Limited Run Breaking Bad WALTER WHITE GLASSES by Magnoli Clothiers

[Westies14]

God, those Wallabees are ugly!

 

[Indy Magnoli]

Thanks guys! We'll start working on v2 right away...

 

[J EM]

The research here is so impressive! Thanks for putting in all that work!

The one thing I think could still be refined is the shape of the lenses. I appreciate that this shot of Walt from the desert is a long lens, pretty straight on, and that makes it a useful reference (especially with a pair of glasses with similar features on hand). At the same time, it's a bit low-resolution. When we're looking to visually dial in millimeters, the blurry edges can fall on either side of the line.

With the curvature, we're getting into an organic shape. A sculptor or 3D modeler would reference as many angles as they could find and test their look against those images. While we may not have any shots as perfectly straight on and flat as that promo image, we have a lot of high-res images which aren't too wide-angle and are pretty front-on. Looking at these, I think the schematics posted so far may come dangerously close to flat/straight walled on the outside - *maybe* the inside too?

This shot might be a 50mm lens - nothing too wide or too close - and he's straight to camera. The glasses are tilted only slightly down, judging by the arm. I think this demonstrates something key to the "feel" of the frames - that there's a subtle curvature to the lens the whole way around. I don't think it ever really flattens out:

Hi. I agree that the top edge of Walt's frame is not at flat as it appears on my attempt at drawing out the shape. Here's a close up.

I did take into consideration the blurred pixels an image gets when up-scaling it. What I did was convert 53x43 mm to pixels, in Illustrator, and re-scaled the image. With consideration of the blurry pixels, I chose the pixels where I thought the lens started, and settled on 150 by 122. 150x122 are not exactly 53x43 when converted to mm. I think 52.917mm x 43.039mm are safe to round to the nearest mm.

The reason the top of the edge appears flatter on my attempt has to do with the frame anchor points, specifically the inner ones (nearest to the nose.) This part of the frame, the inner top corner, gave me the most trouble. I had a hard time drawing that part out and what I posted is the best I've gotten. If I, or whoever attempts at drawing out the shape, can't determine the correct position of the anchors, then the frame shape will not be what we want. The current placement of the inner anchor (at the side, not the corner) lies on half a millimeter. If we manipulate this anchor, possibly place it lower, then this corner of the frame will be curved more and not be as flat – at least that's my theory.


That shot of Walt's face is unfortunately the best Indy Magnoli has in determining the closest shape. Also, the more Walt's head is tilted, the rounder/more curved the frame appears.


I also want to say that looking at the image beiderbeck posted, the top edge of the lens appears flatter.


Also, here's a gif of "proof" of dimensions using the weathered screen used pair (I hope it works because last time I uploaded a gif here, it was a static image)

@SquidMan
You might be interested in this, if you are unconvinced of a few dimensions I provided to Indy Magnoli. Look at the document box on the left.

I want to clarify a few things here.

On top of being blurry, this image adds a black pixel to mostly everything in the frame. That makes a few things appear thicker/wider than they are (it, on top of improper scaling of the image, are why I thought the frame thickness was 1.5mm and depicted it as such in past gifs)
Anyhow, the temple length. I measured from the edge of the hinge piece itself towards the end of the temple tip. Each piece that makes up the hinge is 4.5mm. I think it's safe to say that the temple length is 137mm as it is on my frame.

The hook. In this image, it appears wider and the wire appears thicker. It's approximately .75 on my frame. Also, note the gap between the wire and the hinge piece. The gap I provided to Indy Magnoli (between the hinge edge of the temple and the placement of the wire) was 16.75. Minus 4.5 and we'll have 12.25. I provided what my frame shows, but it's possible that it's placed differently on Walt's frame. Only off by a quarter of a millimeter.

The overall frame width. I did previously mention that I rounded to 140 because that's a standard front frame size, and because that was the length I got when measuring the frame front shot of Walt's face.

 

[Westies14]

Nice! I was actually looking mostly at the outside vertical though, beneath the hinge. It’s certainly subtle, but I think it never quite levels out as much as the current drawing. Thoughts?


Sent from my iPhone using Tapatalk

 

[J EM]

Nice! I was actually looking mostly at the outside vertical though, beneath the hinge. It’s certainly subtle, but I think it never quite levels out as much as the current drawing. Thoughts?

oh. ok. I thought you were noticing the problem I think I notice. lol

What you're noticing can be remedied by lowering the outer-center anchor point - near the hinge. On my image it's placed 15.5mm down- from the top of the lens edge. Anyway, you were correct. The anchor was too high. I'll post a new image later.

I reworked the shape and was pleasantly surprised. even managed to match about half of the frame with a different image of Walt



Edit: @Indy Magnoli, I don't know if this matters, but it's about the bridge thickness. I notice that my aden frame has a bridge that's 2mm wide (only at the center.) It's made from a 1.6 wire that is flattened. From the front view, the bridge on my frame is 2mm wide. From the top view, bridge is flattened to 1mm thick. I think Walt's bridge is the same, but all flattened. And, again, that hinge corner piece ... I'm really sorry I keep posting incorrect measurements. Last one I mentioned about the corner hinge piece (15mm) was also incorrect. I don't know how I messed that one up so much, but it's 17mm and I'm dead sure about it this time.

 

[Westies14]

I don't think what I'm describing would be addressed by lowering the anchor point near the hinge. I may not be describing it well, so I made an image to try and illustrate it a bit more clearly. Again, this front-on, long lens image is useful for many things in reverse-engineering these, working measurements back and forth with standard sizes and that Aden pair with similar features. The shot simply isn't high enough resolution to precisely pattern its lens shape though, and no single image should ever be exclusively relied on to nail down an organic shape like that border. I'm psyched to have all of the details like the hinges and the hook matched to the original, but the lens shape is part of Walt's face - I think getting that right is the most important part.

Looking at the whole collection of reference images, Walt's lenses never seem to flatten out as much as the current template does in places. They always have a subtle curve running. Even our long, flat, front-on shot allows for it - it's just too blurry to use for precise tracing. There's no clear edge to the wire rim. Still, if I were to trace it, it looks like the current template is running flatter than the edges I see. I think even this image supports the "subtle curve throughout" that we see in all the high res shots:



Thoughts?

 

[SquidMan]

Looking at the whole collection of reference images, Walt's lenses never seem to flatten out as much as the current template does in places. They always have a subtle curve running. Even our long, flat, front-on shot allows for it - it's just too blurry to use for precise tracing. There's no clear edge to the wire rim. Still, if I were to trace it, it looks like the current template is running flatter than the edges I see. I think even this image supports the "subtle curve throughout" that we see in all the high res shots:

View attachment 843220

Thoughts?

Agreed, the current schematics of the lens shape is maybe a tad too "boxy" and should be very slightly more curved. The difference is one of milimeters but does have a subtle importance to their accuracy.

 

[J EM]

I don't think what I'm describing would be addressed by lowering the anchor point near the hinge. I may not be describing it well, so I made an image to try and illustrate it a bit more clearly. Again, this front-on, long lens image is useful for many things in reverse-engineering these, working measurements back and forth with standard sizes and that Aden pair with similar features. The shot simply isn't high enough resolution to precisely pattern its lens shape though, and no single image should ever be exclusively relied on to nail down an organic shape like that border. I'm psyched to have all of the details like the hinges and the hook matched to the original, but the lens shape is part of Walt's face - I think getting that right is the most important part.

Looking at the whole collection of reference images, Walt's lenses never seem to flatten out as much as the current template does in places. They always have a subtle curve running. Even our long, flat, front-on shot allows for it - it's just too blurry to use for precise tracing. There's no clear edge to the wire rim. Still, if I were to trace it, it looks like the current template is running flatter than the edges I see. I think even this image supports the "subtle curve throughout" that we see in all the high res shots:

Thoughts?

Lowering the anchor point below the hinge, and then moving the bezier handles accordingly. Prior to lowering the anchor on the most recent gif I posted in this thread, all I did was move the bezier handles. It didn't suffice with the anchor placed 15.5mm downward from the top edge of the lens. I lowered the anchor, moved the bezier handles and got closer to the outline of the frame. This was the best I can do, and to be honest, I think I messed up the top inner corner - or maybe I just prefer what I had on the old version in the previous gif.

Spoiler

(corners are not finished near the lens)

I agree that the shape is the most important. The reason Walt's frames don't ever look flat at the sides is because in most high resolution images of Walt, he has his head tilted down. This reduces the flat appearance of the side of the lens. I showed this with 3 versions of the same frame. The bottom one is tilted about 25 degrees. The middle one would be what the real physical frame would have. If you still think it looks flat, it all depends on what Indy Magnoli does with that anchor just below the hinge. The other things in your image (what's in green) are due to my inability to properly draw the shape. The shape of the bottom inner corner depends on two anchors; the placement of the inner anchor near the bridge, and another anchor along the bottom corner. The frame corners can always be raised if you think the placement is too low. I believe it currently lies at 8.5mm downward from the top lens edge. Placing it 8mm downward looked too high. Placing it at a quarter seems a bit unusual to me.

I don't have 3d modeling software to model the frame. I wish I did. Anyway, consider this. Look at high resolution images of Walt. You'll likely have to rotate them so that the corners are leveled with each other. Or maybe you can just trust your eye. Try determining where the middle of the lens is - the middle of the vertical measurement. Follow this all the way to the lens edge and you'll notice that the lens on Walt's frame "pops out" more a little higher than the middle your eye determined. [strike]This rules out the outer anchor being placed right at the middle of the lens. Right at the middle (of a 43mm vertical lens) would be the only way for the lens to not ever appear flat, but this trick I'm trying to describe to you tells us that the anchor is not right at the middle of the lens. The outer anchor might be placed exactly in the middle between the center point of the lens and the bottom edge of the hinge. Edit: Seems like this may not actually be the case at all. bbl. [/strike] yes, it was the case after all. In order to get it closer the the rounder shape of Walt's, the bezier handles have to be moved by the half millimeter. As always, the bottom inner corner needs work, but this is round enough while still remaining within the blurred pixel of the frame. I understand it may not be best to rely on one low quality image, but knowing the dimensions of the lens, the shape can be achieved if it's worked out half a millimeter by half a millimeter, and focusing on the darker concentration of pixels. This headshot of Walt is still the best photo for this (since no one has offered to 3d model it), unless someone were to post hd screenshots of Walt looking straight at the camera, with very minimal head tilt.

the outer anchor was lowered by 2mm (from 15.5) and the bezier handles were placed within the half millimeter where the shape looks closest.

 

[J EM]

nvm. I attempted to edit the previous post and it double posted instead.

 

[Westies14]

I didn't understand originally that you meant a bezier handle/anchor, as in a vector drawing, when you first suggested that. I thought you were referring to some physical anchor, the hinge or some other engineering aspect of the frames.

I do agree that the vertical center is not the widest point of the lens shape! I'm not clear on why that precludes the outline from having a VERY slight, subtle consistent curve beyond what's there now. In every trace over the one shot we're referencing, I see a drift between the overlay and the blurred edge of the wire rim. That's the issue with this image, above and beyond the idea of using one image exclusively to reference the shape: The shot doesn't have the resolution to see the edge of the frame clearly. Wherever we place the line, we'll see pixels representing the wire and pixels representing his skin on both sides of it. I do realize we're talking about the tiniest of adjustments - I would be very happy with replicas built to your current drawings - but even in this key shot, I'm seeing a more consistent curve than the drawing illustrates.

While other images may have his head or the frames themselves tilted downward by 5 or 9 or 15 degrees, this shot is inherently inconclusive and open to interpretation due to its low resolution. Any resampling. any tiny distortions in the scan or placement of the original film/image could throw the shape off somewhat significantly at this scale. We'd be magnifying and exaggerating an inherent bug.

In addition to vertical tilt, any degree to which Walt's head or the frames are rotated side to side can have the effect of flattening or exaggerating the curve of the side walls. This is why in this shot (and every shot) a trace of one lens won't overlay the other side perfectly when flipped, and why it's important to reference as many images as possible until the frames "feel" right. I don't think they ever feel quite this flat-sided, even in the straight, long image.

If there's something I'm not getting about the way you're saying optometry measurements work (fully plausible!) or anything else that makes this thinking nonsensical for some reason, I apologize for being slow to understand what you're trying to explain to me!

I'm not a skilled 3D modeler myself, but I'll see if anyone I know might be able to attempt creating a model which we could consider from different angles and focal lengths. Thanks again for all of your hard work and info sharing in trying to make these as right as possible!!!

 

[J EM]

Can you take this image and highlight/point out the area of the frame where you mostly notice the consistent curve?



It's not that you don't understand how optometry works, drawing out the shape from an upscaled image might just not be the way to go. I'll draw it out from the original image size, then resize the vector.

 

[Westies14]

I think with the resolution, drawing it at scale or blowing it up first is a six of one/half dozen of the other situation. In that last grab, I think the lack of resolution is actually blocking/sampling pixels in a way which suggests a straight line. In that case, it would be a lack of available pixels/real estate/data with which to illustrate a subtle curve. It’s not that I can demonstrate that it’s consistently curved using that image - I can’t. The resolution of the image isn’t capable of showing us the exact shape of the lenses either way. Whether we’re squinting at a small trace or tracing our best guess at a blurry edge with it blown up, we’ll always be interpreting/guessing because the line isn’t clear.

My best way of showing it visually is up a couple posts - I drew green lines on the edge as I interpret it (best guess) next to the outline as you’ve interpreted it. I think it’s absolutely brilliant how you’ve been able to use it for sizing reference, but I think we need to throw the kitchen sink at it to find the right flow of the form.

We won’t be able to reference anything perfectly. Some shots will have unfortunate angles, steep or minor. Focal length will play tricks on us. Some shots, like our main one, will be low resolution.

To me, taking the reference as a whole, the lens shape is so close in the current draft but just needs to be soooo subtly rounded. I can’t prove it - I just wanted to open it up for discussion. I would love to be proven wrong, and I wouldn’t balk at a pair made to the current drawing by any means. I’ve been hoping for these for a long time and just want them to be as right as humanly possible!

Thanks again.




Sent from my iPhone using Tapatalk

 

[J EM]

I think with the resolution, drawing it at scale or blowing it up first is a six of one/half dozen of the other situation. In that last grab, I think the lack of resolution is actually blocking/sampling pixels in a way which suggests a straight line. In that case, it would be a lack of available pixels/real estate/data with which to illustrate a subtle curve. It’s not that I can demonstrate that it’s consistently curved using that image - I can’t. The resolution of the image isn’t capable of showing us the exact shape of the lenses either way. Whether we’re squinting at a small trace or tracing our best guess at a blurry edge with it blown up, we’ll always be interpreting/guessing because the line isn’t clear.

My best way of showing it visually is up a couple posts - I drew green lines on the edge as I interpret it (best guess) next to the outline as you’ve interpreted it. I think it’s absolutely brilliant how you’ve been able to use it for sizing reference, but I think we need to throw the kitchen sink at it to find the right flow of the form.

We won’t be able to reference anything perfectly. Some shots will have unfortunate angles, steep or minor. Focal length will play tricks on us. Some shots, like our main one, will be low resolution.

To me, taking the reference as a whole, the lens shape is so close in the current draft but just needs to be soooo subtly rounded. I can’t prove it - I just wanted to open it up for discussion. I would love to be proven wrong, and I wouldn’t balk at a pair made to the current drawing by any means. I’ve been hoping for these for a long time and just want them to be as right as humanly possible!

Thanks again.

I started this just a while ago.

I took the BBS1 poster gain - the original uncropped image being at 2700x4307 - rotated it so the frame corners are leveled, and started working directly from it. As you can see, I tried selecting the shape, again, pixel by pixel, closest to the outline of Walt's frame. I settled on 75x61 pixels. When converted to mm, it becomes 26.458x21.519. When resizing the vector to 53mm, I get this

I think I nailed the dimensions again.

I'm not sure how much more round it is, in comparison to the old ones, but we'll see later when I attempt to post a comparison to my old drawing. Does this new shape - which still needs work along the inner curve - look closer to you? I can get closest to the shape of Walt's frame if I decrease the units I'm working with. I was working with half a millimeter increments on the last image I posted, but it seems that's still too large. Working pixel by pixel may be the way to go.

 

[Westies14]

That one “feels” much more like the lens, to me!

I still think it’ll be worth cross-referencing with as many looks at the glasses as possible, seeing if anything jumps out or inspires a change or seems difficult to reconcile when comparing to different angles.

I feel like I may not be fully understanding your process. While this one has a much closer feel, there’s still a lot of guesswork involved if I am reading it correctly. Whether you’re drawing with a pen tool, moving handles in 1mm or 1px increments, aren’t you still just approximating/interpreting by eye where the edge of the wire frame falls? Or did you have a much higher res version of the image, with a sharp edge to the frames? Apologies if I’m not grasping something of the technique.

I think this is a huge inprovement either way! I understand we’re talking about hairs’ breadths of movement along the path. One specific note: the bottom left corner section seems to bulge slightly, pushing out just a little from the flow of the (really nice!) continuous shape around the lens.

Great work!

 

[Westies14]

I see that you did have a larger one - I can’t see it full res on my phone, but it does look like a clearer edge! Still worth looking at them against other pics, but it seems like a much stronger starting point!

 

[Westies14]

Perhaps see what happens when you trace both lenses by your method, flip the left one and see how the paths align? The average of the two could be an even better starting point for refinement.

 

[J EM]

That one “feels” much more like the lens, to me!

I still think it’ll be worth cross-referencing with as many looks at the glasses as possible, seeing if anything jumps out or inspires a change or seems difficult to reconcile when comparing to different angles.

I feel like I may not be fully understanding your process. While this one has a much closer feel, there’s still a lot of guesswork involved if I am reading it correctly. Whether you’re drawing with a pen tool, moving handles in 1mm or 1px increments, aren’t you still just approximating/interpreting by eye where the edge of the wire frame falls? Or did you have a much higher res version of the image, with a sharp edge to the frames? Apologies if I’m not grasping something of the technique.

I think this is a huge inprovement either way! I understand we’re talking about hairs’ breadths of movement along the path. One specific note: the bottom left corner section seems to bulge slightly, pushing out just a little from the flow of the (really nice!) continuous shape around the lens.

Great work!

I think I understand the "boxy" look you and Squidman were noticing. It mostly had to do with that anchor point below the hinge, and even one on the bottom of the lens. They were moved and the result was better.

What I did last time was take that image of Walt and resized it. Despite the blurriness and low quality, I selected the pixels where I thought the outline of the frame would be. I ended up with the right selection (when converting the pixels to mm) but because the image was low quality, that subtle curve on the side appeared to be straighter than it was. I was working off that large image all that time. This time, I didn't resize the image. I did the same thing again, working on a much smaller, sharper image, selected pixels, drew the shape along the more visible pixels, and got the result I needed when resizing the vector - and the subtle curve was preserved. For this latest attempt I did round to the nearest quarter mm to relocate the anchor points. I wasn't that far off to begin with. That's...really my process in drawing out the shape. Decrease the units I'm working with for a more precise shape.

This is the latest attempt.


This was resized from the shape I got in the previous images I posted, after adjusting it to pinpoint that corner that gave me a hard time.

Also, I forgot to mention before. All the thanks should go to Indy Magnoli. All I can do is try to draw the shape so he'll have something to compare against in case his schematic needs to be drawn closer. If his is off, I may be able to describe how and what needs adjustment, because half a millimeter will make a lot of difference; it's easy to get an inaccurate shape.


I'll see what shape that creates. be back later.

 

[Westies14]

Haha I still think the technique is flawed if I’m understanding it... Approximating the edge from a small image takes as much “interpretation” as deciding where to put the points on a blurry enlarged edge - maybe more! There’s no more data in the image at either size, though something about the enlargement felt misleading apparently. It’ll never be the same as confidently tracing the edge of a crisp hi-res image.

BUT it doesn’t really matter! I think you’ve managed a wonderful shape here. I’d be so happy with that latest drawing in glasses form!!!

 

[Westies14]

Going through the series looking for any interesting reference and found some images I thought worth sharing! It can be difficult to call a shot "straight on" due to the frame's warp - sometimes we're looking straight down the barrel of one arm while the other is clearly at an angle. Still, there's a lot to consider in Walt's opening confessional video as he's mostly looking straight into the lens for a long time, and you can get a look at the lens shape at many different vertical angles in a row. Here are a few where the arms read mostly straight, particularly that right (his left eye) lens we've been tracing:



And this one is actually angled UP slightly more than usual, as we can see the bottom of the arm on the right - the right side of that same lens still looks quite rounded:



This is when we return to the opening scene at the end of the pilot:



The shape will change by the tilt or turn of Walt's head, by different focal lengths on the camera's lens, by the distance between the frames and the camera. One thing I'm not sure we considered (or even need to consider) when looking at the main poster image is compensation for the inherent tilt of the frames, the angle at which the arms sit in relation to the front-facing profile. How steep is that angle?

I'm not sure what program you're working in, but let me know if you think this is something you (or Indy Magnoli?) could use if I did this in illusrtator:

If potentially useful/usable, I'd be very willing to go through the series, promo images and the auction pair's images to grab as many clean, useful, straight-on looks at the lens shape as I can find. I'd trace both right and left lenses, giving us dozens of independently-created profiles for each side, flip the right lens outlines to match the left, and finally find the average/median path of all the profiles overlaid. I think this would give us our best possible chance at resolving their subtle and unique shape without having a genuine pair on hand. J EM or Indy could take this profile and apply the measurements J EM's determined from research of optometry standards and his Aden glasses' features before locking in a final design for manufacturing.

If this would be welcome and useful, just say the word! If not, I'll be quite happy buying a few pairs that look like J EM's last drawing.

So excited for this. I hope I'm not dragging down anyone else's enthusiasm (especially you, J EM - I appreciate your work here more than I can express!). Can't wait to see these become a reality!!!

 

[Indy Magnoli]

While you guys sort this out, here are some options for the acetate tips and the frame color:

Let me know what you think...

Kind regards,
Magnoli