Best DPI for Motion, and more!

What does DPI have to do with video? Nada. Take a 2000x2000 pic at 600 DPI and compare it to a 2000x2000 pic at 150 DPI in FCP and you will note no difference. The biggest thing you need is pixel dimensions. If your still is 500x500 and 600DPI...it will not look great in FCP when you zoom in on it. Conversly, if a still is 2000x2000 and 72DPI, it will look great.

In my experience.

Shane

I'm working with 14th 15th cen. text/illustrations and I only get to scan them once.

What you should do is get the best/largest quality scan possible and disregard any of the issues of FCP or Motion. If you get a good/large scan, you can resize later if needed for the application you will use. And downscaling in Photoshop or other app should not be too difficult.

I am saying this because of the nature of the material and the one bite at the Apple, as it were, that you will get.

Sort of like if you have one chance to capture some old audio recordings where the tape is about to go, you would not capture at a low mp2 quality.

You can do it in Motion, with After Effects. Having Photoshop to separate the images is a plus.

And DPI is meaningless in FCP land. You want PIXEL DIMENSIONS. 4000x4000 for HD (or higher). 2000x2000 for SD...depending on how far in you want to move.

I haven't seen RIDING GIANTS but i have seen THE KID STAYS IN THE PICTURE and they used After Effects.

Shane

You should scan them at the highest workable resolution, at least 300 DPI. You can always downsample in Photoshop or FCP, but you can't upsample without losing quality.

I've worked on several long-form historical docos and always was thankful we got the highest res scans we could, at least 300, typically 600 DPI. If space is an issue, you can store them as JPEG at 90 or 95% quality.

Once scanned, you have to decide what workflow you'll use to determine what other preparation they may need. Going straight into FCP with hi-res photos will work, but can add unnecessary overhead, its probably best to downsize them in Photoshop to a pixel size just greater than you'll need in FCP. For example, if you are doing a 720p piece and a particular image will be zoomed in on by a factor of 2, you should downsample it so that it will cover 1560 x 1440, which is twice the 1280x720 pixel dimensions of 720p.

If you are going to use one of the pan/zoom plugins or standalone apps, then use the original scan resolution, they will do a better job of downsampling and deflickering if you do and there's no need to prep them in Photoshop.

While I'm at it, I'll put in a plug for my Ken Burns Effect plugin, Pan Zoom Pro. You can get a demo version and info about it here: http://www.lyric.com/fcp-plugins/

Cheers,
John.
Lyric Media

One last thing, it is pixels, but when scanning get the highest DPI and make sure it is optical DPI, not software based (though there are good ones.)

Example, you have an 8 x 10, scan it at the highest optical level at 100% (check your scanner) if it is 300, 600 or more.

If you want the entire page viewed, COPY the image and scale it down to fit in the canvas, but later on when you want to point out a line or paragrah, you will have more pixels and detail to be able to cut sections and get more pixels for the same area.

Also try to save the original scans in a uncompressed format if you can - again I am looking at this as a rare chance to access these documents. Storage is relatively inexpensive (and compared to video storage almost a non-issue 🙂 )

Normally I use FCP to keyframe the animation ... for example, here's what I did earlier this week to make the headline jump off a newspaper clipping:

The client had a newspaper clipping which I scanned (I forget the exact resolution). In Photoshop, I cleaned up the image a bit (old newspaper blotchiness, etc), then using the Select->Color Range command, I selected the black print only. Next, I feathered the selection about 1.2 pixels or so and copied the selection.

Next, I started a new PS document with a transparent background and pasted the selected text onto it. Saved both and imported them into FCP. With the original newspaper image on V1, I animated it (Viewer window, Motion tab) to make it look like it was falling away from the camera. Once it reached a certain point, I placed the new PS image (text only) on V2, scaled it to match the text on the newspaper image below, then quickly scaled up the text image only (V2). I also used a little distortion (keyframed) to give the movement a little 3D effect.

Its kinda hard to explain in writing, but the end result was that it appeared that the text jumped off the page at the viewer ... kind of an 'in your face' type of thing.

And then there are other times I'll scan an image twice; once for what I'd normally use, and the second time, I might scan a smaller area of the image at a higher resolution. One example I've done using this method was a rather large portrait collage of a medical school graduating class. My client was one of the doctors and the project was a docu-tribute to him.

The original photo was one of those that has an oval matted head shot of each graduate that you often see framed and hanging on the wall. The photo was about 16" x 20" so I actually had to do 4 scans (one for each quadrant) and then stitch them together into one document in PS - but I'll count that as one scan for this example. The second scan was of just the one particular graduate, so the scan was of a much smaller area - about 1.5" x 2". Needless to say, I used a higher resolution setting for this scan.

In FCP, I placed the full image on V1; fading in with a slight scale up over the duration. At the point in time I wanted the smaller image to "lift off," I placed it on V2 and scaled it to match the size and movement of the image below. As it "lifts off," I increased the rate of scale while at the same time slightly reducing the opacity and ramp up a bit of Gaussian Blur to the image on V1 as it remains in the background. Then they both transitioned to the next scene together.

Make sense?

After Effects is probably a much better tool for this type of thing, but I'm just now trying to get myself reacquainted with that program.

-DH

DPI, as you say, has nothing to do with video but it is the normal unit used when scanning photos, which is what we were talking about. By recommending 300 or 600 DPI scans, you will get 2K pixel and up resolutions for 4x6 photos and up, which are generally good enough for most panning & zooming tasks.

Strictly speaking, you want to look at the photo, figure what portion you'll frame the closest, measure that in inches, and then divide that into the final video pixel dimensions to get the required scanning DPI. For example, for a 2" section intended for DV, that would be 640 / 2 = 320 DPI (square pixels) scanning res.

Often you have 3rd parties or other agents generating scans, and you have to tell them what res you want, and they typically only understand scanning resolutions which are in DPI.

John
http://www.lyric.com/fcp-plugins/

> You should scan them at the highest workable

resolution, at least 300 DPI.

Sorry, but this is rubbish - and this link will prove it:
http://www.scantips.com/no72dpi.html

Also - READ SHANE's POST above!
DPI is meaningless in Video.

Andy

G5 Quad 8GB. 250+500 GB HDs. G-Raid 1TB. FCP 5.1.1. Shake 4.1. Sony HVR Z1E Mac OS X (10.4.7)
"I've taught you all I know, and still you know nothing".

What are you talking about? Again, when I referred to DPI it was only with respect to scanning of photos, not video resolution. Obviously, DPI has no relevance to video because the physical dimension of the output display is completely variable, but it certainly does when you are scanning a photo which has a fixed dimension.

Since most (all?) scanning software uses DPI as the unit of resolution, it is important to understand the relationship between scanning DPI and it's effect on images you will embed in video. I was merely providing photo-scanning rules of thumb in the first post (a 300 DPI scan will generally give you workable pixel dimensions for images going to video), and in the follow-up email I provided some math that allows you to figure exactly the scanning DPI you should use.

This is all completely consistent with the scantips.com article on 72 DPI, I certainly didn't recommend scanning at 72 DPI but you can't avoid understanding DPI when it comes to scanning photos for video.

Finally, when I said "scan at the highest workable resolution", again this is a rule of thumb. The higher the scanning res, the more creative options you have at edit time, for tight crops, long zooms, interesting pans and so on, because you want to avoid cropping past 100% on the scanned image.

John

http://discussions.apple.com/message.jspa?messageID=3039641#3039641