Making 37K antique views available for digital 3D viewing #viewing #SPM #alignment #historical #mpo


 

There is an interesting academic project which documents the authors' attempt to make the information captured in historical stereo cards more accessible to modern analysis and research. I encourage folks to at least have a look at the 2 minute video that explains the project, or the longer eight minute talk which goes into some of the technical details discussed in their paper.

In their paper, the authors observe:

One limitation of stereographs is that you need special viewer hardware to experience the stereo effect. To address this limitation, we also introduce a visualization that you can enjoy from a regular 2D display, by synthesizing a continuous 3D camera path, similar to what you would see if you were present at the scene itself (Sec. 3)

Their solution is a very sophisticated wiggle-gram, which certainly is well done. But I believe most of the folks on this list would rather see these images in stereo. I know I would.

So I've started on the project of doing just that, taking the images provided by the authors and turning them into digital 3D formats I can view on my 3D TV and CINERA PRIME viewer. But I also want to make sure that there will be a way of sharing these images with others, either in the formats I've created for myself, or in a format that others can use as a starting point for their own digital format needs. The researchers have done what I would consider the hardest work to make this possible:

As a result, we have found that state-of-the-art stereo rectification and correspondence algorithms often fail to produce good results on antique stereographs. To address this problem, we introduce a new rectification algorithm that leverages the unique properties of antique stereo cameras.

But since the authors were not interested in publishing conventional digital 3D images, the downloadable L/R images are not organized for easy use in creating matched pairs. Fortunately the computer skills needed to fix this, as well as leveraging the supplied meta data, are right in my programming wheel house. I'm excited to start working on this project and have already made significant progress. Stay tuned.

...BC

PS: Thanks to Ian Walker for the original posting about the KeystoneDepth project on March 18th -- which happens to be my birthday -- so this was a great birthday present!


John Clement
 

I prefer viewing them on my 3D monitor or my other 3D viewers after restoring them.

 

John M. Clement

 

From: main@Photo-3d.groups.io <main@Photo-3d.groups.io> On Behalf Of Bill Costa
Sent: Monday, March 22, 2021 8:44 PM
To: main@Photo-3d.groups.io
Subject: [Photo-3d] Making 37K antique views available for digital 3D viewing #3dtv #alignment #historical #mpo #SPM #viewing #workFlow

 

There is an interesting academic project which documents the authors' attempt to make the information captured in historical stereo cards more accessible to modern analysis and research. I encourage folks to at least have a look at the 2 minute video that explains the project, or the longer eight minute talk which goes into some of the technical details discussed in their paper.

In their paper, the authors observe:

One limitation of stereographs is that you need special viewer hardware to experience the stereo effect. To address this limitation, we also introduce a visualization that you can enjoy from a regular 2D display, by synthesizing a continuous 3D camera path, similar to what you would see if you were present at the scene itself (Sec. 3)

Their solution is a very sophisticated wiggle-gram, which certainly is well done. But I believe most of the folks on this list would rather see these images in stereo. I know I would.

So I've started on the project of doing just that, taking the images provided by the authors and turning them into digital 3D formats I can view on my 3D TV and CINERA PRIME viewer. But I also want to make sure that there will be a way of sharing these images with others, either in the formats I've created for myself, or in a format that others can use as a starting point for their own digital format needs. The researchers have done what I would consider the hardest work to make this possible:

As a result, we have found that state-of-the-art stereo rectification and correspondence algorithms often fail to produce good results on antique stereographs. To address this problem, we introduce a new rectification algorithm that leverages the unique properties of antique stereo cameras.

But since the authors were not interested in publishing conventional digital 3D images, the downloadable L/R images are not organized for easy use in creating matched pairs. Fortunately the computer skills needed to fix this, as well as leveraging the supplied meta data, are right in my programming wheel house. I'm excited to start working on this project and have already made significant progress. Stay tuned.

...BC

PS: Thanks to Ian Walker for the original posting about the KeystoneDepth project on March 18th -- which happens to be my birthday -- so this was a great birthday present!


Bill G
 

Bill, great post.  I have been to the UCR campus several times over the past 5 years to explore their stereo views.  Their collection is vast and is open to the public by appointment.  As stated, many of the views have issues, like most photographs from the 1800s.  I agree with your assessment, the task force is doing the heavy lifting correcting the images. I love your idea of extracting the left and right image for more conventional 3d use. 

A few questions...

I assume you will be writing some type of program to extract the L n R image from their wigglegram? 

Is this project completed, or is it on going?  If on going, what % of the 37K images are completed by the task force?

Are the wigglegrams accessible to the public yet?  If so, where?

Bill G


On Mon, Mar 22, 2021 at 6:44 PM Bill Costa <Bill.Costa@...> wrote:

There is an interesting academic project which documents the authors' attempt to make the information captured in historical stereo cards more accessible to modern analysis and research. I encourage folks to at least have a look at the 2 minute video that explains the project, or the longer eight minute talk which goes into some of the technical details discussed in their paper.

In their paper, the authors observe:

One limitation of stereographs is that you need special viewer hardware to experience the stereo effect. To address this limitation, we also introduce a visualization that you can enjoy from a regular 2D display, by synthesizing a continuous 3D camera path, similar to what you would see if you were present at the scene itself (Sec. 3)

Their solution is a very sophisticated wiggle-gram, which certainly is well done. But I believe most of the folks on this list would rather see these images in stereo. I know I would.

So I've started on the project of doing just that, taking the images provided by the authors and turning them into digital 3D formats I can view on my 3D TV and CINERA PRIME viewer. But I also want to make sure that there will be a way of sharing these images with others, either in the formats I've created for myself, or in a format that others can use as a starting point for their own digital format needs. The researchers have done what I would consider the hardest work to make this possible:

As a result, we have found that state-of-the-art stereo rectification and correspondence algorithms often fail to produce good results on antique stereographs. To address this problem, we introduce a new rectification algorithm that leverages the unique properties of antique stereo cameras.

But since the authors were not interested in publishing conventional digital 3D images, the downloadable L/R images are not organized for easy use in creating matched pairs. Fortunately the computer skills needed to fix this, as well as leveraging the supplied meta data, are right in my programming wheel house. I'm excited to start working on this project and have already made significant progress. Stay tuned.

...BC

PS: Thanks to Ian Walker for the original posting about the KeystoneDepth project on March 18th -- which happens to be my birthday -- so this was a great birthday present!


Bill G
 

a few additional observations...

I agree with your assessment, these are some of the best wigglegrams I have seen. 

Does the team make the wigglegram software available to the public? 

One of the issues with wigglegrams for high volume conversion is filling in the hidden areas.  I noticed some of their wiggle grams did a poor job at filling in the hidden area, and others were much better.  However, even the ones which did a poor job, surprisingly, it does not detract from the 3d effect.  I assume this is because the hidden areas typically represent a very small % of the total image area.  Just like some forms of disparity in 3d viewing, our brains seem to provide a reasonable leeway for imperfections before it becomes annoying or a distraction.





On Mon, Mar 22, 2021 at 6:44 PM Bill Costa <Bill.Costa@...> wrote:

There is an interesting academic project which documents the authors' attempt to make the information captured in historical stereo cards more accessible to modern analysis and research. I encourage folks to at least have a look at the 2 minute video that explains the project, or the longer eight minute talk which goes into some of the technical details discussed in their paper.

In their paper, the authors observe:

One limitation of stereographs is that you need special viewer hardware to experience the stereo effect. To address this limitation, we also introduce a visualization that you can enjoy from a regular 2D display, by synthesizing a continuous 3D camera path, similar to what you would see if you were present at the scene itself (Sec. 3)

Their solution is a very sophisticated wiggle-gram, which certainly is well done. But I believe most of the folks on this list would rather see these images in stereo. I know I would.

So I've started on the project of doing just that, taking the images provided by the authors and turning them into digital 3D formats I can view on my 3D TV and CINERA PRIME viewer. But I also want to make sure that there will be a way of sharing these images with others, either in the formats I've created for myself, or in a format that others can use as a starting point for their own digital format needs. The researchers have done what I would consider the hardest work to make this possible:

As a result, we have found that state-of-the-art stereo rectification and correspondence algorithms often fail to produce good results on antique stereographs. To address this problem, we introduce a new rectification algorithm that leverages the unique properties of antique stereo cameras.

But since the authors were not interested in publishing conventional digital 3D images, the downloadable L/R images are not organized for easy use in creating matched pairs. Fortunately the computer skills needed to fix this, as well as leveraging the supplied meta data, are right in my programming wheel house. I'm excited to start working on this project and have already made significant progress. Stay tuned.

...BC

PS: Thanks to Ian Walker for the original posting about the KeystoneDepth project on March 18th -- which happens to be my birthday -- so this was a great birthday present!


Ian Walker
 

Wow, this sounds like a fabulous project - can't wait to see the results and so happy that my contribution might have played a small part in kicking off such a mighty endeavor! :-)

(Oh. and a Happy Birthday - March 18th is also my wedding anniversary so I have passed on your thanks to my wife for letting me "play on my computer" on such a hallowed day :-))


KenK
 

On Mon, Mar 22, 2021 at 08:44 PM, Bill Costa wrote:

this project and have already made significant progress. Stay tuned.

This is great news. Do you plan a demo at some point or are you going to release all 37k in one fell swoop? 


 

JMC: prefer viewing them on my 3D monitor or my other 3D viewers after restoring them.

For my TV MPO formats work well since both my DVD player and TV recognize it and automatically do the right thing -- going into 3D mode. Right now on my CINERA I use the provided photo gallery app. For the images to display correctly, I format the 3D image as SBS JPG with borders. Of course there are lots of digital 3D formats, simple SBS, squashed under/over squashed SBS, interlaced, etc. Which formats work best with your devices/software?


KenK
 

On Tue, Mar 23, 2021 at 04:56 AM, Bill G wrote:
Are the wigglegrams accessible to the public yet?  If so, where?
 
Yes. See Bill Costa's first link (repeated below) - the top menu bar has Explore and Download options. The Explore option shows the wigglegrams
 
On Mon, Mar 22, 2021 at 08:44 PM, Bill Costa wrote:
There is an interesting academic project


 

BG: I assume you will be writing some type of program to extract the L n R image from their wigglegram?

Don't have to. That's what is so great about the KeystoneDepth study, all of the images are provided as individual L and R PNG files. Not only that, the files are available in 3 different stages of conversion, plus the wiggle-grams. More on this later.

BG: Is this project completed, or is it on going? If on going, what % of the 37K images are completed by the task force?

Well the KeystoneDepth study is complete and all of the images are available for download. My project, which I guess should have a name, is in progress.

BG: Are the wigglegrams accessible to the public yet? If so, where?

If you go to the KeystoneDepth download page they are available using the _gifs links. At least thet is what I assume. I used the [pairs] links to download the L/R pairs. Be aware these are massive downloads. But you can use their explore page to search and view images from the collection.

BG: Does the team make the wigglegram software available to the public?

Don't know. I'm not going to be paying much if any attention to the wiggle-grams in this thread or my project. For discussions about the wiggle-grams, see the original thread and 3D-photo-inpainting which is a fork from that thread.

..BC


 

KK: Do you plan a demo at some point or are you going to release all 37k in one fell swoop?

I'm going to do some short posts about work done so far and plans for next steps. I want to do short posts to make them easy to read and comment on. One of the things to work out will be where to park the images so I imagine that will start with some small subset that people will be able to try to make sure it works well. My goal will be to make all the images available, but there are just so many images that I could imagine a long tail of the project would be an group effort of doing some sort of curation. Remember sturgeons law -- a large number of these images are not really going to be of much interest. The trouble is, of course, what is interesting to me may not be interesting to you.

...BC


John Clement
 

The various schemes to make the 3D accessible by fly by wigglegrams may have a big unintended effect.  It may harden the opinions against real stereoscopic presentations and do more to kill 3D.  One of the joys of viewing some of the old pictures in digital 3D is the ability to zoom in and peruse details.  The Civil War photo negatives are high resolution and can be magnified quite a lot.  Unfortunately the released images may have limited resolution which prevents zooming.

 

If you have a panned image with reconstructed missing detail, it is then possible to make a good lenticular print for glasses free viewing at a variety of angles.  That might be a better use of the 3D images.

 

A large number of stereoviews are available in high resolution scans from the Library of Congress.  The NY Public Library also has a number, but they do not release many of their scans free to the public.  The LC has only scanned a fraction of their holdings, but what is available is fantastic, however just raw images.  That is OK as I can usually fix most of the defects, but some have proved to be more work than I was willing to do.  The debate over original vs restored can be resolved by offering both.

 

John M. Clement

 

 

Bill, great post.  I have been to the UCR campus several times over the past 5 years to explore their stereo views.  Their collection is vast and is open to the public by appointment.  As stated, many of the views have issues, like most photographs from the 1800s.  I agree with your assessment, the task force is doing the heavy lifting correcting the images. I love your idea of extracting the left and right image for more conventional 3d use. 

 

A few questions...

 

I assume you will be writing some type of program to extract the L n R image from their wigglegram? 

 

Is this project completed, or is it on going?  If on going, what % of the 37K images are completed by the task force?

 

Are the wigglegrams accessible to the public yet?  If so, where?

 

Bill G

 

On Mon, Mar 22, 2021 at 6:44 PM Bill Costa <Bill.Costa@...> wrote:

There is an interesting academic project which documents the authors' attempt to make the information captured in historical stereo cards more accessible to modern analysis and research. I encourage folks to at least have a look at the 2 minute video that explains the project, or the longer eight minute talk which goes into some of the technical details discussed in their paper.

In their paper, the authors observe:

One limitation of stereographs is that you need special viewer hardware to experience the stereo effect. To address this limitation, we also introduce a visualization that you can enjoy from a regular 2D display, by synthesizing a continuous 3D camera path, similar to what you would see if you were present at the scene itself (Sec. 3)

Their solution is a very sophisticated wiggle-gram, which certainly is well done. But I believe most of the folks on this list would rather see these images in stereo. I know I would.

So I've started on the project of doing just that, taking the images provided by the authors and turning them into digital 3D formats I can view on my 3D TV and CINERA PRIME viewer. But I also want to make sure that there will be a way of sharing these images with others, either in the formats I've created for myself, or in a format that others can use as a starting point for their own digital format needs. The researchers have done what I would consider the hardest work to make this possible:

As a result, we have found that state-of-the-art stereo rectification and correspondence algorithms often fail to produce good results on antique stereographs. To address this problem, we introduce a new rectification algorithm that leverages the unique properties of antique stereo cameras.

But since the authors were not interested in publishing conventional digital 3D images, the downloadable L/R images are not organized for easy use in creating matched pairs. Fortunately the computer skills needed to fix this, as well as leveraging the supplied meta data, are right in my programming wheel house. I'm excited to start working on this project and have already made significant progress. Stay tuned.

...BC

PS: Thanks to Ian Walker for the original posting about the KeystoneDepth project on March 18th -- which happens to be my birthday -- so this was a great birthday present!


Bill G
 

BC, thx for the link and added infor.
When I did a search and downloaded some of the LR images
I found the resolution in the range of approx 600 x 700 pixels per image.
Is this the level of resolution they all are?   I wonder if higher rez scans exist.  Many of the glass plates offer much more resolution than this.
BG


On Tue, Mar 23, 2021 at 6:06 AM Bill Costa <Bill.Costa@...> wrote:

BG: I assume you will be writing some type of program to extract the L n R image from their wigglegram?

Don't have to. That's what is so great about the KeystoneDepth study, all of the images are provided as individual L and R PNG files. Not only that, the files are available in 3 different stages of conversion, plus the wiggle-grams. More on this later.

BG: Is this project completed, or is it on going? If on going, what % of the 37K images are completed by the task force?

Well the KeystoneDepth study is complete and all of the images are available for download. My project, which I guess should have a name, is in progress.

BG: Are the wigglegrams accessible to the public yet? If so, where?

If you go to the KeystoneDepth download page they are available using the _gifs links. At least thet is what I assume. I used the [pairs] links to download the L/R pairs. Be aware these are massive downloads. But you can use their explore page to search and view images from the collection.

BG: Does the team make the wigglegram software available to the public?

Don't know. I'm not going to be paying much if any attention to the wiggle-grams in this thread or my project. For discussions about the wiggle-grams, see the original thread and 3D-photo-inpainting which is a fork from that thread.

..BC


 

JMC: A large number of stereoviews are available in high resolution scans from the Library of Congress.

I have looked at those and have thought it would be a great project to convert them to a digital 3D format. The problem is I don't know how I would create a program to automatically and reliably extract the left and right images from the card. The KeystoneDepth webpages offer the card images already separated, so I can easily take it from there using readily available tools like SPM and my own modest programming abilities. At some point I will want to approach the authors of KeystoneDepth to inquire about the availability of the software they used with perhaps the goal of using that tool on the Library of Congress archive and others. In the meantime perhaps somebody in this group could reach out to the KeystoneDepth authors about the software availability while I continue work on the current archive? While the KeystoneDepth images may not be high-rez, they are certainly acceptable and worth working with.

BG: When I did a search and downloaded some of the LR images I found the resolution in the range of approx 600 x 700 pixels per image. Is this the level of resolution they all are? I wonder if higher rez scans exist. Many of the glass plates offer much more resolution than this.

Well remember that we are talking about 37,000 images here, so there is the matter of having enough disk space! Perhaps the way to look at this trove of photos is as a set of really good 3D thumbnail images which can be used to cherry-pick photos of particular merit that deserve being revisited at a higher resolution (assuming it is available). At this point, even if I had access to them, 37,000 high-rez images would simply be too much data to handle. As the availability of cheap storage and CPU speed continue to increase, this will be less of in issue in the future, but for now I'm running with what we've got today. (There's never time to do it right, there's always time to do it over.)

...BC


John Clement
 

The LC site has images in a variety of resolutions.  The cards are not at very high resolution, but scans of the glass negatives are usually extremely high.  For space reasons I have had to reduce some of them to 4k+.  When I use them as substitutes for the same image on a View-Master reel, the result is astoundingly good.  It is possible to extract a wide screen image by some trickery.  The R image has extra on the R side and the L has more on the L side.  Normally this would be cropped off, but with careful editing it can be moved to the extra R image info, is moved to the R hand side of the L image.  That way you can get the normal looking through a window effect with more of the scene in view.

 

Each scene on a card or negative requires personal hand work for best effect.  Maybe someday there will be an AI program to handle the work but until then HI is the only way.  I have only restored LC images which I could substitute for the View-Master images, plus a few to fill in the VM views.  VM made some good selections for their packets so why not use their work!  I hope I have sparked a bit of interest in the VM reels by putting them on YouTube in 3D.  3D needs all the help it can get to interest enough people so it will stay alive for the rest of my life.

 

John M. Clement

 

From: main@Photo-3d.groups.io <main@Photo-3d.groups.io> On Behalf Of Bill Costa
Sent: Tuesday, March 23, 2021 11:17 AM
To: main@Photo-3d.groups.io
Subject: Re: [Photo-3d] Making 37K antique views available for digital 3D viewing #3dtv #alignment #historical #mpo #SPM #viewing #workFlow

 

JMC: A large number of stereoviews are available in high resolution scans from the Library of Congress.

I have looked at those and have thought it would be a great project to convert them to a digital 3D format. The problem is I don't know how I would create a program to automatically and reliably extract the left and right images from the card. The KeystoneDepth webpages offer the card images already separated, so I can easily take it from there using readily available tools like SPM and my own modest programming abilities. At some point I will want to approach the authors of KeystoneDepth to inquire about the availability of the software they used with perhaps the goal of using that tool on the Library of Congress archive and others. In the meantime perhaps somebody in this group could reach out to the KeystoneDepth authors about the software availability while I continue work on the current archive? While the KeystoneDepth images may not be high-rez, they are certainly acceptable and worth working with.

BG: When I did a search and downloaded some of the LR images I found the resolution in the range of approx 600 x 700 pixels per image. Is this the level of resolution they all are? I wonder if higher rez scans exist. Many of the glass plates offer much more resolution than this.

Well remember that we are talking about 37,000 images here, so there is the matter of having enough disk space! Perhaps the way to look at this trove of photos is as a set of really good 3D thumbnail images which can be used to cherry-pick photos of particular merit that deserve being revisited at a higher resolution (assuming it is available). At this point, even if I had access to them, 37,000 high-rez images would simply be too much data to handle. As the availability of cheap storage and CPU speed continue to increase, this will be less of in issue in the future, but for now I'm running with what we've got today. (There's never time to do it right, there's always time to do it over.)

...BC


Boris Starosta
 

Bill, I notice that at least some fraction of the stereoviews are tagged with keywords.  For example, searching for Zeppelin (or airship) produces a couple dozen hits.  Will you have a way to keep the keyword tags with the images?  How many of the 37K images do you observe to be tagged?  Keywording is very labor intensive... cannot be automated; so a large fraction of the value in this collection will be in the image keywords/tagging.
Boris

--
––––
Contemporary work in the Stereoscopic Arts:  www.patreon.com/retroformat
––––


 

On Tue, Mar 23, 2021 at 03:09 PM, Boris Starosta wrote:

at least some fraction of the stereoviews are tagged with keywords

Turns out that there is meta data for most of the images. It is supplied as a JSON data file which is keyed by the image's file number. The JSON file is easy for me to parse and slice and dice as needed. I haven't look in detail yet at the EXIF data within the images themselves, but I think it would be advantageous if at least some of the data in the JSON file could be applied to each file's own EXIF data. Although keep in mind that a lot of image tools will unfortunately strip or loose such data during processing. In any case it is something that I am aware of and in particular I'm hoping to be able to use the available meta data to do some sort of organizational grouping of the files themselves. Right now they are pretty much an amorphous mass with no apparent organization that I can see as of yet.

...BC


 

On Tue, Mar 23, 2021 at 01:20 PM, John Clement wrote:

Each scene on a card or negative requires personal hand work for best effect

That is something I am specifically avoiding with this project. My goal is to use as much automation as possible to process the images. This is not to say that there are not some images that are worth being carefully hand curated and edited. In fact it is my hope that having easy and effortless access to actual 3D versions of all of the *KeystoneDepth images will make it easy to find those images worth the special handling. In my initial viewing of the first 1,000 images, it is clear that there are a lot of duplicates and, quite frankly, duds. But every once in a while there is a gem. It would be good if we could somehow crowd source the identification of the gems. Something to consider later in the project.

...BC


Bill G
 

BC, that was my experience when  viewing many of the slides on the UCR campus...
3/5 min. were duds, or of no significance or interest, like a landscape shot with a horse in it....it would look the same today.  But I did not go through all 37K of them! 
As JC pointed out, like the LC, the 37K does have a significant amount of glass slides of good resolution, again, not sure the %.
I hear u on the resolution and storage issue.  I guess my thought was, to go through all this trouble, and not do higher resolution scans, at least on the glass plate images, seems like a waste.  As u point out though, maybe higher resolution scans do exist, but were not released due to storage size limitations and download time.  The staff there used to answer the phone and were quite helpful, at least pre covid.
BG


On Tue, Mar 23, 2021 at 2:27 PM Bill Costa <Bill.Costa@...> wrote:

On Tue, Mar 23, 2021 at 01:20 PM, John Clement wrote:

Each scene on a card or negative requires personal hand work for best effect

That is something I am specifically avoiding with this project. My goal is to use as much automation as possible to process the images. This is not to say that there are not some images that are worth being carefully hand curated and edited. In fact it is my hope that having easy and effortless access to actual 3D versions of all of the *KeystoneDepth images will make it easy to find those images worth the special handling. In my initial viewing of the first 1,000 images, it is clear that there are a lot of duplicates and, quite frankly, duds. But every once in a while there is a gem. It would be good if we could somehow crowd source the identification of the gems. Something to consider later in the project.

...BC


Martin Schub
 

Bill Costa writes:

> The problem is I don't know how I would create a program to automatically and reliably extract the left and right images from the card.

Bill, you're welcome to the program I wrote for Robert Thorpe, which automatically extracts the image portion of a whole-card scan, removing the left and right margin and the bottom margin.  We discussed it here some time back.


Boris Starosta
 

Crowdsourcing the ID of "gems" and adding keywords would be a great way to add value to the collection.  No single person could do all of it - simply for the sheer volume of the work, and also because a single individual could bring only their own unique history and preferences and expertise to the images.  If a crowd could contribute, then you'd have the interest and expertise of all those people who wanted to contribute, and all the manpower...  I wonder if a wiki-like system would work.  Each image got a "page" on the wiki, and people could add information as they paid a visit to the image (and maybe that information could be up- or down-voted by other visitors to provide a ranking to it).  Information like: keywords on content in the image, meta-keywords pertaining to image quality (e.g. "pseudo"  "scratched"  "boring"  "abstract"  "flat"), recommendations for, or even opportunity to "fix" image problems?  (I suppose that last option could always exist in the form of people submitting a URL to "their fixed/modified version" of the image, stored elsewhere)

A huge project, but also a huge opportunity for our community.
--
––––
Contemporary work in the Stereoscopic Arts:  www.patreon.com/retroformat
––––