Aspect ratio of a photo?

[Myra Wildmist]

I'm just starting out, playing with photos, so sorry if this is a dumb question. I read this help entry, but I don't think it addressed my question: http://wiki.secondlife.com/wiki/Texture_aspect_ratios

When I take a photo or get one from someone else, is there a way to know what the proper aspect ratio (e.g. 16:9, 4:3, 16:10) is, without having to go through trial and error?

Maybe a related question. I don't see anything about screen resolution or aspect ratios in Viewer 2. Does a photo I take in sl inherit these properties from my computer's graphic settings?

Thanks for any help.

[Peggy Paperdoll]

Aspect ratio is the wide vs the hieght and it's fixed ratio.  1 : 1 is both sides equal to each other.  1 : 2 is the height is two times the width........and so on.  You're computer monitor has is fixed aspect ratio (typically 16 : 9 for HD or widescreen and 4 : 3 for "standard" on older monitors).  What that means for photos you take in SL is that if you take a "full screen" snapshot the image will be very close to your monitor's aspect ratio.  The problem arrises when you want to upload that snapshot back into SL.  Second Life only allows images (also called textures) at aspect ratios that are at powers of two..........and 16 : 9 and 4 :3 will never be at a power of two.  Each dimension of a texture or image has to be at one of the powers of two that the SL software allows (those are 8, 16, 32, 64, 128, 256, 512, and 1024).  To avoid distortion and retain the best image quality you should use an image editor to resize to the powers of two before uploading to SL.  The SL software will resize for you if you try to upload your image without resizing but it's pretty poor and a little unpredictable how how it will resize for you.

 

Depending now how your resize you image you can either upload without any distortion or resize to have as little as possible distortion (then placing your image on a prim cube and sizing the cube to get of the distortion).

 

Hope that wasn't too confusing.

[Zena Juran]

You can select preset and custom height and width sizes in pixels of a picture you take in SL using the Snapshot Preview box. but nothing in the viewer that tells you what the aspect ratio is.

If you receive a photo from someone only they can tell you what the pixel size was that they used in taking the pic.

Determining aspect ratio is up to you.

 

These references may help you...

1. Limits

2. Aspect Ratios

[Rolig Loon]

---

Myra Wildmist wrote:

When I take a photo or get one from someone else, is there a way to know what the proper aspect ratio (e.g. 16:9, 4:3, 16:10) is, without having to go through trial and error?

---

If you have the texture in your invenory, double click to open it.  The pixel dimensions should be displayed in the lower left corner.  From those, you ought to be able to calculate the aspect ratio easily. 

---

Myra Wildmist wrote:

Maybe a related question. I don't see anything about screen resolution or aspect ratios in Viewer 2. Does a photo I take in sl inherit these properties from my computer's graphic settings?

---

When you take a photo in SL, you can change its default dimensions in the Snapshot window that opens. (If it doesn't open a window, type CTRL + Shift + S when you want to take a photo.)  Those dimensions determine what its aspect ratio is.  You can change the default resolution for photos in your Advanced (CTRL + D) or Develop (CTRL + Q) menu.  Look for the parameter labelled High Resolution Photographs.  (The specific spot and name vary from one viewer to another, but the function is there in all of them.) 

The actual dimensions and resolution of any texture will of course always depend on the graphics card and monitor being used to view it. 

[Myra Wildmist]

That all seems a little odd. Maybe the aspect ratio isn't important to a photo. I think I can see that, but I would think you would want to maintain the proper dimensions when a photo is uploaded to sl. Having to upload a photo in sizes that are powers of 2 seems very limiting and not very user friendly.

Is there a technical reason for this?

[Rolig Loon]

---

Myra Wildmist wrote:

That all seems a little odd. Maybe the aspect ratio isn't important to a photo. I think I can see that, but I would think you would want to maintain the proper dimensions when a photo is uploaded to sl. Having to upload a photo in sizes that are powers of 2 seems very limiting and not very user friendly.

Is there a technical reason for this?

---

The business about powers of 2 is important.  It applies to any textures that you upload to SL, regardless of how they were created.  When you take a photo, it is saved on your hard drive with whatever dimensions you specified when you took it (see my previous note).  You'll find that it's automatically saved with power of 2 dimensions, though, even if that means SL has added some white space around the edges.  You shouldn't need to worry when you upload it to SL, therefore, unless you have done some cropping and resizing by post-processing in Photoshop/GIMP/Paintshop Pro. 

The technical reason for all of this has to do with the challenge of forcing each user's graphics card to render a zillion textures with a zillion odd dimensions.  Limiting the possibilities saves a lot of work.

[Peggy Paperdoll]

Here is a very simple, no sweat way to get your texture uploaded to a power of two.

 

Say your image is saved to your hard at a 16 : 9 aspect ratio.  That is very close to a 2 : 1 aspect ratio (twice as wide as it is tall).  Using an image editor of your choice (Photoshop, Paintshop Pro, GIMP, even Paint.NET) open a new "image" (a canvas) at 2048 pixels wide by 1024 pixels tall..........you'll need to resize it before uploading but the point is to get your image at some dimension that is at a power of two.  Then open your image that you saved to your hard drive as a layer.  The will place your desired image on top of the background on the canvas your just opened.  Depending on your save options you chose when you saved the snapshot to your hard drive, the image will either fill the canvas completely with the edges off the canvas (but still there.......just off the canvas) or it will not fill the canvas completely with a "border around it (your canvas filling to space around the image).  If the image is larger than your canvas you can move it around to include everything you want in the final image and exclude anything not important to the picture.  You can then flatten your image to complete the "crop" or not.....your choice.  Next resize your image to 1024 by 512 (keep the maintain aspect ratio box or option checked and you'll only need to adjust one dimension.........the other will change to keep the aspect ratio the same).  If you want a smaller image to upload then resize to 512 by 256.........just use powers of two is the only important thing to remember.  Then save to an accepted format for upload (TGA, PNG, BMP or JPEG.......JPEG is a lossy format and will affect the quality). 

 

Done, you're finished and ready for upload.  No distortion.

[Myra Wildmist]

ummm but there is. 16:9 is not 2:1. so there's some loss there. sorry, did I miss something?

[Peggy Paperdoll]

I think I worded it a little awkward............no, unless you count the edges being cut off if your saved to hard drive image is larger than your canvas size.  The 16 : 9 aspect ratio can never be changed to become a power of two.  I suggested a canvas size of 2048 by 1024 because that would most likely be wider in pixels than your monitor's width resolution.......if your monitor's resolution in width is larger then there's two ways to deal with that.  First open a canvas at some power of 2 that is larger than your monitors width resolution.........say 4096 (the next power of two larger than 2048).  Make the hieght 1/2 of the width.  And do what I explained.  You will absolutely have to resize your final saved image to not more than 1024 as the largest dimension.........that is locked into the software and there's not way around it.

 

The other way (the way I would do it).  Open the canvas as I explained in my other post (2048 by 1024).  Open as a layer your snapshot.  If the image is smaller than the canvas (there's your background showing all the way around it) then using the height (the smaller dimension) resize the layer (different from resizing the image) to 1024..........that will leave a border at the top and bottom of your picture.  Now resize the image (not just the layer) to 1024 by 512 (exactly as I explained earlier). Save to one of the formats accepted by SL and upload.  No loss of any of your snapshot and at the highest, undistorted resolution possible for use in SL (assumng you did not save to JPEG also at the highest quality).

 

It's the nature of the software you have to deal with in SL..........that's the best anyone can do since there are no workarounds that will improve the results.

 

Editing because I made a big big boo-boo

Instead of resizing the smaller dimension to 1024, resize the LARGER dimension to 2048. 

I think I really confused you now...........I'm so sorry.

[Myra Wildmist]

Thanks, Peggy. I understand. I can crop my photos. I does seem like a bit of an oversight on LL's part. I get that pictures are treated like textures, so maybe, technically, that's problem. Maybe LL needs to rethink the way they handle photos.

[Peggy Paperdoll]

It's not so much what LL did as it is the way graphics are rendered.........it's the most efficient way as Rolig mentioned.  There are literally millions of textures available in SL and if they were all allowed to be any random aspect ratio, the lag would be so great that no one would be able to use SL.  If LL were to delete all but about 10% of the textures then it might lessen the lag to a point that maybe a few high end computers would run with a tolerable amount of lag........the rest of the users would just be out of luck and probably not even able to stay on SL for more than a few minutes before they crashed

 

The powers of 2 are the way computer work..........keeping the textures at that same power of 2 makes the software able to work.  I'm not sure there is an alternative. 

[Rolig Loon]

As I should have emphasized in my earlier posts, the powers of 2 restriction is not an LL creation. It's a requirement of OpenGL, the graphics language that underlies SL.  It's basically a way for limiting the effort (and hence time) that your graphics card would need for rendering the large number of images it has to handle all the time.  Those of us who have worked with textures in SL for a long time don't find it a difficult constraint to deal with. 

[Zena Juran]

If I happen to make an object that has a rectangular face that I want to apply a custom non-seamless texture to, I figure out the aspect ratio of that face and create a texture in PS with that ratio.

SL will convert it to the closest power of 2 upon upload but when I apply it to the face of the object, it will have the perfect aspect ratio.

Sometimes there is a little math involved with texturing objects in SL but it's not all that complicated. And it all depends on how detail oriented you want to be. Seamless textures that are stretched to weird ratios will usually go unnoticed to the average viewer while pics of places and avatars will need a little work to look "right".

I find it worth the small effort to at least get the ratios correct for profile, classified, and parcel pics especially if you are a merchant advertising your craftmanship to everyone.

[Peggy Paperdoll]

I have drawn textures that I intend to use as a desktop back ground for my computer.  My native resolution for my monitor is 1600 by 900 pixels (16 : 9).  I have uploaded a few of those textures that I think I might like to place on a prim to put on a wall or maybe even to give to someone who I think might like it.  I do all the drawing, along with all the other layers I might use to get the final texture (image) at the 1600 by 900 size (I created my canvas for my "project" to that size).  Once I'm finished I save the texture and all the layers involved to my image editor's native format (that being XCF for GIMP).  Then I resize the entire image to the closest power of 2 for each dimension (horizontal and vertical) to upload to SL.  I first unlink the "keep aspect ratio" option (that means I'll size each dimension independently) and size my texture to the width of 1024.  Next I'll size the height dimension to 512 (the reason I use those powers of two is that 16 : 9 is very close to 2 : 1.  16 : 9 is not a power of 2, 2 : 1 is).  Yes that distorts the image but I have control over how it's distorted and the quality of the resizing (the resample rate).

 

When I upload the texture it's at the proper powers of 2 so the SL software does nothing but convert it to JPEG2000 format and saves it to the servers.  I then make my prim by rezzing a cube and sizing on the horizontal and vertical axis's to a 16 : 9 ratio.  A large picture would be 4 meters by 2.25 meters..........a smaller more reasonable size would be 2 meters by 1.13.  The thickness doesn't matter too much.........about .10 works for most picture prims.  When you apply the texture it will be "undistorted" and be at the same aspect ratio as your original.  In the purest sense you have lost some quality due to the downsizing (for the uploaded file) and then effectively upsizing back to the original size when you placed the texture on the prim.  The loss in quality will be almost negligible............and there is no noticeable distortion.

[Korgi Lerwick]

Most of the answers given have explained the business of aspect ratio pretty well. I'm assuming we're talking about snapshots that you've saved to your hard drive, edited and then uploaded into SL.

The key thing to bear in mind is what you intend to do with your finished pic. There are 3 main scenarios:

* a pic for one of the 3 tabs on your profile. There it's important that the aspect ratio or SHAPE of your pic matches the tab where you want to display it. That's been well covered already in this discussion.

* a pic for an image viewer or photo frame where the dimensions are pre-decided. Again the trick is to match the shape of your pic to the shape of the viewer or frame.

* a pic you plan to add to a modifiable prim or frame. There you can make your pics any shape you like so long as you modify the prim or frame to be the SAME shape! For example if I create a pic which has dimensions of, say 1600 x 1150 pixels then I would create a prim which has x and y dimensions of 1.6 m x 1.15 metres. Then I can resize the whole prim to its final size and the shape remains precisely correct. There's no quality loss whatsover and this works perfectly. (Of course SL will resize your uploaded image file to the nearest powers of 2 but this doesn't matter so long as the prim shape and the original image shape are identical!)

This is one of the most misunderstood aspects of SL photography  and it's sad that so many pics which are good appear stretched or squashed.

Good luck Myra!!!

 

Korgi Lerwick

http://www.fotoscopesl.com