Page 4 of 7 FirstFirst 1 2 3 4 5 6 7 LastLast
Results 91 to 120 of 198
  1. #91
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Boxes:

    Name:  2018 0716_095520.jpg
Views: 730
Size:  112.7 KB

    Name:  2018 0716_100054.jpg
Views: 729
Size:  110.3 KB

    The grim battle continues. I think some of these are perhaps slightly better than previous attempts. I noticed that when ghosting lines towards a vanishing point, my hand tends to veer way off course.

    I have become distracted by Proko's basic intro to figure drawing, but I think this is probably a good thing. Perhaps it is better to first get a general overview before diving into details of anatomy. And thus, a bunch of gesture sketches:

    Name:  2018 0713_104701.jpg
Views: 730
Size:  92.7 KB

    Name:  2018 0716_101958.jpg
Views: 730
Size:  93.6 KB

    Name:  2018 0716_103531.jpg
Views: 730
Size:  92.4 KB

    Name:  2018 0716_115203.jpg
Views: 729
Size:  99.9 KB

    Name:  2018 0716_120249.jpg
Views: 728
Size:  97.3 KB

    Not exactly the most elegant looking things, but according to Proko's video, these are very important to do. They're kind of fun too, so I'll do lots.

    Lastly, a sketch of a handkerchief, done from life:

    Name:  2018 0716_142741.jpg
Views: 727
Size:  202.0 KB


  2. Hide this ad by registering as a member
  3. #92
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Ah, at last CA is back up. Time to catch up with posting stuff, which I'll do over several posts. Some boxes:

    Name:  2018 0717_095621.jpg
Views: 717
Size:  104.4 KB

    Name:  2018 0717_100027.jpg
Views: 716
Size:  96.7 KB

    Gesture sketches:

    Name:  2018 0717_103555.jpg
Views: 715
Size:  99.9 KB

    Name:  2018 0717_104911.jpg
Views: 715
Size:  107.8 KB

    Robo beans:

    Name:  2018 0717_120857.jpg
Views: 714
Size:  124.4 KB

    Name:  2018 0717_122302.jpg
Views: 715
Size:  111.3 KB

    And beginning to explore the head:

    Name:  2018 0717_142408.jpg
Views: 713
Size:  109.7 KB

  4. #93
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    The usual boxes, mostly all wrong:

    Name:  2018 0718_102359.jpg
Views: 712
Size:  90.1 KB

    Name:  2018 0718_102909.jpg
Views: 713
Size:  68.6 KB

    Gestures:

    Name:  2018 0718_110513.jpg
Views: 710
Size:  83.8 KB

    And exploring reference photos via mannequins:

    Name:  2018 0718_124007.jpg
Views: 709
Size:  90.2 KB

  5. #94
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    And more boxes, gestures and mannequins:

    Name:  2018 0719_101612.jpg
Views: 709
Size:  92.3 KB

    Name:  2018 0719_102108.jpg
Views: 711
Size:  92.9 KB

    Name:  2018 0719_103859.jpg
Views: 708
Size:  91.0 KB

    Name:  2018 0719_122920.jpg
Views: 705
Size:  99.5 KB

    Name:  2018 0719_163824.jpg
Views: 705
Size:  91.5 KB

    Yikes, that last one is rather skew.

    Next week, more of the same. Rather a bit of a battle, but I can't say I'm not learning anything. Even with the boxes, which I just don't get, I nevertheless feel I'm beginning to get a sense of things existing in 3D space, as opposed to just being contours to be copied.

  6. #95
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    And the usual - first a set of boxes:

    Name:  2018 0723_094105.jpg
Views: 681
Size:  101.6 KB

    Name:  2018 0723_094651.jpg
Views: 682
Size:  98.0 KB

    As skew as always. At this point, I was just about ready to give up on even trying to get the boxes right, and start to just go through the motions of completing the 250 boxes. But then, standing in the shower, where so many profound discoveries are made, I suddenly realized what the error probably is, that I am making over and over again. Well, I'll have to experiment a bit to see. Will do so tomorrow and see what happens.

    For the rest, more gestures, and analyses of reference photos via mannequins. These are pretty skew too, but I think that is to be expected at this point in my development.

    Name:  2018 0723_100454.jpg
Views: 680
Size:  85.3 KB

    Name:  2018 0723_105332.jpg
Views: 679
Size:  95.4 KB

    Name:  2018 0723_122742.jpg
Views: 682
Size:  112.9 KB

    Name:  2018 0723_124719.jpg
Views: 678
Size:  103.8 KB

    Name:  2018 0723_155252.jpg
Views: 679
Size:  86.1 KB

    Proko advises doing the same figure again, getting it as good as one can, rather than doing lots of different ones, so perhaps I should give these figures another go tomorrow and see what happens.

  7. #96
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Yesterday: as noted here before, I sometimes have little choice but to spend some time in front of a TV. Well, why not put the time to good use? I took along a small sketch pad and scribbled some mannequins, some from imagination, and some by taking a mental snapshot of characters on TV and then trying to recreate them as mannequins. It may well be a good exercise:

    Name:  2018 0724_092815.jpg
Views: 664
Size:  67.6 KB

    Name:  2018 0724_092838.jpg
Views: 662
Size:  74.1 KB

    I played around a bit with the little brain wave I had yesterday about why my boxes never work out, and my experimentation seemed to confirm my thoughts:

    Name:  2018 0724_093647.jpg
Views: 662
Size:  94.6 KB

    I start with a Y-shape as suggested in the box exercise description. The two top arms will now point to vanishing points on a line. And here's the bit I never realized before: the upright arm must be perpendicular to the line between the two top vanishing points. If it isn't, the box will be skew and there will be no way in hell you can create three sets of four parallel lines as the exercise requires. Well. Took me more than two hundred frickin' boxes to reach that little insight.

    Anyway, it turned out that while simple enough in theory, it's not so easy to get it right in practice:

    Name:  2018 0724_092851.jpg
Views: 665
Size:  133.9 KB

    Name:  2018 0724_095642.jpg
Views: 663
Size:  85.0 KB

    At least now I know what to be on the lookout for, and why things go wrong.

    Once again did some gestures. These are rather bad. It wasn't my day for gestures, it seems:

    Name:  2018 0724_105633.jpg
Views: 661
Size:  91.7 KB

    Name:  2018 0724_105646.jpg
Views: 662
Size:  67.9 KB

    And more mannequin versions of reference photos. I redid one of the previous ones; it looks a bit better now:

    Name:  2018 0724_122436.jpg
Views: 659
Size:  109.9 KB

    Name:  2018 0724_170403.jpg
Views: 659
Size:  99.8 KB

    Not sure how many of these one should do, but I think I'll do plenty - they still leave much to be desired, and I think they really help one to think of the body as a 3D shape rather than a flat one.

  8. #97
    chrismartin12 Guest
    beautiful sketches and it inspires me to draw one.

  9. The Following User Says Thank You to chrismartin12 For This Useful Post:


  10. #98
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Quote Originally Posted by chrismartin12 View Post
    beautiful sketches and it inspires me to draw one.
    Thanks. I have much to learn. But a sketchbook thread is a good way to help one keep in the discipline of regular drawing.

    Today's batch:

    Boxes:

    Name:  2018 0725_104219.jpg
Views: 646
Size:  136.4 KB

    Name:  2018 0725_104703.jpg
Views: 644
Size:  134.5 KB

    And mannequinizations of reference photos:

    Name:  2018 0725_111339.jpg
Views: 644
Size:  79.7 KB

    Name:  2018 0725_123615.jpg
Views: 644
Size:  74.9 KB

    Name:  2018 0725_130359.jpg
Views: 644
Size:  75.7 KB

    I shall do plenty more of these, I think...

  11. #99
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Boxes:

    Name:  2018 0726_101601.jpg
Views: 624
Size:  123.1 KB

    "Mannequinalyses" of reference photos:

    Name:  2018 0726_110924.jpg
Views: 625
Size:  102.8 KB

    Name:  2018 0726_125825.jpg
Views: 626
Size:  105.3 KB

    Name:  2018 0726_151902.jpg
Views: 624
Size:  105.8 KB

    And some from imagination. I started with somewhat random "robo beans" and then added limbs in similarly semi-random manner. Rather crude, but better, I think, than I could do even a month ago:

    Name:  2018 0726_144028.jpg
Views: 623
Size:  107.7 KB

  12. #100
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Just for fun - a sketch in ballpoint and colored pencil:

    Name:  2018 0728_135724.jpg
Views: 603
Size:  242.0 KB

  13. #101
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Battled it out with one hell of a headache today, which is why I got little done and the boxes are even worse than usual:

    Name:  2018 0730_114109.jpg
Views: 590
Size:  135.0 KB

    Name:  2018 0730_144923.jpg
Views: 588
Size:  113.3 KB

    Well, that makes 250 of them. It was a frustrating exercise, because while it was great at diagnosing my problems, it seems to have done nothing at all to effect a cure. I learned this much: if I ever want to create a scene with 3 point perspective, I'll construct it the old-fashioned way, with a ruler!

    Tried out Posemaniacs. The muscle anatomy on the figures is ridiculous, and they would actually have been more useful with smooth skin instead. I would definitely not recommend them for learning anatomy, but they do seem to be useful for creating mannequins; I skipped the more distorted poses:

    Name:  2018 0730_152413.jpg
Views: 587
Size:  95.4 KB

  14. #102
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Mannequins constructed from photo in a newspaper, whole waiting for potatoes to boil. It's much more difficult to do from clothed figures!

    Name:  2018 0731_084931.jpg
Views: 578
Size:  69.0 KB

    Done with 250 boxes challenge, on with the 250 cylinders. Here I immediately ran into a snag: I simply cannot make head or tails of the videos that explain it. Particularly the one about drawing circles in perspective; what Unfortunate seems to say is that to draw a perfect square in perspective, simply construct a proper ellipse, and, erm, to draw a perfect ellipse, simply construct a perfect square in perspective first. Catch-22. Or something. I will have to go look around for alternatives. I played around a bit with cylinders in boxes anyway; one sometimes learns more through experience than following explanations.

    Name:  2018 0731_115652.jpg
Views: 578
Size:  75.5 KB

    Name:  2018 0731_121456.jpg
Views: 579
Size:  88.1 KB

    Some gestures:

    Name:  2018 0731_123053.jpg
Views: 577
Size:  79.9 KB

    And more mannequins:

    Name:  2018 0731_125513.jpg
Views: 577
Size:  100.8 KB

  15. #103
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    He is incorrect. You cannot rely on ellipses to get you a square (not calling it perfect, because every square is "perfect"), it's the other way around. Because our tool is linear perspective, all we really can know for sure is where to put lines, the rest (curves) has to be approximated with lines. [Edit: he also relies on minor axes for ellipses, which are not so straightforward and are just a rule of thumb: http://www.conceptart.org/forums/sho...83#post3711583)


    Now the way it works, before you start the drawing, you can decide which one of rectangles drawn in perspective represent the square (works in 1pp and 2pp; 3pp is more complicated); when you do there's a viewpoint (https://www.handprint.com/HP/WCL/IMG/LPR/perspec1g.gif) from which the quadrilateral is the representation of how one of the squares looks like behind the picture plane. It is a single point, and from now on everything must be drawn to relate to this point (in principle). Such a point exists for whichever [edit: it can actually only be so narrow; explained in the next post] rectangle in perspective you will call a square, however the more unrealistic it appears to you, the less convenient the viewpoint, hence try and make it look "right" from where you're looking.

    We also have enough information to draw the farther square (however finding the square on the side of this box would require more work, you cannot just invent it).

    Name:  1.png
Views: 568
Size:  46.2 KB

    Now, what do we know about circles? (1) When a circle is fit into a square it touches the midpoints of each side of the square. (2) In perspective, it is some kind of an ellipse (let's say we know if from experience).

    Brianworx sketchbook

    It follows that we need to fit an ellipse inside the square so that it hits the four points. If you need more help placing the ellipse, you can use Loomis' division method.

    Name:  2.png
Views: 566
Size:  53.8 KB

    Now finish the cylinder with tangents to the circles (ellipses) through the right vanishing point.

    Name:  3.png
Views: 565
Size:  57.0 KB

    To draw more similarly oriented cylinders, you'd need to find proper quadrilaterals for the square, for which we intersect the line of the diagonal of the available square(s) with the perpendicular with the horizon line through the vanishing point used to draw the squares. This gets you a vanishing point for diagonals of each square drawn to the left vanishing point, which you use as follows:

    Name:  4.png
Views: 565
Size:  104.8 KB

    Name:  5.png
Views: 565
Size:  101.4 KB

    For 3pp the process gets more complicated, and is even more impractical. In practice people use minor axis approximation of the orientation of the ellipse, apparently roughly fitting it inside eyeballed squares (which is what gets them the approximation of the angle of the ellipse).
    Last edited by onemax; August 2nd, 2018 at 02:47 AM. Reason: Corrections

  16. The Following 2 Users Say Thank You to onemax For This Useful Post:


  17. #104
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    Actually, one mistake. You can make the square as "narrow" as you like as long as this red point doesn't end up more to the right than the right VP. If it does, there's no corresponding viewpoint from which the quadrilateral is a square. But in any case it starts looking way too narrow much earlier.

    Name:  6.png
Views: 548
Size:  60.7 KB

  18. The Following User Says Thank You to onemax For This Useful Post:


  19. #105
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Quote Originally Posted by onemax View Post
    He is incorrect. You cannot rely on ellipses to get you a square (not calling it perfect, because every square is "perfect"), it's the other way around. Because our tool is linear perspective, all we really can know for sure is where to put lines, the rest (curves) has to be approximated with lines.
    Thanks for the explanation. Yesterday I took a long walk, during which I managed to work out at least the basics of this for myself; based on what you wrote, there is always rather more to it, though I think one can get away with minor inaccuracies if the basics are sound.

    I did indeed also think that in 3 point perspective, this whole thing will get far more complicated, but I have a feeling that one will in practice actually seldom need to use very obvious 3pt. Usually, 2pt will dominate the picture, and one might be able to make small adjustment for the 3rd point by eyeballing. Not too sure though, and I don't want to worry too much about it just yet, otherwise the information becomes a bit overwhelming - one thing at a time! :-)

    Others in this thread highly recommended to me that I should go through that draw-a-box website, and indeed, it really does seem like it contains much that is useful. The artist who runs it seems to know what he's doing too; his art is pretty good. But boy, does he struggle sometimes to explain what he means!

    Anyway, makes me think I should perhaps not follow his procedures and exercises in too rigidly religious a manner. But I will nevertheless take note of them, and spend some time drawing at least some cylinders and ellipses so I get a feel for how they are done. Ellipses are particularly important because they are used so often, and in my own drawing and painting thus far, the thing that very often ruined my picture was getting ellipses all wrong.

  20. #106
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    Quote Originally Posted by brianworx View Post
    though I think one can get away with minor inaccuracies if the basics are sound.
    Yes, because while in principle a picture is drawn for one single point to be viewed from, in practice it's not so critical at all: https://en.wikipedia.org/wiki/Perspe...l)#Limitations As we know a regular picture looks good from many viewpoints, hence something in the picture not drawn exactly in relation to the viewpoint of another element won't necessarily stand out.

    And, really, for actual drawing today, free-hand perspective is the most important to master:

    https://www.artstation.com/artwork/ldVmo
    https://www.artstation.com/artwork/OlmK8
    https://www.artstation.com/artwork/v89od

  21. The Following User Says Thank You to onemax For This Useful Post:


  22. #107
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Quote Originally Posted by onemax View Post
    Yes, because while in principle a picture is drawn for one single point to be viewed from, in practice it's not so critical at all: https://en.wikipedia.org/wiki/Perspe...l)#Limitations As we know a regular picture looks good from many viewpoints, hence something in the picture not drawn exactly in relation to the viewpoint of another element won't necessarily stand out.

    And, really, for actual drawing today, free-hand perspective is the most important to master:

    https://www.artstation.com/artwork/ldVmo
    https://www.artstation.com/artwork/OlmK8
    https://www.artstation.com/artwork/v89od
    Those were all done freehand? They're very good! The draw-a-box exercises are supposed to develop one's ability at it, though I'm not too sure they have actually helped me much to improve, though they sure did diagnose all my errors!

  23. #108
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    Quote Originally Posted by brianworx View Post
    Those were all done freehand? They're very good!
    Well, I mean, I doubt he measured something out. Maybe used a ruler here and there (though it doesn't even look he used that).
    It seems like in this case it's all in the practice over some perspective basics (square multiplication, minor axis approximation for ellipses, etc.). I'm saying that not being good at drawing beautifully in perspective myself but knowing the technicalities well. Doesn't look like it matters in the end whether you know that or not if your goal is just cool drawings. But well, you asked.
    Last edited by onemax; August 1st, 2018 at 08:39 AM.

  24. The Following User Says Thank You to onemax For This Useful Post:


  25. #109
    Join Date
    Jun 2007
    Location
    Toronto, Ontario
    Posts
    3,534
    Thanks
    326
    Thanked 1,329 Times in 1,059 Posts
    Quote Originally Posted by brianworx View Post
    Others in this thread highly recommended to me that I should go through that draw-a-box website, and indeed, it really does seem like it contains much that is useful. The artist who runs it seems to know what he's doing too; his art is pretty good. But boy, does he struggle sometimes to explain what he means!
    Be aware that in the http://www.drawabox.com website, to draw a box is a tool, instead of an end. It is a lot more about body mechanics and muscle memory, than about drawing boxes in perspective. Nowadays, beyond the most simple case, drawing in perspective is an academic discipline, completely replaced by 3D computer skills.
    Grinnikend door het leven...

    Sketchbook Blog

  26. The Following User Says Thank You to eezacque For This Useful Post:


  27. #110
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    And just to clarify it a little further: he's not entirely correct because (1) he relies on minor axes, which aren't reliable http://www.conceptart.org/forums/sho...83#post3711583 and (2) what is a square depends on the viewpoint.


    In his video https://www.youtube.com/watch?v=yFjrSddZiv4 he indeed constructs a square and circle fit inside if for one of orientations, but it's far from the only possible solutions here.


    To see that, we reverse the point called measuring point (mp), used for both measuring slants (explained why here: http://www.conceptart.org/forums/sho...92#post4009692 | how they are used in practice is explained here: http://www.conceptart.org/forums/sho...76#post4050176) and lengths of lines in perspective (which can't be measured directly with a ruler against the drawing). So we extend the line of the diagonal of the square to hit the vanishing line through VP2. Now, since the diagonal is at 45 degrees from the ground, we can find the mp by finding a point where the slant with the horizon line would be 45 degrees.
    Name:  1.png
Views: 503
Size:  72.0 KB
    Now, when a viewpoint is known, the mp is found by taking a compass placing it on VP2 (in this case) and drawing a circle through the viewpoint until it meets the horizon line; the point of intersection is the mp for VP2. Since here we know the mp for VP2 and need the viewpoint we do it backwards (drawing a circle centered at VP2 through the mp to hit the viewing semicircle [which consists of all possible viewpoints for the two vanishing points]) to find the viewpoint.
    Name:  2.png
Views: 502
Size:  77.6 KB
    Since the mp is a point used also for measuring lines in perspective, we may "check" that what he drew is indeed a square for the particular viewpoint. bc' here measures the line bc in perspective, ab needs no special tools to be measured and can be measured directly. We see that ab=bc' and since bc'=bc, ab=bc. So this is a square, indeed, and the ellipse, hitting the necessary midpoints of the square, represent a circle in perspective for the particular viewpoint.
    Name:  3.png
Views: 503
Size:  82.9 KB
    So what he found is a square oriented at angle θ to the picture plane, as shown.
    Name:  4.png
Views: 503
Size:  94.4 KB
    Should the farther line of the square have been chosen differently, repeating the process we would have found a square oriented at a different angle and a different viewpoint, and an ellipse fit inside properly would have represented a circle for that new viewpoint.
    Name:  6.png
Views: 503
Size:  91.1 KB
    Name:  7.png
Views: 503
Size:  93.0 KB
    He's right about that other ellipse not being circle for any viewpoint. If we tried to repeat the process, we would not be able to hit the viewing semicircle if we assumed that the rectangle is a square. You can see the vanishing point for the diagonal is so high above that drawing a line from it to the horizon line at 45 degrees and then using a compass to hit the viewing semicircle would be impossible (hence my correction/rule of thumb in #104).
    Name:  8.png
Views: 502
Size:  48.2 KB

  28. The Following User Says Thank You to onemax For This Useful Post:


  29. #111
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Quote Originally Posted by eezacque View Post
    Be aware that in the http://www.drawabox.com website, to draw a box is a tool, instead of an end. It is a lot more about body mechanics and muscle memory, than about drawing boxes in perspective. Nowadays, beyond the most simple case, drawing in perspective is an academic discipline, completely replaced by 3D computer skills.
    Probably not for traditional painters, though! :-)
    Also, I would think that with all manner of stuff where computers have taken over, it is nevertheless a good idea to learn the old-fashioned ways too, even if only the basics.

  30. #112
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Onemax: At my present level of knowledge, all of that might as well have been written in Klingon. But I appreciate everybody's attempts to help. :-)

  31. #113
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Scribbles after photos in a newspaper:

    Name:  2018 0801_093244.jpg
Views: 485
Size:  145.1 KB

    Name:  2018 0802_085739.jpg
Views: 484
Size:  58.9 KB

    Bunch of cylinders:

    Name:  2018 0801_094456.jpg
Views: 727
Size:  67.0 KB

    Name:  2018 0802_090754.jpg
Views: 484
Size:  79.7 KB

    Found a nice website with 3D models that can be rotated: https://sketchfab.com/
    Drew some rib cages:

    Name:  2018 0802_100306.jpg
Views: 486
Size:  112.9 KB

    And simplified heads from various angles:

    Name:  2018 0802_123811.jpg
Views: 484
Size:  126.9 KB

    More mannequins:

    Name:  2018 0802_104613.jpg
Views: 485
Size:  93.9 KB

    I watched some of Proko's videos on hands, and tried some simplified hands from reference. All pretty awkward, but actually already rather better than anything I managed before watching his videos, which gives me confidence that he knows what he's doing when it comes to explaining to others how to go about doing stuff:

    Name:  2018 0802_143922.jpg
Views: 484
Size:  106.0 KB

    And a bit of informal scribbling just for fun - one should relax a bit now and then:

    Name:  2018 0801_154738.jpg
Views: 482
Size:  431.2 KB

  32. #114
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    The follow-up was just to provide a bit of a proof to whomever finds it doubtful. Sorry again for thread hijacking.

    For this one cylinder, the viewer would have to be very close to the screen looking somewhere over the farther cylinder peaking to the left and down for it to look right. Hence why it looks bad from a more natural position; so it's better to make the front rectangle wider. Here's a slower introduction to this whole viewpoint concept: http://www.mathaware.org/mam/03/essay6.html

    Brianworx sketchbook
    Last edited by onemax; August 2nd, 2018 at 12:03 PM.

  33. The Following User Says Thank You to onemax For This Useful Post:


  34. #115
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    [QUOTE=onemax;4061767]The follow-up was just to provide a bit of a proof to whomever finds it doubtful. Sorry again for thread hijacking. [quote]

    No problem; I appreciate all feedback.

    For this one cylinder, the viewer would have to be very close to the screen looking somewhere over the farther cylinder peaking to the left and down for it to look right. Hence why it looks bad from a more natural position; so it's better to make the front rectangle wider. Here's a slower introduction to this whole viewpoint concept: http://www.mathaware.org/mam/03/essay6.html

    Brianworx sketchbook
    Not sure which one you mean (the top one?). In previous sheets (post #102), I ran into an unexpected problem when constructing the boxes to draw the ellipses in. I started by drawing the rectangle closest to the viewer; I had no idea how much it should diverge from a square, so I guessed - all I do know is that the top and bottom lines must be shorter than the front and back ones, because perspective squashes distances a bit, and it is after all a square. This much I learned from hard experience earlier in this thread!

    But when I then constructed the rest of the box, I saw to my consternation that the more distant rectangle is now way too stretched out, which means my near rectangle wasn't squashed enough. Or so it seemed. So for the last one, I started by first constructing the far rectangle, and making sure its top and bottom lines are shorter than the front and back ones. I then worked "backwards" to the rectangle closest to the viewer, and the way it worked out was simply the way it worked out: a quite squashed up one, and indeed when I then inserted the ellipse, the whole thing looked a bit wrong to me. I can't work out how to fix that.

    I'll go check out the link you refer to...

    EDIT: Okay, I have read through that, and he addresses precisely, exactly, the problem I have been wondering about for ages: how to set up that first line. The article is something of a revelation, in principle. But I am not at all clear how I would apply all that to my cylinder above to fix it.
    Last edited by brianworx; August 3rd, 2018 at 01:54 AM.

  35. #116
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    Yes, the top one.

    Name:  Untitled-1.png
Views: 462
Size:  248.1 KB

    then worked "backwards" to the rectangle closest to the viewer, and the way it worked out was simply the way it worked out: a quite squashed up one, and indeed when I then inserted the ellipse, the whole thing looked a bit wrong to me. I can't work out how to fix that.
    The front one looks bad because the implicit viewpoint for it is as shown (and found as I explained above). And if we find the 60 degree cone of vision for it (https://i.pinimg.com/originals/c9/36...673914fce8.jpg and look up youtube videos on cone of vision), which is area fairly free of distortions, you see your front square/circle is outside of it and therefore distorted. All your drawing should really fit inside this dashed circle. Also, the more spaced your vanishing points are, the bigger the cone of vision, hence the famous rule of thumb "space your vanishing points". Another rule of thumb, if you don't know where explicitly your 60 deg c.o.v. is, is to not draw too far away from an object that looks undistorted; once stuff starts getting distorted you're getting out of the 60 deg c.o.v. So your sheet in principle should look something like this, with vanishing point(s) far outside of it (which is why perspective is pain in the ass, boring and impractical, except in theory):

    Name:  Untitled-2.png
Views: 461
Size:  60.0 KB

    EDIT: Okay, I have read through that, and he addresses precisely, exactly, the problem I have been wondering about for ages: how to set up that first line. The article is something of a revelation, in principle. But I am not at all clear how I would apply all that to my cylinder above to fix it.
    This is for one point perspective, and easier to understand. I just wanted to point out what I meant by viewpoint, and that it is an actual point in front of the picture plane (but which we draw below the horizon usually, as we can't draw it on the air). For 2pp it's on the viewing semicircle. I'll send you some info on 2pp/3pp.
    Last edited by onemax; August 3rd, 2018 at 08:28 AM.

  36. The Following User Says Thank You to onemax For This Useful Post:


  37. #117
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Quote Originally Posted by onemax View Post
    The front one looks bad because the implicit viewpoint for it is as shown (and found as I explained above).
    It is not clear to me how you found the implicit viewpoint, but for the moment, it doesn't matter all that much, because all the detail and theory is probably more than I can take in at this stage anyway. What I do take away from this is that parallel lines should not converge to the vanishing point too quickly, because it will introduce distortions outside of a fairly small part of the picture. In fact, for some kinds of pictures one can dispense with vanishing points altogether, and simply keep parallel lines parallel - I seem to recall having seen such pictures, and they didn't look too bad, but it does depend on the picture.

    You'll be amused to hear that this stuff is now buzzing in my head to the point where it literally gives me sleepless nights. Last night I suddenly awoke, realizing that I have run into a new problem: it seemed to me that a circle seen in perspective cannot possibly be an ellipse, because the perspective "squashes" the far half of the circle smaller, therefore the two halves of the original circle will no longer be symmetrical around the major axis.

    A sketch seemed to clear up this problem: as we tilt the circle, the major axis actually shifts towards the viewer:

    Name:  2018 0804_080039.jpg
Views: 437
Size:  62.3 KB

    The circle in square ABCD touched the sides at points 1, 2, 3 and 4, and its center line ran from 3 to 4. Tilt it, and you get trapezium ABCD with an ellipse inscribed into it, and the ellipse still has to touch the sides at the original points. But if the original center line now becomes the major axis of the ellipse, the two halves are no longer the same size, and thus it cannot possibly be an ellipse.

    Except it is, because as we move the curved lines of the ellipse past points 3 and 4 closer to us, they slightly "bulge out" into the larger half of the trapezium, and create a new major axis. Or at least, so it seems to me - if I am wrong on this point I have a new thing to be hopelessly confused about.

    They don't tell you this stuff in introductory perspective books! I think I know why too: it can easily make the book too long and too complicated, and most of us don't want a treatise on mathematics; we just want to know the rules so our drawings will look right. But it might be useful for perspective books to briefly touch on it, even if only to introduce some rules of thumb like "don't converge your parallel lines too quickly", perhaps with a demonstration sketch to illustrate how distortions will happen, even if the book doesn't explain why they happen.

    Anyway, enough rambling - back to the drawing board. As always, thanks for the input. :-)

  38. The Following User Says Thank You to brianworx For This Useful Post:


  39. #118
    Join Date
    Nov 2013
    Posts
    617
    Thanks
    457
    Thanked 215 Times in 169 Posts
    From Joseph D'Amelio - Perspective Drawing Handbook

    Name:  image.jpg
Views: 432
Size:  313.6 KB

  40. The Following User Says Thank You to onemax For This Useful Post:


  41. #119
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Quote Originally Posted by onemax View Post
    From Joseph D'Amelio - Perspective Drawing Handbook
    Glad to see my reasoning was correct. :-)

  42. #120
    Join Date
    Jun 2018
    Posts
    144
    Thanks
    57
    Thanked 10 Times in 10 Posts
    Had all manner of stuff that had to be done, not least of which is to arrange for having some work done on my teeth, 'cause I'm tired of living in almost constant pain.

    Anyway, tried out the overlapping forms exercise from draw-a-box. As the video notes, it's tricky, and I found it very difficult to imagine ways in which the forms would intersect.

    Name:  2018 0806_102249.jpg
Views: 398
Size:  83.8 KB

    I tried out some 30 second gestures. It was an absolute nightmare; this kind of frantic scribbling goes against my laid back nature. Result: one can hardly recognize the scribbles as stick figures. I'll need a great deal of more practice:

    Name:  2018 0806_103017.jpg
Views: 396
Size:  85.2 KB

    Name:  2018 0806_103635.jpg
Views: 397
Size:  72.1 KB

    Name:  2018 0806_104343.jpg
Views: 397
Size:  65.3 KB

    Name:  2018 0806_105038.jpg
Views: 396
Size:  66.0 KB

    Name:  2018 0806_105631.jpg
Views: 395
Size:  66.4 KB

    And further anatomical bits and pieces. I didn't really have a good day, and somehow everything was just all tentative and awkward:

    Name:  2018 0803_111050.jpg
Views: 396
Size:  72.7 KB

    Name:  2018 0803_133142.jpg
Views: 395
Size:  82.7 KB

    Did these simplified heads from models at sketchfab. I find them extremely difficult, actually even more difficult than it would have been to draw a real head, I think. Not sure why this should be so, but it is what it is:

    Name:  2018 0806_121824.jpg
Views: 395
Size:  71.3 KB

    Awkward hand studies:

    Name:  2018 0806_124535.jpg
Views: 395
Size:  76.8 KB

    I have been neglectful of Proko's advice for fast improvement, which is to rather draw the same thing over and over until it has improved, than to draw different things. I'll try to remember this for future hand studies, and rather do six studies of one hand pose than moving on to different ones.

    Another mannequin:

    Name:  2018 0806_144251.jpg
Views: 394
Size:  87.6 KB

Page 4 of 7 FirstFirst 1 2 3 4 5 6 7 LastLast

Similar Threads

  1. SketchBook: Sketchbook, Sketchbook, Rolly Polly Sketchbook...
    By Teechan in forum Sketchbooks
    Replies: 5
    Last Post: August 17th, 2015, 05:41 PM
  2. Replies: 23
    Last Post: January 19th, 2013, 03:19 PM
  3. SketchBook: *MOD's please move to sketchbook section* Ulvdemon Sketchbook 030805
    By ulvdemon in forum Sketchbooks
    Replies: 11
    Last Post: February 18th, 2011, 12:07 PM
  4. Replies: 12
    Last Post: August 27th, 2009, 11:38 AM
  5. SketchBook: Unknown Epiphany's Sketchbook/brand new sketchbook(Updated February 27th)
    By unknown_epiphany in forum Sketchbooks
    Replies: 195
    Last Post: February 27th, 2007, 10:43 PM

Members who have read this thread: 54

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Designed by The Coldest Water, we build the coldest best water bottles, ice packs and best pillows.