Lin Evans

Hi Albert, Thanks very much for testing this - I think this is the first time this one was tested on Vista as I recall. Best regards, Lin

Hi Andrew, Here is the PTE for the best I've been able to achieve using the aforementioned techniques. See my answer to Al for details. My new link Best regards. Lin

Hi Al, Indeed - an interesting one! The transition and timings seem to be key, but even with precise settings the effect is only "near" perfect. Something seems to be missing. Here is the best I could come up with. I used the same image here three times. The smoothness as in a continuous motion was best obtained by using a 1000 (1 second) transition but pulling the ending keypoint on the first slide back to the start of the transition. The second slide starting point was pushed into and slightly past the keypoint end on the timeline for slide one. The "curve" was altered on the non-linear transition smooth to give a six percent rise rather than the twenty percent default and "perspective correction" was set on. One way of achieving the imperceptible transition from slide to slide is to stop the zoom as with the transition between slides two and three. This is about as good as I've been able to achieve. One thing which will make doing this on a panorama quite different and "better" I think is that all areas of the pano will be in full depth of field focus. In my Kachina sample, the limited DOF on extreme macros makes the transitions from slide to slide much more obvious as the slides differ greatly depending on whether I was using a wide angle (great DOF) or macro lens. In fact to get "sufficient" DOF on the final slides I had to combine nine different photos into one with each photo having a different area of the Kachina in focus via Helicon Focus. So, the bottom line is that this technique, in so far as imperceptible changes from slide to slide, is better accomplished where there is infinite DOF. Of course for panoramas which were taken primarily at infinity focus this will present less of a problem. http://www.lin-evans.net/pte/infinite.zip Best regards, Lin

At the transition points, the way to avoid this is to use linear zoom and pull the slide two Pull the start point for slide two into the ending point for slide one and with linear the continuation is smooth but you have to also adjust the size for slide two to compensate for the shift point. This leaves only the starting and final slides where you need or might want to use smooth and you get around this issue by duplicating the starting and ending slide. Put duplicate slides for number one and set the smooth on the first and linear on the second so you are continuing your initial zoom into slide two. I won't be able to post a sample of this tonight but tomorrow I should have, hopefully, a working example. I have to test this with PTE but I've been doing this with PSG for years so I "think" it should be possible with P2E. Of course there is a slight differene in that PSG has transitions which allow this smooth continuation. That is they have a Fade in and a Fade out as separate transitions where PTE has a "Fade-in/Fade out. This "may" present a problem unless the manual adjustment of the smoothing line and timing can compensate. The issue is that we need a continuation in the zoom during the brief transition time between end of zoom and start of zoom for the respective slides. The only way I know to avoid this apparent "jerk" is to begin slide two while slide one is still zooming but to diminish slide one with possibly an opacity shift at the end of its zoom. More testing will be necessary. Best regards, Lin

Hi Al, Except there are no "objects" other than the slide itself. The thing I'm trying to line up are better described as a "subject" in photo 2 which is the same subject as in photo 1 but taken from either a different focal length or a different position and reduced in size at the start of the second photo by zooming out. This allows a greater zoom level on photo 2 than on photo 1. This progression then continues until the final slide. It's easy enough to copy slide two to slide one, set the opacities in such a way as to line them up then write down the pan and zoom numbers for the temporary "object" (slide two) then exit PTE without saving, open the PTE file back up, go to slide two, click on the first keypoint and insert the numbers from the ending keypoint of the temporary slide two which matches the perfect fit from slide one. It's easy, but very time consuming. Having a very precise, thin line grid with numeric references could greatly speed up the process. The reason for not using the subject matter as a PNG object is that where I'm going with this is for zooming in on panoramas such as the Cliff Palace example posted earlier. This means to perform a really tight zoom, I would create multiple photos of areas of interest using long focal lengths then seamlessly "patch" them in for tight zooms to see ultimate detail, then switch back seamlessly to the panorama when the zoom out reaches the point where the black monitor borders would appear should the zoom out go further. If you look into Zoomify, you see that what it does is take a giant file and break it down into file segments. This could be 20 or so for a rather small panorama or as many as 5,000 or more for something like the 13 gigapixel pano of Harlem. Then as the viewer zooms in or out, these smaller files are switched in and out to allow an infinite zoom (to the limits of the focal length of the original frames). With Zoomify there is an "out-of-focus" period (very short) where they hide the switch process. Microsoft has a beta version of a very similar process except it's very smooth without the obvious out-of-focus period. I'm not certain how they do this. Microsoft uses some 1.4 gigabyte examples, but have nothing nearly as impressive in terms of file size as the Harlem example done with Zoomify. I don't think this would be too difficult to do with PTE. It would take some time to do a really impressive example because of the necessity to match the zoom very carefully. But the nice thing about it is that for the panorama examples there wouldn't be the obvious changes as in my Kachina example because the number of frames used for the panorama would be masking the changes. That is one would already be zoomed in to the 100% point for the original before dropping in a "patch" frame for the increased detail capabilities. I wasn't able to do an example outside because we have had 100+ mph winds and I couldn't keep the tripod locked down - LOL. Perhaps next week I can get some real samples to work with. It's necessary to very carefully match photos for this so that the perspective doesn't shift which would make it difficult to do. With my kachina samples I had already done a detailed macro several months ago and just did some quick handheld tabletop photos of the same subject from "approximately" the same position to play with. Ultimately its not nearly as accurate as I would like. Best regards, Lin

Hi Andrew, Because they are on different slides and there is no way to display them simultaneously. The ending size for the object on slide one must precisely match the beginning size of the object on slide two. It can't be close but must be absolutely identical or there will be a percepitble "jerk" when the zoom begins on the second slide. A precision grid is absolutely essential to do this. If one were using the identical image on two slides then it would be easy to match the zoom and pan numbers, or if there were a way to visually do it as you suggest, but having both objects on the same slide defeats the purpose of conservation of video RAM and would be, for all practical purposes the same as just having one huge file. It's possible to copy slide two as an object on slide one, line them up visually with each set to limited opacity, write down the numbers from the objects and animations blocks for object two, delete object two and copy the numbers from the match into slide two but tha'ts a lot of extra steps and it would be much quicker to have a pop-up grid with very fine lines and reference numbers. Best regards, Lin

Hi Fred, Thanks! Actually, I would have loved to not have to use some fade transitions in the earlier slide transitions, but I just didn't have anything with sufficient precision to match up the tiny subjects from slide end to next slide beginning. But as I told Ken, I'm satisfied after the test that it's indeed 'possible" to do this type thing. There were a few completely undetectable transitions on some of the later slides which were large enough for me to "fudge" the subject into proper alignment. Al Robbins earlier produced a very nice grid, but for doing this correctly we need something resident in the program so we don't have to load and unload a PNG grid. I made one earlier as well, but it doesn't have nearly the precision needed. Best regards, Lin

LOL - one of the benefits of being "seim-retired"... Of course being broke most of the time is the downside - HA! I'll have to fix those scratches on her table - hope I didn't put them there... The infinite zoom idea will definitely work. I had to use some fade transitions in the earlier frames because I didn't have any grid precise enough to do perfect matching, but with a perfect match on the end-slide, begin-slide sequence a "quick, no transition" is virtually undetectable as it was in several of the later images where they were large enough that I could more closely (on in some cases perfectly) align slide content position. I'm satisfied that this is indeed possible and would work to good advantage for things like panoramas where one might want to zoom in very tightly on particular areas. The key is in segmenting the huge files into "byte" sized pieces which can be seamlessly integrated with PTE both on zoom in and zoom out. Perhaps Igor will be able to give us a precise grid with reference numbers for the next release (5.1, etc.) and this will make it much easier to match things up. Of course the photography must be amenable, but that's another issue. Best regards, Lin

O.K., this is a first attempt at an "infinite zoom" sort of after the effect of Zoomify. Several issues became aparent. First, we definitely need a precision grid to do this type thing well. It's necessary to match image positions and sizes nearly perfectly from end of slide one to beginning of slide two, etc. Second, we need nearly perfect photography so this first attempt is not "perfect" because I had to quickly take an number of frames to attempt to match a macro I did several months ago. It was not entirely successful, but not bad considering. Third, we need to use consistent zoom proportions from slide to slide. For example, if you allot 30 seconds to each slide, then the percentage of zoom per slide must remain very consistent or there will be issues. In this sample there are seven slides. It's not feasible to reduce a huge high resolution slide to a small size then zoom in to 100% to achieve this effect because of the too large RAM loading on the video card. In Zoomify, and the new Microsoft HD zoom software, a huge image is broken down into manageable segments which are then switched in and out as the zoom progresses. I attempted to somewhat simulate this by matching 100% end size to reduced start size per slide. The subject was a tiny wooden carved Kachina doll made by a Native American Artist. The height of the carved portion not counting the tiny wooden base is about one centimeter, so very, very tiny. The final macro was done by combining nine images each taken with a slightly different focus point for infinite depth of field. Helicon Focus was used to achieve the image used in the last couple frames. The rest of the photos were taken hand held tonight with three different lenses. 15-30mm, 70-300mm and 80-400mm. Anyway, here's the first attempt. About 22 meg zipped executable and a bit over 5 minutes of zoom.... 6/10/07 - I've modified the program slightly for better registration between transitions and made it visually a little more appealing I think... Lin My link to Kachina Zoom

Hi Cor, I don't think you need a new computer, just a new video card. The computer itself has quite sufficeint power to run the slideshows without problem, but the MX460 doesn't. As I told someone else a few days ago, if you have an AGP (advanced graphics port) slot on the bus (run dxdiag to see) you can get an older ATI Radeon 9800 Pro on the web for a bit over (and sometimes under) $100 U.S. - certainly cheaper than a new computer and will run virtually any slideshow smoothly without problems. Best regards, Lin

Here's my take on it. If you are going to spend the time and effort to scan images from 35mm negaitve or transparency, you should go ahead and scan them at a dpi which will recover all the detail possible from the slide or negative within reason. By within reason, I mean that an expensive drum scan at a bit over 5,000 dpi "might" still pull a bit of extra detail from the transparencies or negatives "if" ond only "if" the originals were done with very fine grain media. In the case of "average" or "typical" 35mm frames, there is little probability that scanning at over 4,000 dpi would be of any value, but scanning at lower resolutions will not get all the latent detail in a properly focused original on fine grain film or transparency. The problem is that since there is a good deal of time and effort involved, why not create a digital original which you can later print at whatever maximum size the individual frame will allow considering subject matter and grain? You can "always" resample the images down for your slideshow to whatever pixel count you need or want, but you can't go the other way (interpolate) and get any more detail and quality than in the original scan. Just my $02 on this... Best regards, Lin

Hi Giel, Very nice! I love the happy mermaids!! Beautiful images of impressive rigging. Best regards, Lin

Hi Silvia, You would ask me that! LOL Actually, I didn't write it down when I made the text, I just picked one which looked nice to me. I spent about 30 minutes trying to match it and it may have been "Calligraph421 BT" in itallic. I created the text as a PNG file in Photoshop and I used the Layers Drop Shadow and Contour to give the characters some dimensionality so they don't look "exactly" as they would if I hadn't applied these effects. Best regards, Lin

Hi Judy, Thanks Judy! Those are small discs of adobe brick laid out on cement slabs. They are used as color samples by the rangers when mixing and matching sandstone and mortar for making repairs which have to be done on occasion to keep the integrity of some of the buildings and walls. Best regards, Lin

Hi Neil, Good to hear. Both Al's and my pano should be smooth with most systems. Best regards, Lin

Hi Al, The same base algorithm developed by the authors of "autostitch" at UBC is used in Autopano Pro but there is very sophisticated automatic blending, etc., to take care of the problems which you needed to address by cloning. This method is licensed now to three different pano software developers. Autopano Pro was used to stitch the 13 gigapixel Harlem photo. Best regards, Lin

Hi Andrew, Not a problem in this particular case because the subject was quite far away and typical of Bryce Canyon images - primarily the beautiful rock cliffs. The same was true for Peter Grote's Annapurna frame done with the Betterlight scanning camera. To get great results from many, many overlapping frames, it's necessary to choose the subject matter more carefully. I've used PTAssembler and PTGui and both have strengths and weaknesses like most stitching programs. The reason I like Autopano Pro is the combination of power and simplicity. It has excellent blending uses three types of interpolation, can match odd frames which are very difficult for other programs and you can simply throw all your images into a single folder, even ones not part of the stitch, and the program will automatically find, order and stitch every possible pano. It has some fantastic features. I'll get back with another test a bit later. Best regards, Lin

A suggestion would be to create a quick and simple slideshow with music which does this and let some other users try it on their systems. Are you by any chance using a SoundBlaster Audigy sound card? The reason I ask is that I get some of this through my headphones with my SoundBlaster, but when I play the same slideshow on other systems the sound plays perfectly. The sound on this system plays fine on CD's and such but there is something not quite right about the SoundBlaster. I don't have this issue with any of my other seven computer systems. Best regards, Lin

If I understand right you are looking for a "universal" or default setting? If not, each button's size can be set in one of two ways. You can either simply drag the bounding rectangle until the size is suitable, or you can do that then look at the zoom numbers and duplicate them for each button. The "problem" for me is that if the text lies on the button itself, the button will change sizes depending on the "quantity" of text characters. For example, if you have the default word "Button" on the button, it will be larger than the same button with the simply the letter "B" on it. So you can't change the font size directly, except by choosing a different font, but you "can" make the text on the button invisible then add your own text which you can size as you did the button with the bounding rectangle. The you can set set the text to an "action" if you wish. It's a bit kludge, but is a work-around. It would be nice if there were a way to simply set a default size for all buttons and change the font sizes to keep the same button font but use a smaller font size for longer button I.D. text. It's something I asked about in an earlier post but Igor has been very busy with trying to get version 5 out so we may have to wait until the first upgrade for him to address this question. Best regards, Lin

A question from a forum member prompted me to write this brief reminder. If you have the need to automate running several of your slideshows sequentially from a CD or from the computer in a sequence without intervention from the user you probably realize that this can be done by simply setting the last slide in each show to run the next program in sort of a "daisy chain" sequence, but what if you would like to do this without having each executable be encumbered with a code which runs some other file? Perhaps you would like to use these executable files without having to modify each to fit a particular purpose or one-time use?? The use the "invisible menu". This is easily done by taking advantage of the fact that PicturesToExe can make any object or slide invisible by setting the opacity to zero. So how can this be used to create the scenario above? Just create a brief executable file containing a few "invisible" slides (set opacity to zero). Set the time to a couple seconds or so for each invisible slide, set the effect to "quick, no transition," and set the slides to run an external progam which would be the slidshows you wish to sequence, etc. This "invisible menu" then runs the first invisible slide, executes the desired program, returns invisibly and runs the second invisible slide which in turn calls the next desired program, etc. Thus you have circumvented modifying each slideshow with a slide which calls another show and eliminate the need to make changes to several slideshows. Lots of possibilities here folks! Best regards, Lin

Hi Peter, I'll touch on a few of them. First let me say that Memories On TV Pro, Proshow Producer and Proshow Gold are very, very nice programs. Each are feature rich and there are things such as drop-in video clips, adjust images within program, some sound editing capabilities, etc., which Producer and Gold can do which are not currently possible with PTE. Also with Producer there are advanced masking features, Flash and Presenter output for web and some dynamite transitions which are very nice. These are also things which are not possible with PTE. Now for PTE's features not available with Producer, Gold or any other commonly available presentation slideshow software. You already know about hardware rendering and image quality. Memories On TV Pro doesn't make executable files, of course so that's one feature. Producer and Gold do, so that's a shared feature with PTE, but of course without high resolution output by hardware rendering. Three of the very powerful features not shared with other programs are parent/child relationships, off-center rotation and object calls from within the program. Let's start with off center rotation. Normally, in a slideshow program images rotate on their center mass axis. So if you choose to rotate an image (let's use a silver dollar as an example) the dollar would "spin" clockwise or counterclockwise so many rotations in such and such a time span. Of course you can also do this with PTE. But PTE goes much farther in that you can move the axis of rotation anywhere you desire. With this feature coupled with unlimited keypoints (keyframes in producer lingo) you can describe circles, bezier curves, elipses, etc. For example, let's say you wanted an object or small image to circle around the perimeter of the monitor rather than rotate on its own axis. You simply move the center of rotation to the center of the monitor and when rotate is used the object or image circles around on the screen. Now let's discuss the parent/child relationship. This feature lets you make any object a child of any other object. This can mean child, grandchild, great grandchild, etc. A "child" takes on the attributes of the parent. So if you rotate the parent, zoom the parent, pan the parent, etc., the child follows this action precisely. But, if you wish the child can also perform its own independent zoom, pan, rotate in conjunction with the animation of the parent. So how does this work in a practical example?? Let's take the above off-center rotation example. When you move the axis of rotation to the center of the screen with the object at the top or sides, the objec rotates around the screen. But when it reaches the bottom the orientation is naturally reversed or "upside down". What if you wanted it to remain upright while rotating? The you could simply duplicate the object as a child of the original, set the rotation for the child to the polar opposite of the parent, make the parent invisible with the unlimited opacity capability and when the object rotates only the child is seen and the orientation stays upright regardless of the position on the screen. This happens because the parent is rotating 360 degrees in either a clockwise or counterclockwise direction and since the child follows this animation, the normal rotation for the child would be as the parent. But if you set the rotation for the child exactly opposite that of the parent without changing the child's default center-mass axis of rotation, the child will mirror the circle described by the parent but rotate on its own center mass in the opposite direction at the same rate as the parent thus the net effect during the circle movement on the screen is zero rotation or stasis in regards to it's center-mass axis regardless of the position of the child object on the screen. Another possibility is to rotate a cube and see all six sides rather than the three sides possible with Producer or Gold. Let's say you want a helicopter to fly level across the screen with the tail rotor rotating normally but with the main rotor blade rotating horizontally? With the parent child relationship it's easily accomplished. You can't do this with Producer or Gold. What if you want to animate a picture and rotate it on not only the X and Y axis independently, but also on the Z axis? This too is possible with PTE. Finally, the ability to use object calls from within the program. This gives you the possibility of running any external executable file from within your slideshow program then returning and continuing on seamlessly with the slideshow. You can easily make your own menus and do things not possible with Producer or Gold such as immediately jump to any slide number from a menu called from within PTE. Pop up a menu slide from an invisible button, click on slide 23 and immediately go to that slide, etc. Control your slideshow manually, automatically or both by a navigation bar. Sure you can pause a Producer slideshow, but try to random access to any slide in the show. Then there is total control over zoom, pan and rotate characteristics. PTE shares llinear and smooth zoom with the others, but it also has non-linear accelerate, decelerate, and smooth which the user can adjust to any curve. So complete control over this between any two keypoints. Also there is "perspective zoom," another feature not found in the others. The bottom line is that Producer and Gold and Memories on TV Pro are essentially programmed approaches while PTE is a low level tool to let the user create their own "program". It's not nearly as feature rich as the other three but in many ways much more powerful from a "low level" programming perspective. Best regards, Lin

Hi Igor, Thanks - I told Fred that this might be the case. I asked him to run dxdiag and see if his bus supported an AGP slot. If so, I gave him a link to a place to buy an inexpensive ATI Radeon 9800 Pro card which should work very well with larger and difficult animations. Best regards, Lin

Hi Andrew, The Cliff Palace pano is a stitch of five images. Right now the current largest stitched image contains over 2,000 images and is 13 gigapixels in size. The first gigapixel stitched image was done by Max Lyons of a scene in Bryce Canyon, UT. Max used his old D60 (six megapixel) Canon to make 196 separate overlapping captures. He then had to write his own software to handle the stitching of the +2.00 gigabyte file. The result was a beautiful 1.09 gigiapixel image which was printed uninterpolated at 300dpi by LightJet technology. The largest print size possible for the LightJet used could not do it in on a single page so separate prints put the entire image together for the final ten by twelve foot print which Max displayed at PMA in Las Vegas three years ago. I spent about forty five minutes examining the beautiful print which was absolutely breathtaking in detail. At the same PMA show another beautiful huge print was displayed by Peter Grote which he took in a single scan with his Betterlight scanning back. This one was taken in a single frame and consisted of over a 500 megapixel panorama file of a scene in the Annapurnas (mountain range in the Himalayas}. Peter was using a motorized pano head. He printed it at an uninterpolated 180 dpi on an Epson 9600 inkjet. The print was about 17 feet by 4 feet and done from roll paper so in one continuous piece. After carefully comparing the prints, Max actually had more detail from his stitched image. So the bottom line is that it is entirely possible to have incredible resolution gained from overlapping multiple stitched images. The problem in displaying them (up to 2 gigabytes in size for all current Windows compatibility) is not PicturesToExe, but the computer hardware. Just stitching these huge images in an overwhelming task. Once they are stitched, the file size must be small enough to be amenable to the Windows operation system being used. For overall compatibility that would be 2 gigabytes of file size. But this would only be possible for an animated (pan, zoom, rotate) still image if the GPU (graphical processing unit) in the Video card could process the image and if the video card had sufficient RAM to handle such a load. With present technology this would only be possible if the images were broken down into bite-sized (byte-sized) pieces. So for performing an "infinite" zoom sort of as you see on Google Earth, you would need to mix different still images ending one with level of zoom in where the next begins. This can be done, but it requires careful mix and match techniques. To a degree, doing a single zoom in on a reasonably highly detailed large panorama is quite possible, but likely to exceed the hardware capabilities of many users. Of course if you are only making the slideshow to be displayed on a projector using your own equipment, then whatever limitations that applied would only be applicable to your own system. Actually, the reason I posted this test is to approximate finding workable limits for this kind of application for general use. Best regards, Lin

Hi John, Thanks! Good to hear that it works well on the Nvidia FX5700 - so far I've only had one response (email) from someone who has had trouble, but time will tell I guess. Best regards, Lin