably, which is surely yet another reason to avoid them ?

 

Well yes of course. But again I my comment was in the context of what was shown. If you cap off some irregular object (like an extruded spline) with an n-gon and you're not going to later have to do anything with that object, in that case the n-gon doesn't really matter. You can even still bevel the edges quite well. That doesn't mean I'm advocating leaving random n-gons in your mesh.

Of course I'll believe it that sites like Turbosquid won't accept n-gons. Just saying but there are other issues with this mesh that seem to be more pressing.

And to the OP: I'm not trying to put down your work (or hijack your topic, sorry about that), but I do think it would help to read up on some basic modelling best-practices. Cerbera and Vector here are this forum's resident modelling guru's so they know what they are talking about. :lol:

Link to post
3 minutes ago, King of Snake said:

 

Hey I'm not disagreeing with any of these points really. As I said I my comment was the context of the techniques he was showing, his specific statement that the model would not be seen up close as well as his comment that "the most important thing" was to get rid of n-gons. To me it shows a mistaken priority and a lack of understanding of why people say you should model all quads or have good topology.
He only mentioned "maybe selling it on turbosquid" as an aside at first and the comment about making it more detailed than other models was made after my initial comment.
Of course I don't disagree that having good topology is something to strive for, quite the opposite! But as you say yourself just making something all quad is not the same as making a good model and I see quite often beginners (even more beginner than me ;)) that seem to have read somewhere that n-gons and triangles are bad but do not know why. That's why I'm trying to say "all quad" should not be the goal or principle in itself. It should be a natural element of using good technique.

 

 

Well yes of course. But again I my comment was in the context of what was shown. If you cap off some irregular object (like an extruded spline) with an n-gon and you're not going to later have to do anything with that object, in that case the n-gon doesn't matter. You can even still bevel the edges just fine. That doesn't mean I'm advocating leaving random n-gons in your mesh.

Of course I'll believe it that sites like Turbosquid won't accept n-gons. Just saying but there are other issues with this mesh that seem to be more pressing.

And to the OP: I'm not trying to put down your work man, but I do think it would help to read up on some basic modelling best-practices. Cerbera and Vector here are this forum's resident modelling guru's so they know what they are talking about.

Not disagreeing with you're points either, as they are valid in a sense that caping a cylinder with an n-gon isn't going to break anything, just that i wouldn't do it ;) , just looking back at some of my older work, i thought that was good at the time, i bow my head in shame at the state of it, in comparison to what im able to do now, i think we can agree that the best way to approach a model, is one in which will be the easiest for you to work with, and get completed with the least amount of hiccups along the way, i remember being dubious of sud b modelling to begin with now it's the only way i model :D 

 

 

Link to post
23 minutes ago, Cerbera said:

Here's a practical example that demonstrates the sort of topology you could have used on one of those floral plates. This version I made using regular poly modelling, symmetry, and being very minimal on the polygons, but it is still all quads (or will be when symmetries are combined and centerline edges dissolved).

It took 5 minutes, which is arguably faster than making the splines in the first place !

 

starquads.thumb.jpg.71c42a9feb9c71f171c2a2de9c8322ba.jpg

 

I appreciate that it might take someone with less experience more time to do it this way, but if you train yourself from the ground-up to model things in quads then you become as fast at doing that as lesser skilled people are at doing it the lazier spline-based ways.

 

CBR

 

 

1.30 min :P ... i agree with everything you said, but there are cases where an extruded spline is way faster, no matter how skilled you are in poly modelling. of course i'm just arguing speed here, not beauty ;) btw, poly modelling something that takes most regular humanoids like myself at least 20 minutes, i'm in awe that you modelled it in 5 minutes o_O ... also it's not about being lazy, sometimes neither budget or deadline allow things to be properly modelled, so it's always good to have different tools in the pocket to skin the cat.

Screen Shot 2017-12-28 at 17.10.12.png

Link to post
6 minutes ago, everfresh said:

1.30 min :P ... i agree with everything you said, but there are cases where an extruded spline is way faster, no matter how skilled you are in poly modelling. of course i'm just arguing speed here, not beauty ;) btw, poly modelling something that takes most regular humanoids like myself at least 20 minutes, i'm in awe that you modelled it in 5 minutes o_O ... also it's not about being lazy, sometimes neither budget or deadline allow things to be properly modelled, so it's always good to have different tools in the pocket to skin the cat.

Screen Shot 2017-12-28 at 17.10.12.png

i reckon i could do it in 3 ;)

Link to post

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...