Best Graphic Import Resolution

I need to show a lot of paper documents in my video - (showing the entire 8.5 x 11 doc, zooming into them, zooming into certain words, 60minutes style)

I scanned these documents at a high resolution and want to save them and set them up ready for import into FCP. What resolution should I be saving these as, and what format? They are all black and white so should I be converting to greyscale for any reason? I'm looking for nice crisp text on my NTSC monitor (of course) weather I zoom in or zoom out. Is there a good res that will cover my needs?

FCP 4.5. Standard Def.

Thanks!

I don't know what "60 Minutes style" means, but in my experience, it is more impactful to do a 3D Camera move on the paper and instead of zooming into a word, have that word or sentence fly off the page and out to a CU while dimming the bkgd. If you are going to zoom into a "word" on a document, that document will have to be scanned in at a VERY large size (3 to 5 times larger).

- Joey



When life gives you dilemmas...make dilemmanade.

If you're looking for the resolution setting for your initial scans, think more in terms of overall image size in pixels. Screen resolution is set at 72dpi, so you wouldn't think you'd need to scan at a high dpi like you would for print (300dpi +). HOWEVER, the higher dpi you scan in, the more detail you'll have if you enlarge the documents on screen.

Assume you're going to keep the scale at 100% (I never enlarge on the scanner if I can avoid it, to reduce artifacts). You'll get more detail if you bump the resolution: an 8.5x11 doc scanned at 72dpi is only 612x792 pixels, but the same doc scanned at 300dpi is 2550x3300 pixels. To fill the screen with only one word you'd have to enlarge the 72dpi version a lot, turning it into mush. But the 300dpi image will be big enough that you only have to zoom in just a little and the image will look better. I would do some test scans at different dpi, then do your test animations and see which resolution comes out the cleanest after the image is enlarged. Try to narrow it down, as too much resolution will be overkill and you'll wind up with lots of big files.

HTH,
JK

I guess what I'm asking is what res should I import them as. I know 72 dpi. Then in FCP - 100% (actual size) at 72dpi renders the nicest image. Will say scaled to 42.3% render the nicest image also? or should I reimport a different file closer to the final size I need it in FCP?

Remember reading that most efficient size is 1800 by 1350 at 72 dpi according to FCP guru Larry Jordan who's done some experimenting -- that's what I've used for a while and it works well for slow zooms into a subject without becoming pixilated and at the same time not slowing everything down when you get to that graphic in the timeline. Andy

Logan,

Scaling down to a smaller size is not a problem. It's when you enlarge past 100% that the image starts to fall apart. If you really want to zoom in on single words as you said, you'll need a larger image.

PS very good article by Jordan at this site

[www.bosfcpug.org]

Yes, that's a good article. Make sure you scroll down to the comments below to read the discussion about zooming in on graphics, which is not covered in the main article text:

"Also, if you want to zoom in, you need to import an image at a larger size than full-screen so that you have room to zoom without pixelization. For me, if the image is full-screen with no moves, I import it at 720 x 540 x 72. If I plan to pan or zoom, I import it at 1440 x 1080 x 72. This allows me to zoom to double the image size without problems. Image quality is good."

The Larry Jordan comments quoated above are of course correct. but also note that if you create and import in the resolution native to the sequence in which you are working, and aren't doing any motion on the GFX, no rendering will be required.

Format is sometimes an issue too. I find Tiff to work best, sometime JPG works OK, but the lossy compression can make an image noisy and just bad looking.

72 dpi is not really a hard and fast rule for creating the GFX - it is the resolution of the output, but you're under no obligation to create graphics in that res. If you create them at 300 dpi, for example, you'll just end up with a larger and more detailed graphic file (probably one which is way larger than you actually need, and which you'll probably have to make smaller in FCP, though, and will likely require a render). It becomes a sort of relative thing when one gets into it.

I find knowing Photoshop and how it can interact with FCP indispensible for this. Knowing how you can use Image size and Canvas size, and applying batch actions can save an enormous amount of time.

Playing around with different settings when you make the GFX may be able to yeild results which suit you better than generic ones, and will slip right into the timeline with little or no effort.

btw - I'm just finishing up a lecture piece that's over 7 hours in length and has over 800 powerpoint slides in it (I'm afraid to do a final count).



HarryD

You might want to look in to this as well

[www.ampede.com]

"Ampede PDF is a fully-integrated Final Cut Pro plugin that renders vector content from Adobe Illustrator(1) and PDF files frame by frame, so the image is always sharp regardless of scale, movement, or rotation. Vector content can be seamlessly mixed with QuickTime content, and the geometrics are fully keyframeable using the standard Final Cut Pro interface you already know how to use."

Hi

I admire Larry's work - his seminars, newsletters and the new magazine bring a lot of help to many of us (I think Larry knows).
This article is great and very helpful. But he is wrong with the "video forever and ever is 72 dpi" and this dpi discussion may confuse people in connection with video. "dpi" is the abbreviation for "dots per inch". Video does not have any "dpi", otherwise we may got a preset like "HD 1920x1080 50i @ 300 dpi for a 19" monitor with effective resolution of 1280x720".
Simple example: working with SD video (720x486) this video will have pretty close to 72 dpi on a 14" broadcast monitor with 800 lines resolution, using a 20" monitor with the same resolution would lower the dpi to around 50 dpi, taking a 14" consumer one with around 320 lines of rez would give another dpi, looking at the viewer of your camcorder again another one, the FCP viewer will give you a bunch of different dpi, depending on you monitor and so on.
Video has a fixed resolution depending on the format you're working on.

BUT when you do scans (or create artwork) you have to take into account the size of the artwork, the kind of the artwork (digital photo, vector, print, photo), the final size within your video and the pixel aspect of the video format you're working with - and when doing scans that will have something to do with scanner "dpi" settings to achieve the final resolution of X by Y to fit the video setup, but - again - nothing with "dpi" of video.
We got this discussion a few weeks ago and Loren Miller and me were "fighting a little" about the approaches on how to understand and handle; the end was that we were not so far away.
This "72 dpi", which visually was handled only a long, long time ago where "72 dpi " was the WYSIWG setting for 14" BW Mac monitor (there was the choice of 72 or 75 dpi in PageMaker). This changed when monitors became bigger and allowed to use several resolutions. WYSIWG and the sense of "72 dpi" has gone since those days. And sorry John, screen resolution is not 72 dpi nowadays, just something depending on your monitor - just take a ruler and a calculator. Though Quartz could handle that to match the feeling of it nowadays. But you're right pointing out that the final resolution in pixels is the most important thing.

So coming back to our video, as an example a simple HD still should have a final resolution of 1920x1080 at whatever dpi. You might create or scan it at a bigger size and than sample down to get a better quality, you also may squeeze it to match the pixel aspect of your current format (like 1440x1080), but it's just a fixed size. Some programs might take the "dpi" settings into account for display - not for quality!
Things become a little bit more complicated when you want to zoom (or pan) with the still, but it's not that complicated to calculate. If you want to zoom 250% for example just multiply the 1920 and the 1080 by 2.5 and you got the needed final resolution in square pix.
As Loren stated correctly the "dpi" are relevant when creating stills by scanning artwork. And it needs to be understood where you want to go and how things must match. Not every resolution will work with every software.

Loren posted a link to a nice tutorial on LAFCPUG
www.lafcpug.org/tutorials/basic_scanpro.html
There is also a short "how to" on a tool I did for some of my customers
[www.spherico.de]
Maybe those help to understand a little bit better.

Regards
Andreas

This is all true. 72 dpi might be a nice number to get one started with a scan, but really the only relevant thing is the actual size, and to work in whatever dimensions are called for by the final output. SD DV means 720 wide by 480 high (or in some cases 720 by 534, to account for pixel aspect distortion). HD can mean 1920 X 1280, or 1280 X 720. Or you may need something else for a pan and/or zoom. My students all hated to do any math, but there's no easier, nor faster way to make a good graphic happen.

And as I said before, keep in mind that working in resolutions not native to the sequence may require a render. Non-native resolutions may also cause unexpected things to happen when dropping into the timeline.

And then there's pixel depth and bits... :-)



HarryD

> 72 dpi might be a nice number to get one started with a scan, but really the
> only relevant thing is the actual size, and to work in whatever dimensions
> are called for by the final output.

I'd suggest differently here. Don't scan at 72dpi. When you're scanning graphics, you should take into account things like whether you'll need that scanned file for non-video-related graphics, print, artwork, etc. I'd scan at a high dpi and picture dimensions, then make smaller copies of the files at 72dpi and large image size for the video work. Otherwise if you need to make artwork from the scans, you have to do them again, and nobody ever accused scanning of being fun.

FYI - the final version of Ampede PDF doesn't work in Tiger nor FCP5. The beta test version is supposed to.



HarryD

That's a good point, Derek. And Photoshop's Image Size and Canvas Size are perfect tools for resizing and changing resolution. If you have a lot of scans, batch automate the process to resize to video and go have a cold one. but if one scans at a low res (like 72) don't expect to bring it back for print in PS by resizing to 300 dpi. Alternatively, if one scans for video only at 300 dpi, you'll have a big file that will require a render in FCP, and won't look any better.

But if one is scanning for print, you're right - dpi absolutely does matter.

And I think this is what we're all saying - know what the project is first!

thanks,



HarryD

Andreas,
Thanks for your post. Good stuff.

BTW, Harry. It's not 720 x 534 for square pixels, it's 720 x 540.

[docs.info.apple.com]

I used to be a 720 x 534 guy too. Adobe and Apple have two different ratios that are both within the same spec.



Kevin Monahan
Social Support Lead, DV Products
Adobe
Adobe After Effects
Adobe Premiere Pro
Adobe After Effects and Premiere Pro Community Blog
Follow Me on Twitter!

Kevin Monahan wrote:
> Andreas,
> Thanks for your post. Good stuff.

Thanks for the warm words.
If anybody is interested I can release the small calculator for all - for free.
Let me know.

Regards
Andreas



Some workflow tools for FCP [www.spherico.com]
TitleExchange -- juggle titles within FCS, FCPX and many other apps.
[www.spherico.com]

Kevin Monahan wrote:


>
> BTW, Harry. It's not 720 x 534 for square pixels, it's 720 x
> 540.

Cool. I've heard both. But I can't draw a perfect circle anyway. :-)

Do you happen to know why/how it got changed? Because as I'm sure you know it definitely used to be published that it was 720X534 (i think that was back in the analog-to-digital CCIR601 days).

To be truthful, in the most recent version of software like Photoshop, I create things in 720X480 and they look fine, which surprised me. The software seems to finally know what to do with my files. I wish i did.



HarryD

O.K. I guess here's what confuses me...

Most all of the posts above - as well as Larry Jordan's article posted above indicate that importing should be done at a certain size (720 x 480 or 540 x 72 - etc. etc. depending on who you're talking to), in fact John K. posted above that:

"I import it at 720 x 540 x 72. If I plan to pan or zoom, I import it at 1440 x 1080 x 72."

However, this is confusing to some of us trying to learn more - I cannot find a way to set my scanner to 720 x 540 x 72 - I have tried to force a target size and the software doesn't like it - it always jumps one number up or down when you change the other one...

I guess the question is, if one cannot set one's scanner with these settings what's to be done? Set at 300 dpi and scale it down in photoshop image size or what?

Or is the formula mentioned above by Andreas the answer?

"If you want to zoom 250% for example just multiply the 1920 and the 1080 by 2.5 and you got the needed final resolution in square pix."

Could someone lead a novice in the right direction???

Thanks,



Kid

I just removed my ramble. Mr. Mustard re-posted his question and got a much shorter and eloquent answer....

zzzzzzzzzzzzzz



Post Edited (08-19-06 01:11)

_______________________________________
SCQT! Self-contained QuickTime ? pass it on!

> ?
> However, this is confusing to some of us trying to learn more -
> I cannot find a way to set my scanner to 720 x 540 x 72 - I
> have tried to force a target size and the software doesn't like
> it - it always jumps one number up or down when you change the
> other one...
>
That's what I wrote in my small calc tutorial, not any scan software will support any dpi.

> I guess the question is, if one cannot set one's scanner with
> these settings what's to be done? Set at 300 dpi and scale it
> down in photoshop image size or what?
>
In your case with a 8.5 x 11 doc the 11 will be the width I guess. Having a given square pix size of 720x540 (D1) the result will be 720 dots / 11 inch = 66 Dots Per Inch. Now scan the whole doc at 72 dpi, the closest match, as this nearly any scan software supports (some only support a lowest setting of 75). The result though will end up in a 792x594 document. Since the width is the main factor here - again I assume - the file must be scaled (resampled) to 720 in width. This will set the height to 540 as well. Having a SD DV/DVD format 720x534 the file needs to be cropped a little bit.

> Or is the formula mentioned above by Andreas the answer?
>
> "If you want to zoom 250% for example just multiply the 1920
> and the 1080 by 2.5 and you got the needed final resolution in
> square pix."
With the above said, it's quite easy to calculate the scan settings for a zoom.
For example if you want to zoom into a part of the doc which is 1" wide, you'll get a zoom factor of 11(11/1 with the given doc size of 11x8.5). So the above calculation will tell you that you need a file with the dimensions 7920x5940 (which is not supported by FCP).
Now what to do with the scanner settings? Quite easy 7920 dots / 11 inch = 720 Dots Per Inch - dpi. Again the problem that the scanner software might not support this dpi setting. So we take the next one above the 720, like 1200. This will blow up the file dimension to 13200x10200 and we need to resample it after scanning to the correct size.
Coming back to FCP: the max. supported dimensions are currently 4000x4000. So make a backward calculation. 4000/720 = 5.5 is the maximum zoom factor you can achieve. This will give a max part of the doc with 2 inch width. The scanner setting can be calculated as same as above: 360 dpi. The closest setting here might be 400. That will result in a 4400x3263, which has to be resampled.
To achieve the 11 time zoom you can either take that doc and zoom it, since the interpolation up to 200% is quite good in FCP or: scan the doc with the settings calculated for the "real zoom 11". Make a copy and resample to FCP max file width, then within the original file create a region with 4000x3000 around the center of the zoom and crop. Those two files now can be layered (and keyframed) in FCP to get a high quality zoom.

>
> Could someone lead a novice in the right direction???
Hope the above helped to understand the procedure.

Regards
Andreas

The short answer is, you can *scan* at any dpi or pixel dimensions as long as it's greater than 720 by 540 and do a resize in photoshop, if that's your workflow. In fact, you can scan at any size and then do a resize in FCP (under the motion/scale tab) if that's what you want. The latter is not the best course, but it does work. Heck, you can poutput tp pdf and bring that in if you choose. Lots of options here. you need to play around and see what works for you.

Keep an eye on the graphic after import to make sure it doesn't look squished - there's another issue at play we haven't discussed, namely square vs rectanguler pixels. Which shall remain for another day.

I second the motion to get Rick Harrington's book, I think it's called Photoshop for NLE or something like that. Excellent book, relates graphics to video very well. It's not like a PS book for print output; it's video oriented. It also offers tips for using PS to get the best use of it in a video setting. It discusses in detail everything here and a lot more. I used in courses I taught and it served the students well.



HarryD