Problems With Still Photographs In PE-11

[Avagadro1]

I’m new at video, but an old-timer with Photoshop (the full version). I have a 64-bit Windows 7 set up.

I just purchased Adobe Premier Elements-11 and Steve Grisetti’s book. My projects will all be standard definition, converted from VHS, but for now I’m merely practicing, practicing, practicing. Unfortunately, I’m running into TWO SEPARATE roadblocks when importing still photographs into my test projects.

PROBLEM 1:
I read very clearly that Steve advises no photos greater than 1000 x 750 pixels. Of course he is also quite intuitive when he assumes this will at first be perplexing to an old Photoshoper who works only with 16-bit-dept PSD files, but I get it about video being different than Photoshop’s high-ppi creating continuous-tone images. So I took a 5"x7", RGB, 16-bit, 360ppi, image in portrait mode and downsampled it (in Photoshop) to 1000 pixels at its height (of course lower at its width), RGB, 8-bit. I tried three formats: PSD, tiff and jpeg, as Steve’s book states that all three will work in PE-11.

When I imported the PSD and dragged it to my PE video time line, the image in the Monitor Panel appeared somewhat blurred and distorted. Well, being new at this, I’ve noticed that the program sometimes has a pop-up telling the user that “Rendering” may make things better. Although I have no idea what “rendering” actually does, I nonetheless rendered. The program crashed.

Then I tried the tiff. PE popped-up a failure/error dialog that the image could not be imported.

I then tried the jpeg. This imported without a problem. However, the image in the Monitor Panel was still not as crystal clear and sharp as in Photoshop or a jpeg viewer. So I rendered (I don’t know why I rendered, but it seemed the thing to do.) No crash.

So, accepting that I possess a very good knowledge of Photoshop and still-image techniques but little or no knowledge of video or PE-11 -- despite having read and reread pages 54 -55 of Steve’s book -- can someone help me to identify my failures and misunderstandings, and more generally to help me to better utilize still photos as part of PE-11 projects? Perhaps there is a distinct thread on either Muvipix or Adobe Forums that I have not myself found, that focuses on these issues.

PROBLEM 2:
I read Steve’s discussion on page 56 of his book to TURN OFF the “Frame Size” function.
When my 5" (width) x 7" (height) photo (in portrait mode, previously downsampled to 1000 pixels in height), appears in the Monitor Panel when the “Frame Size” function is ON, the image properly appears as a proper, rectangular 5x7. When I turn off the “Frame Size” function, in the Monitor Panel both the top inch or so and the bottom inch or so of the image are cropped away, thereby making the remaining image appear square rather than rectangular, while losing the important top and bottom areas of the photograph. If go back and turn “Frame Size” back on, then the image properly appears as a rectangular 5" (wide) x 7" (high) photograph.

Since Steve’s advice was clear, and turning the function of or on is quite easy, I must be missing something. Any idea what that might be.

Thank you,

Howard


HOLD IT. WAIT A MINUTE. I'M EDITING MY POST.

Steve's book states to resize photograps to no more than 1000 pixels x 750 pixels.
Since my project (though it is just a test project until I learn the program better) is a standard definition NTSC-based project, should I instead have downsampled my photographs to 720x480? Could this be the source of some of my problems in dealing with still photographs in PE-11, or is there more to this matter? I don't know why this would be the source of my problems if Steve's recommendation allows for up to 1000x750 pixels knowing that NTSC is 720x480, but I am at a loss.

One more consideration. After some further research, I am now aware of a function that I was not previously aware existed in Photoshop: changing the pixel aspect ratio (PAR). I just read that VIDEO (such as NTSC 720x480) uses a PAR of .91, while all of my Photoshop images have a PAR of 1. Photoshop has a function under the "View" menu to change the PAR to the .91 video standard. Converting any of my Photoshop images to PAR .91 makes then appear taller and narrower (a distorted view) in Photoshop. Nonetheless, for any image that I want to import into Premier Elements, should I first change the PAR to .91? ALTERNATIVELY, should I not do this, perhaps because PE-11 does this automatically in the background anytime a "regular" PAR 1.0 image is imported into PE?

All quite confusing.

Thanks again.

[Paul LS]

Yes, images brought into Premiere Elements can be rendered in order to improve the preview. Note Premiere Elements 11 has a quick render mode (probably default) that will not give as good quality as the longer render. That said once exported you should not see any issues. I am surprised you are having issues rendering PSD files, I use them all the time.

Regarding scaling to frame size, if you turn this off then the photo will go in at its native resolution superimposed on the project frame size. As your project frame size is smaller than you photo resolution the photo over fills the screen. You can use under Properties: Motion.. Scale & Position to manually scale the photo to fit the frame or to pan and zoom around the photo.

1000x750 is perfect for a standard definition video as it allows you to pan and zoom around the frame as described above. if the photo resolution matches the project frame size then any zooming into the photo would mean you are digtially zooming and thus reducing the quality.

Regarding PAR.. bring the photos in with a PAR of 1... Premiere Elements will automatically take care of the PAR for the video.

[Steve Grisetti]

Great answers, Paul!

The reason we recommend you use photos at 1000x750 is that this keeps the photos small enough that the program doesn't implode trying to downsample the images, yet it gives you a little extra, if you'd like to pan and zoom a bit.

If you don't intend to pan and zoom, you can go to Applied Effects for the clip, open the Motion properties and set the Scale to 67%. Or you can just turn on Scale to Frame Size and not mess with scaling at all.

There are actually two factors that affect how your photos look in the Monitor panel. One, as Paul says, is whether or not you've rendered your timeline. The other is the quality level you have set for your Monitor. You can set it to Highest quality by right-clicking on the Monitor and you'll get a cleaner, clearer look at your video.

Although, ultimately, neither rendering your timeline nor changing the quality level on the Monitor has anything to do with how your final output will look. If you've optimized your photos and, optionally, right-clicked on each photo and set Field Options to Flicker Removal, your output will give you results that look as good as a 720x480 slideshow can possibly look!

[Avagadro1]

Thank you both for your responses. I shall experiment using every suggestion you've offered.

There is an extensive discussion of the Adobe Premier Elements Forum about using still images in PE. The primary responding expert was a fellow named “Hunt” (obviously very knowledgeable). On sub-subjects regarding Adobe Premier and NLE, I struggled to fully understand (as did most of the posters). But as the discussion related to still images, pixels, resizing, etc. -- what we may call Photoshop matters -- I understood every word, and then some.

I would summarize the very extensive discussion as follows:
Use Photoshop to resize all still images PRIOR to importing them into Premier Elements (or Premier Pro, for that matter). For standard definition video, resize to 720x480 pixels, but even where zooming is expected, no more than 1000x750 pixels. In one post there was a link to an instructional video for batch resizing multiple images to the desired dimensions. Also strongly advised was to keep images in the native Photoshop format: “PSD” files, notwithstanding the larger size as compared to jpegs.

But one sticking point remained, at least for me. The expert in those posts to the Adobe Premier Forum repeatedly stated that, in Photoshop, he also changes the pixel aspect ratio (“PAR”) of still images from 1.0 to .91, as the latter is the video standard. When I practiced this in Photoshop, it, of course, distorted the images, making them taller and narrower than proper. I thought that perhaps I should not change the PAR in Photoshop as it is something that PE does automatically upon importing a still image. But the expert (Hunt) repeatedly stated that he downsizes still images and changes the PAR to .91 while in Photoshop, and prior to importing into Adobe Premier. By contrast, the person in the resizing tutorial mentioned above, though he clearly explained the process of automated resizing of still images to properly fit the Premier time line, said nothing about changing the PAR in Photoshop to 0.91.

Were I Arnold Palmer, I’d have no problem with PAR. But I’m a weekend duffer with a handicap who expects not Eagles or Birdies, but would like a correct PAR now and then. More generally, I simply would like still images to be properly presented in PE, even if a little work in Photoshop is a necessary precursor.

Thanks,

Howard

[Bob]

OK, let's see if I can help clear up some of the confusion regarding PAR.

Yes, you read correctly, NTSC standard definition video uses a frame size of 720x480 pixels. That's true whether you are using full frame 4:3 video or 16:9 widescreen video. The difference between the two is the PAR being used. Full frame video uses a PAR of about 0.91 (the pixels are skinny) while widescreen video uses a PAR of about 1.21 (the pixels are fat). Still photos, for comparison, use square pixels with a PAR of 1.0.

When you use a still photo using square pixels in your video project, Premiere Elements will automatically convert it to one using rectangular pixels with the appropriate PAR for the project. You only need to resize the photo to fit into the video frame (or slightly larger if you plan on panning and zooming). This is probably the easiest way to include still photos into your video projects.

You can, if you desire, work with rectangular pixels in Photoshop. When you create a new document, you can choose from standard video presets that have the correct dimensions and PAR already applied. These documents will also have guidelines present representing the Title Safe and Action Safe areas. Once the new document has opened, you can create your titles or artwork as desired. You can also convert an existing photo with a PAR 1.0 to a video PAR, but it's not that straight forward.

When you change the PAR using the View>Pixel Aspect Ratio... command, you aren't actually changing the pixels. You are only recording a numerical value for the PAR. It's similar to the assign profile command. When you use assign profile, can can say the document is sRGB, for example, but the color values of the pixels haven't actually been changed. You need to use the convert to profile command to change the color values to the named profile. However, there is no corresponding command to change the pixels from one PAR to another. When you applied a PAR of 0.9 in your test, the pixels were not changed and that is why the image appeared squeezed with Pixel Aspect Ration Correction on and normal with correction off. Had the pixels been changed, the image would have appeared normal with Pixel Aspect Ratio Correction on and stretched when the correction was off.

One way to convert the Photo properly is to create a new document using the desired video preset and then use the edit>place command to bring in the photo. The photo will be automatically resampled and converted to the document PAR. You can also drag the photo and drop it on the new document to do a conversion.

If you don't want to create a new document using a video preset, you can modify the existing document. But, that involves unchecking the constrain proportions checkbox on image size and manually stretching or squeezing the horizontal dimension the appropriate amount. Not recommended. Adobe does provide an action to do that for you. If you open the actions panel and click on the menu button, you can select "Video Actions". That will load a number of video related actions. The one titled "DVD Slideshow NTSC Standard" will convert a square pixel photo into a resized photo with PAR 0.9 that will fit the video frame. Note that this action is a bit dated in that it uses the old PAR of 0.9 instead of the newer 0.91 value, but it should be close.

If you use the action on a background image, it will fill the transparent portions with black. If you want to maintain transparency, you need to convert the background to a normal layer before running the action. You will need to save the file as a PSD, flattening the image or saving as a jpeg will lose the transparency.

[Avagadro1]

Thanks Bob,

Though I’m still in the practicing/learning stage of PE-11 and video generally, using Steve’s book I had been moving along. Then came the matter of importing still images into a video project. Per instruction, I Unchecked the “Scale to Frame Size” option and from then on everything went to hell. After several days of research I believe the problem(s) were, and the solutions are, as follows:

I need to import some 5x7 portrait photographs into test video. In Photoshop I downsampled the photos to 1000 x 750 pixels, per ubiquitous suggestions. When imported into PE-11 with “Scale to Frame Size” unchecked, only the central portions of a photograph showed, and blurry and distorted at that. After untold hours on the Internet and several stories on how still images should be downsampled in Photoshop to the ACTUAL scale of one’s video setup, in my case standard definition 720x480 NTSC, it occurred to me that I should not downsample the original scanned image to 1000 x 750 pixels, but to 720 x 480 pixels, since I’m not planning any zooming. So I began afresh and downsampled the original image to 720 x 480. Importing the image into PE-11, the result was still no good. Then it hit me: notwithstanding these being images intended to be printed as 5" wide and 7" tall, in PE-11 they must be NOT 720x480, but 480 pixels high and whatever width that falls to. I started from scratch and downsized the image to 480 pixels high by whatever the width. Imported into PE-11 with “Scale to Frame Size” turned off — a perfect fit, finally! The image was still not sharp as a tack, but I’m hoping it will show well in the final video.

What I may have learned (though I’m still not sure) from the various places I visited was that --

1) Still images should NOT be imported into one’s video program until AFTER they have been resized in Photoshop to the intended dimensions (Photoshop has a much better engine for such purposes than does Premier - Elements or Pro, and doing so also takes the burden of the video program.);

2) If no zooming is intended, it serves no purpose to have any still image larger than 720 x 480 pixels in a standard definition video. As an old Photoshopper this took some time to appreciate. ANY THOUGHTS?

3) In PE-11, the height dimension must be the 480 pixels, not the 720. So for my portrait photographs in portrait orientation, the maximum pixel count in the vertical dimension should be 480. This was a bit of a shocker to an old Photoshopper whose original image was 2100 pixels high (7" x 300 ppi).

What this still leaves is the following:

1) I do not understand the commonly observed 1000 x 750 maximum pixel count for PE-11 (if no zooming is intended). If there are 750 pixels in the height dimension and one unchecks “Scale to Frame Size”, how could such an image correctly fit into the PE-11 Monitor Panel, which is expecting no more than 480 pixels? I've got to be missing somethng here.

2) In PE-11, my still images are not tack sharp as they were in Photoshop. Perhaps they will look better in the final video. We’ll see.


Final point for all Premier Pro or Premier Elements tutorial creators or instructional book writers: If the intended audience is beginners like myself, there should be a FULL and SEPARATE tutorial or chapter on preparing still images and incorporating them into a video project. What’s out there now on this limited, but very important, aspect of the video-making process seems wholly inadequate.

Thanks again,

Howard

[Bob]

What this still leaves is the following:

1) I do not understand the commonly observed 1000 x 750 maximum pixel count for PE-11 (if no zooming is intended). If there are 750 pixels in the height dimension and one unchecks “Scale to Frame Size”, how could such an image correctly fit into the PE-11 Monitor Panel, which is expecting no more than 480 pixels? I've got to be missing somethng here.

Pan and zoom is very popular. 1000 x 750 pixels is merely a guideline to assist users in resizing their photos for that purpose. If you don't intend to pan or zoom the image, you may resize it to fit within the video frame any way you see fit.

The recommendation to uncheck "Scale to Frame Size" is intended to help prevent image degradation. There are two ways to scale a photo in Premiere Elements: "Scale to Frame Size", and the Motion Scale property as described earlier in this thread. "Scale to Frame Size" is actually a pre-scaling that takes place before any other scaling. If you import a larger photo intending to zoom into it and "Scale to Frame Size" is on, the photo is resized to a smaller image which will fit the video frame. If you subsequently try to zoom back out to the original size as imported using Motion Scale, you will be scaling that smaller image up and the image will degrade and become soft or pixelated. If you are going to be zooming into an image, you need to turn "Scale to Frame Size" off. With it off, you can do any desired scaling with the Motion Scale property and you will be scaling the original size as imported.

2) In PE-11, my still images are not tack sharp as they were in Photoshop. Perhaps they will look better in the final video. We’ll see.

Standard and high definition video conform to television standards. Standard definition for the NTSC system is 720 x 480 pixels. That's what is displayed on your standard definition TV and is what DVD uses. High definition is what is displayed on HDTVs and is what Blu-ray discs use. High definition can go up to 1920 x 1080 pixels. In either case, the resolution is far less than the digital images from your camera. To judge the image quality, you need to view it on the intended display device -- a television. Viewing on your pc monitor, especially if you view standard definition enlarged to full screen monitor size, is not going to look as good.

[Avagadro1]

Thanks Bob,

Yes, I figured I had to see the image on a TV screen before judging whether a photographic image imported into PE-11 was "good". Since I'm still learning and practing, I've not yet created any DVDs.

Separately, your excellent explanation (above) seems to confirm the concept that PE-11 is not the best route for resizing an image, but rather it should be resized first, by Photoshop or similar, then imported into PE-11. My understanding is that if I resize in Photoshop and downsample the height dimension to EXACTLY 480 pixels (and allow the width to fall wherever, so long as it does not exceed 720) then PE-11 will do NO RESIZING AT ALL, leaving still images as sharp as possible. Again, that's my understanding, but I've not yet seen my final product.

Thanks again,

Howard

[Bob]

I always use Photoshop to resize my large out of camera images to video resolution prior to importing into Premiere. This is mainly an issue when the differences in size are large. The scaling in Premiere is perfectly fine for the relatively small changes consistent with typical zoom amounts. Just remember, if you plan on zooming someday,use a larger image for those photos so you don't have to scale greater than 100%.

One more comment about resizing. If you import a square pixel image into a standard definition video project, it will be resized even if the pixel dimensions are correct for the videoframe. That's because the pixels in standard definition video are rectangular, not square. Since your photos use square pixels, they will need to be asymmetrically resized with the horizontal dimension squeezed or stretched depending on whether the video frame is widescreen or fullscreen. If you create a new document using the video presets and use the Place command to scale and position the photo in the document, the asymmetrical scaling will be done then and the saved file can be imported into Premiere as is with no additional resizing.

[Avagadro1]

Ah, I never thought about the square vs rectangular pixel issue.

So my exacting efforts to downsample a still image in Photoshop (using the bicubic method) such that height equals 480 (and width is anything under 720), or such that width equal 720 (and height is anything under 480) is NOT going to avoid resizing in Premier. On any existing Photoshop image having squate pixels Premier must nonetheless scale the image to create the .91 PAR.

For those well versed in Photoshop techniques, this raises an issue. In the Photoshop world, after upsampling or downsampling an image, one MUST apply sharpening (via the "unsharp mask"). If that image is imported into Premier and Premier then does its own resampling -- for whatever the reason -- are the still images therein less than well sharpened?

Thanks,

Howard

[Bob]

Yes, resampling will soften the image slightly. That's not always a problem, though. Standard definition is always interlaced and that can be a problem if the image is too sharp. Fine detail on the order of a pixel in vertical size can flicker on and off as the alternating fields display and that can be very annoying. Flicker is normally solved by blurring the image slightly. Try a few typical images and see if the loss of sharpening is an issue. If it's an issue, use the method where you Place the image in a new document with a video preset. You can sharpen that and it won't be resized when imported as it already at the proper dimensions and PAR.

[Avagadro1]

Ah, your suggestion also explains why PE-11 has a "Flicker Removal" function. I guess that function slightly blurs the image.

As I mentioned in an earlier post, dealing with still images in a video-editing program deserves a tutorial or book chapter all it own!

Thanks again, Bob.

[Steve Grisetti]

For what it's worth, I dedicate a large section of our books to that very subject!

We've also got tutorials here at Muvipix on creating slideshows in Premiere Elements that includes these tips and suggestions.

[mfl1]

I had the same issue with stills in portrait orientation being far too tall for my videos (even if I intended to pan and zoom), and I appreciate the suggestion that, rather than using "Scale to Frame Size" to shrink them, it's better to use Applied Effects/Motion/Scale. I find the scale slider to be very awkward to use, though: it's very hard (and time-consuming) to move the slider to precisely the right point. Am I missing something? Is there an easy way to use the slider (as there is in photo processing programs)? Thanks.

Marvin

[Steve Grisetti]

You can also just manually type in the Scale percentage -- which is much easier than using the slider for precise settings.

To fit a portrait photo that is 750x1000 pixels in size into a standard definition video frame -- or to fit a 1500x2000 portrait photo into a high-definition frame -- just set the Scale to 48 to 50%.