IMAGE QUALITY DEGRADATION?

[orizaba]

Just a question:

1. As an example, let us suppose we have a slide (or video). In O&A it is with zoom = 100.

2. Let us put it inside a Mask Container, beeing this with zoom = 70

3. Let us put this Mask Container inside a Frame 1 with zoom = 200

4. At last, let us put this Frame 1 inside another Frame 2, with zoom = 120

5. Let us "imagine" that final result for the slide (or video) is an "equivalent" zoom = 168 (just an example).

My question is:

- Is the final image quality 100% equal to the quality we would obtain if we would apply only one zoom = 168 directly to the slide (or video)?

- Or, each time we apply a zoom (mainly a zoom in, as I think) picture quality is reduced?

- This point has to do with usual down sequence processing of several effects (from top to bottom).

So, I think that quality reduces each time a zoom is applied, I mean, PTE does not process a sort of "counter balance" of successive zooms.

Is this so?

Jose

[fh1805]

Jose,

Have you done a test to investigate? In the attached, do you think the third slide is the same quality as the second?

regards,

Peter

Test Zoom Degradation_Jan5-2013_7-32-57.zip

[orizaba]

Jose,

Have you done a test to investigate? In the attached, do you think the third slide is the same quality as the second?

regards,

Peter

Test Zoom Degradation_Jan5-2013_7-32-57.zip

Thanks Peter,

In your example, in a first glance, I can not really see any difference.

I never made a test like yours, in order to compare.

I just should like to know, under a pure technical point of view, if there is any reason for a quality degradation.

I compare this situation with audio hi-fi:

Human ear can detect a maximum 20 KHz, however, we can not notice any difference between 19 Khz and 20 KHz, but difference is really there!

So, I insist: what happens with PTE? Is there degradation or not? Even if we can not notice.

Regards,

Jose

PS: By the way: does "unsharp mask" (in Project options) apply to video, or only to pictures (slides)?

[davegee]

Jose,

Since the same image is zoomed in and out to whatever percentage is required and is never saved how can there be any degredation?

The deciding factor is the percentage zoom at any moment in time i.e. if the (overall) zoom percentage is 150% (given a 1920x1080 image in a 1920x1080 project) then upscaling is taking place but when the zoom percentage is returned to 100% (overall) then the upscaling ceases and the image is represented pixel for pixel.

Downscaling is always preferable to upscaling and Igor said (a long time ago) that PTE is better at downscaling than upscaling.

If you can determine the (overall) zoom percentage that an image is going to be subjected to then it is possible to calculate the exact pixel dimensions to size the image at in order to gain maximum efficiency in terms of project size.

Or you can do as some people do if their computers are powerful enough - feed it with full sized 36Mp Jpegs at the highest quality - you can't do better than that :rolleyes:/>.

DG

[fh1805]

PS: By the way: does "unsharp mask" (in Project options) apply to video, or only to pictures (slides)?

Jose,

The "What's New" for v7.5 made no mention of Unsharp mask. However, the "What's New" document for v7.0 had this to say:

"Unsharp Mask for Sharp Picture

This feature dramatically increases sharpness of picture. PicturesToExe adds unsharp mask for entire screen in real-time, even for slides with animation."

I interpret the statement "...even for slides with animation" as meaning that unsharp mask is applied to each rendered frame.

regards,

Peter

[orizaba]

Jose,

The "What's New" for v7.5 made no mention of Unsharp mask. However, the "What's New" document for v7.0 had this to say:

"Unsharp Mask for Sharp Picture

This feature dramatically increases sharpness of picture. PicturesToExe adds unsharp mask for entire screen in real-time, even for slides with animation."

I interpret the statement "...even for slides with animation" as meaning that unsharp mask is applied to each rendered frame.

regards,

Peter

Peter,

Thanks.

I had never tried to apply Unsharp, neither to slides nor to video.

Just now, I applied Unsharp = 200 (maximum) to video and difference is great!

Such video is an old Hi8 video, which I digitalized (AVI file).

Original quality in terms of sharpness is not so good, mainly when we are already used to Full HD quality... so, this Unsharp Mask helps a little bit to quality image.

Regards,

Jose

[orizaba]

Jose,

Since the same image is zoomed in and out to whatever percentage is required and is never saved how can there be any degredation?

The deciding factor is the percentage zoom at any moment in time i.e. if the (overall) zoom percentage is 150% (given a 1920x1080 image in a 1920x1080 project) then upscaling is taking place but when the zoom percentage is returned to 100% (overall) then the upscaling ceases and the image is represented pixel for pixel.

Downscaling is always preferable to upscaling and Igor said (a long time ago) that PTE is better at downscaling than upscaling.

If you can determine the (overall) zoom percentage that an image is going to be subjected to then it is possible to calculate the exact pixel dimensions to size the image at in order to gain maximum efficiency in terms of project size.

Or you can do as some people do if their computers are powerful enough - feed it with full sized 36Mp Jpegs at the highest quality - you can't do better than that :rolleyes:/>/>.

DG

DG,

Thanks for your explanations.

Anyhow, I think I am not still convinced...!

You say "... zoomed in and out to whatever percentage is required and is never saved how can there be any degredation?"

This is the point:

You zoom in. Result is "degredation" (even if you don't save the project at this time), is this so?

After this, you zoom out (even to the original percentage). Image which will be zoomed out is not the original one, but it is already a degretaded one instead, is this so?

I say this because PTE (and almost all video editing softwares) execute successive "operations", one after the other, according to "up-to-down" order, is this so?

This is my doubt... Can you clarify?

Of course, I always use original picture size (36Mp), but problem (and doubt) above still exists.

Regards,

Jose

[davegee]

Hi Jose,

I think that we may be misunderstanding one-another?

Any image used in PTE "should" be big enough to cope with the amount of zooming required.

If you use a 1920x1080 image in a 1920x1080 project and apply 200% zooming to it then you WILL get "degradation" or "interpolation effects" during the Zoom. But if you zoom back out to 100% you will, once again, be using the original image at its original "pixel-for-pixel" resolution with no lasting side effect.

If you use a 3840x2160 image in a 1920x1080 project and use "fit to screen" you are applying a 50% zoom to it. You can then zoom to 100% without any interpolation effects being evident. Anything over 100% will cause interpolation effects (degradation) but it will recover if you use a zoom value between 50% and 100% again.

If you did that in an image editor you might not get away with it but PTE is not an image editor.

I hope that I am understanding your question?

DG

[Lin Evans]

Hi Jose,

Actually, PTE uses the original image to do everything with. If you zoom in from say an original 1024x768 to twice it's original size there will be image degradation, of course. However, if you subsequently zoom back out to the original 1024x768 there will be zero image degradation because the original image is always being used, not the degraded zoom. PTE only keeps the copy of the image you load on your slide. All actions including zoom out, zoom in, etc., are done via this image. The zoom-in image is simply being returned to the original in much the same way it is done in Photoshop. You do not loose any quality by zooming out, zooming in, etc., because it's all being done via the original.

Best regards,

Lin

[xahu34]

I am pretty sure that Lin is right. An image is zoomed and rendered only once. Nested frames and containers above have only influence on the mathematics.

Regards,

Xaver

[davegee]

Lin,

I am not sure that the Photoshop example is correct? If you make an image 150% bigger you are introducing intermediate pixels. If you then return the image to 100% I am not sure that you would necessarily lose all of the exact pixels that you gained in the upscaling process?? Something to think about?

DG

[orizaba]

Lin,

I am not sure that the Photoshop example is correct? If you make an image 150% bigger you are introducing intermediate pixels. If you then return the image to 100% I am not sure that you would necessarily lose all of the exact pixels that you gained in the upscaling process?? Something to think about?

DG

DG, Lin and Xaver,

Thanks so much, all of you understood my question quite well, and now everything is clear for me!

I mean... almost...!

Regarding PTE methodology, as DG and Lin explained, things are quite clear for me and I agree: no degredation. Period!

But this last post from DG, concerning Photoshop process...

Does this mean that Photoshop does not follow the PTE process? Always using the original picture?

I do not work with Photoshop, so I am away of this problem, but I would find curious (at least) that Photoshop use "a less clever process" than PTE! Even beeing PTE not an image editor.

Regards,

Jose

[Lin Evans]

Hi Dave,

But that's not what's actually happening. You are not returning "the image," you are zooming out the zoomed-in image until the original (which is the only image actually stored inside the program) is displayed. The image after the zoom in and return to regular 100% size (assuming you began with the original) is the "identical" image you started with. All intermediate images are math computations derived from the original. When the zoom is returned to 100% the original is again displayed. The original is never actually changed in any way. Likewise with Photoshop. Photoshop loads your original and any zooms out or in have no real essence until a save is done. The original remains as you loaded it unless you overwrite.

Dave, from a programming perspective inside both PTE and Photoshop and other image editors other than Photoshop are electronic "pointers" referencing the original file. This is how they know where it all starts and/or ends. So when zooms are done and the pixel dimensions reference the original, the actual original is returned as the value.

Best regards,

Lin

[orizaba]

Hi Dave,

But that's not what's actually happening. You are not returning "the image," you are zooming out the zoomed-in image until the original (which is the only image actually stored inside the program) is displayed. The image after the zoom in and return to regular 100% size (assuming you began with the original) is the "identical" image you started with. All intermediate images are math computations derived from the original. When the zoom is returned to 100% the original is again displayed. The original is never actually changed in any way. Likewise with Photoshop. Photoshop loads your original and any zooms out or in have no real essence until a save is done. The original remains as you loaded it unless you overwrite.

Dave, from a programming perspective inside both PTE and Photoshop and other image editors other than Photoshop are electronic "pointers" referencing the original file. This is how they know where it all starts and/or ends. So when zooms are done and the pixel dimensions reference the original, the actual original is returned as the value.

Best regards,

Lin

So, this clears my only doubt (as per my last post).

Thanks,

Jose

[davegee]

Lin / Jose,

When Lin mentioned PS I thought he was refering to re-sizing. If you re-size an image and then reverse that process you cannot be sure of getting back to where you started (other than by using the History Pallette). Zooming into an image is different - no new pixels are made.

DG

[xahu34]

...

I do not work with Photoshop, so I am away of this problem, but I would find curious (at least) that Photoshop use "a less clever process" than PTE! Even beeing PTE not an image editor.

...

You should not try to compare PTE and Photoshop. Photoshop has parallel layers all of which have the same coordinate system, and there is a particular rule how to evaluate a tower of several layers (including blending modes), and this is the model we need in case of image manipulation.

PTE, having the parent/child mechanism, uses a scene graph model. Each image object has a coordinate system that sticks to this object, and which serves for describing the positions and the relative animations its children. The animation parameters of the object itself refer to the sticky coordinate system of the object's own parent (the screen being the parent of all top level objects). Systems of that kind are used by programs like computer games, and the situation of nested coordinate systems seems to be well supported by DirectX.

Regards,

Xaver

[Igor]

Jose,

I agree with Lin's explanation and with answers of other members.

Image/video object can be placed into hierarchy of embedded containers (frames) with different zooms. PicturesToExe will calculate final (visible on the screen) zoom value. So there will be no losses in quality.

This is true even for mask containers.

[orizaba]

Jose,

I agree with Lin's explanation and with answers of other members.

Image/video object can be placed into hierarchy of embedded containers (frames) with different zooms. PicturesToExe will calculate final (visible on the screen) zoom value. So there will be no losses in quality.

This is true even for mask containers.

Igor,

Thanks, that is good.

That is even better because I already noticed, since long time ago, that quality of PTE zoomed images is far better than same zoomed pictures using a video editor like SONY VEGAS or ADOBE PREMIERE, which I also use.

You know, as you already understood, I work with film and video since more than 50 years.

When PTE introduced video in slideshows I started using PTE to "edit" my films and videos! No more, no less!

With PTE I can frame, zoom, fade, etc., with better quality, that's my conclusion.

That's why I am very sorry indeed that PTE is not a native 64-bit program... I told you before.

Because 32-bit is very much limitative and forces me to have a lot more of a long (and stupid) work not to follow in an out of memory situation.

Please, tell me that you will produce a native 64-bit version VERY, VERY SOON! I have dozens of films and videos to edit... and I am not so young...!

Best regards,

Jose

PS: By the way, can you provide exact parameters of your PTE video converter? I mean, video and audio codecs used, bitrates, etc., which real values - of what? - correspond to quality percentage (from 0% LOW, to 100% HIGH), etc..

At, last, a suggestion: is it too difficult to introduce in PTE (video section) an immage stabilyser? Almost all video editors have this feature. My experience shows that SONY VEGAS stabilizer is far better than any other, it has 3 stabilization degrees and it is very much effective in each degree. This feature in PTE would avoid at least 1 video conversion and consequent quality loss, because, as you say, stabilized immage should be the original one, and after zoomings, fades, etc., it would be rendered only once (on the PTE converted video file). Is this so?

[Igor]

Jose,

I'm very impressed that, as you wrote, PicturesToExe provides better quality than those famous video editors!

However I can't promise that 64-bit version of PicturesToExe will come soon, sorry. Just because we have too much old code written many years ago. All I can say that we spend 20% of our time to rewrite old code to make it cross-platform (for future Mac version) and also it will open way for 64-bit version.

Our more nearest plans are to optimize memory usage that will help a little in that situation with "Out of memory" problem.

Also probably you noticed another topic on the forum for same problem ("Out of memory") in executable shows. But indeed this problem should not occur there. I'm sure these shows don't exceed 2Gb limit of memory and a reason of that problem is different. I have a guess that this problem and yours are linked. Once we solve one problem, it should help for another problem, too.

P.S. Sorry for my complicated explanations, my English is not perfect yet.

[orizaba]

Jose,

I'm very impressed that, as you wrote, PicturesToExe provides better quality than those famous video editors!

However I can't promise that 64-bit version of PicturesToExe will come soon, sorry. Just because we have too much old code written many years ago. All I can say that we spend 20% of our time to rewrite old code to make it cross-platform (for future Mac version) and also it will open way for 64-bit version.

Our more nearest plans are to optimize memory usage that will help a little in that situation with "Out of memory" problem.

Also probably you noticed another topic on the forum for same problem ("Out of memory") in executable shows. But indeed this problem should not occur there. I'm sure these shows don't exceed 2Gb limit of memory and a reason of that problem is different. I have a guess that this problem and yours are linked. Once we solve one problem, it should help for another problem, too.

P.S. Sorry for my complicated explanations, my English is not perfect yet.

Igor,

Your English is perfect, explanations not complicated, thanks for all your information.

What about my own P.S.?

Regards,

Jose

[Igor]

PS: By the way, can you provide exact parameters of your PTE video converter? I mean, video and audio codecs used, bitrates, etc., which real values - of what? - correspond to quality percentage (from 0% LOW, to 100% HIGH), etc..

At, last, a suggestion: is it too difficult to introduce in PTE (video section) an immage stabilyser? Almost all video editors have this feature. My experience shows that SONY VEGAS stabilizer is far better than any other, it has 3 stabilization degrees and it is very much effective in each degree. This feature in PTE would avoid at least 1 video conversion and consequent quality loss, because, as you say, stabilized immage should be the original one, and after zoomings, fades, etc., it would be rendered only once (on the PTE converted video file). Is this so?

Jose,

1. The idea is really interesting, but I can't answer now. Please create a new topic for image stabilization in "Ideas and suggestions for next versions".

2. Audio code is AC3, 256 kbit.

3. For video we use "Quality" parameter which PicturesToExe calculates to bitrate taking in attention width & height of video.

When you save JPEG images you also use "Quality" paramter. With same "Quality" value you get different file size of JPEG for large and small images.

You can use video editor to convert video to MPEG4 video file for best playback in PicturesToExe.

When you include video file to EXE file, PicturesToExe does not convert anymore source video file. We add all effects, 3D, mixing with still images on the fly in real time.