"Mike Kujbida" <kujfam-misleadingspam@sympatico.ca> wrote in message
news:nGwie.1803$Wp.396759@news20.bellglobal.com...
> Henry Padilla wrote:
>> I'm scanning in a book to read on video (the words will change color
>> for the kids to read along and stuff).  But I was wondering, other
>> than "as high as you can get" what dpi should I scan these at?
>>
>> My scanner goes to 2400dpi so "as high as you can" will make a
>> ~800Meg gif. That doesn't help.
>>
>> Tom P.
>
>
> Tom, what it comes down to is to scan as big as you need to in order to
> make
> the print readable on the highest resolution screen that it'll be shown
> on.
> My rule of thumb is that one dpi = 1 pixel.  Therefore, if I'm scanning a
> 4"
> x 6" print for video, I'll scan at 150 dpi.  This works out to a screen
> res
> of 600 pixels x 800 pixels which is just over NTSC res of 720 x 480.
> You don't say if you're bringing these into an NLE or not.    This would
> have some bearing on your settings as video res is only 720 x 480 but a
> typical monitor is 1024 x 768 (or larger).  If you want to fill the frame
> completely, the latter omes are the numbers I would take into
> consideration.
> OTOH, if you're going into an NLE and want to do a pan&scan of the image,
> then, once again, the larger the better.  HTH.
>
> Mike

"Richard Crowley" <rcrowl***@xprt.net> wrote in message
news:118lb17puke5r8e@corp.supernews.com...
> "Henry Padilla" wrote ...
>> I'm scanning in a book to read on video (the words will change color for
>> the kids to read along and stuff).  But I was wondering, other than "as
>> high as you can get" what dpi should I scan these at?
>>
>> My scanner goes to 2400dpi so "as high as you can" will make a ~800Meg
>> gif. That doesn't help.
>
> That translates directly to "how big do you want it to appear on the
> screen"?  (Assuming no diddling with the images after scanning.)
>
> You can even calculate it. You know the vertical and horizontal
> number of pixels. So, for example, if you chose "100DPI" the TV
> screen would show approx 4.5 x 7.2 inches (but remember to
> allow for the "safe action area" and "safe title area").
>
> Assuming also that you have permission to do this from whomever
> holds the copyright to the book.

"Henry Padilla" <padil***@hotmail.com> wrote in message
news:KMydnXvpZbu2FhffRVn-qg@comcast.com...
> I'm scanning in a book to read on video (the words will change color for
the
> kids to read along and stuff).  But I was wondering, other than "as high
as
> you can get" what dpi should I scan these at?
>
> My scanner goes to 2400dpi so "as high as you can" will make a ~800Meg
gif.
> That doesn't help.
>
> Tom P.
>
>
>

"Steve King" <st***@TakeThisOutToReplysteveking.net> wrote in message
news:BZCdnWs6HKh3_hbfRVn-vg@comcast.com...
> "Steve Guidry" <stevegui***@earthlink.net> wrote in message
> news:7JIie.4052$uR4.956@newsread2.news.atl.earthlink.net...
>> Hmmm . . .
>>
>> That's like asking "what flavor is a horse? ",  or "what color is the
>> music?
>> "
>>
>> The two don't go together.
>>
>> Steve
>
> But his problem of what resolution to scan material for inclusion in a
> video project is no closer to a solution than it was before your comment.
> I'm no video editor.  I write, direct, and produce.  However, my
> understanding is that material scanned or saved at 72 dpi looks good on
> NTSC video, when used directly.  If one wants to do a track-motion effect
> to zoom in on a particular part of a graphic image, then the resolution
> has to be proportionately higher so that the resultant expanded image is
> at least 72 dpi.  Example:  an area on a graphic image will be enlarged 4
> times on the video screen; then, the resolution of the original graphic
> should be at least 4 times 72, or approx 300 dpi.  No doubt others will
> offer more accurate and useful information.

>
> Steve King
>

"PTRAVEL" <ptravel88-use***@yahoo.com> wrote in message
news:3f16n1F5d61kU1@individual.net...
>
> "Steve King" <st***@TakeThisOutToReplysteveking.net> wrote in message
> news:BZCdnWs6HKh3_hbfRVn-vg@comcast.com...
>> "Steve Guidry" <stevegui***@earthlink.net> wrote in message
>> news:7JIie.4052$uR4.956@newsread2.news.atl.earthlink.net...
>>> Hmmm . . .
>>>
>>> That's like asking "what flavor is a horse? ",  or "what color is the
>>> music?
>>> "
>>>
>>> The two don't go together.
>>>
>>> Steve
>>
>> But his problem of what resolution to scan material for inclusion in a
>> video project is no closer to a solution than it was before your comment.
>> I'm no video editor.  I write, direct, and produce.  However, my
>> understanding is that material scanned or saved at 72 dpi looks good on
>> NTSC video, when used directly.  If one wants to do a track-motion effect
>> to zoom in on a particular part of a graphic image, then the resolution
>> has to be proportionately higher so that the resultant expanded image is
>> at least 72 dpi.  Example:  an area on a graphic image will be enlarged 4
>> times on the video screen; then, the resolution of the original graphic
>> should be at least 4 times 72, or approx 300 dpi.  No doubt others will
>> offer more accurate and useful information.
>
> Concerns about tracking, panning and zooming are, of course, valid and are
> dictated by the _resolution_ of scanned material.  However, "dpi" is an
> output measure and has nothing to do with resolution.  A 720 x 480 scan,
> whether scanned at 78 dpi or 2000 dpi, will have exactly the same
> resolution, i.e. 720 x 480.
>

"Steve King" <st***@TakeThisOutToReplysteveking.net> wrote in message
news:06KdnSBQjMwNHBbfRVn-uQ@comcast.com...
>
> "PTRAVEL" <ptravel88-use***@yahoo.com> wrote in message
> news:3f16n1F5d61kU1@individual.net...
>>
>> "Steve King" <st***@TakeThisOutToReplysteveking.net> wrote in message
>> news:BZCdnWs6HKh3_hbfRVn-vg@comcast.com...
>>> "Steve Guidry" <stevegui***@earthlink.net> wrote in message
>>> news:7JIie.4052$uR4.956@newsread2.news.atl.earthlink.net...
>>>> Hmmm . . .
>>>>
>>>> That's like asking "what flavor is a horse? ",  or "what color is the
>>>> music?
>>>> "
>>>>
>>>> The two don't go together.
>>>>
>>>> Steve
>>>
>>> But his problem of what resolution to scan material for inclusion in a
>>> video project is no closer to a solution than it was before your
>>> comment. I'm no video editor.  I write, direct, and produce.  However,
>>> my understanding is that material scanned or saved at 72 dpi looks good
>>> on NTSC video, when used directly.  If one wants to do a track-motion
>>> effect to zoom in on a particular part of a graphic image, then the
>>> resolution has to be proportionately higher so that the resultant
>>> expanded image is at least 72 dpi.  Example:  an area on a graphic image
>>> will be enlarged 4 times on the video screen; then, the resolution of
>>> the original graphic should be at least 4 times 72, or approx 300 dpi.
>>> No doubt others will offer more accurate and useful information.
>>
>> Concerns about tracking, panning and zooming are, of course, valid and
>> are dictated by the _resolution_ of scanned material.  However, "dpi" is
>> an output measure and has nothing to do with resolution.  A 720 x 480
>> scan, whether scanned at 78 dpi or 2000 dpi, will have exactly the same
>> resolution, i.e. 720 x 480.
>>
>
> Your last statement is true, of course, about the video image, but it is
> no help to the OP.  When using graphic images in video, dpi is what my
> scanners offer as a means of determining output quality of scanned images.
> My editor doesn't ask for an image 2880 x 1920 pixels, when he wants to do
> a move to a 4x enlargment.
>
> What would you say to help the OP?

>
> Steve King
>
>
>

"Steve King" <st***@TakeThisOutToReplysteveking.net> wrote in message
news:BZCdnWs6HKh3_hbfRVn-vg@comcast.com...
> "Steve Guidry" <stevegui***@earthlink.net> wrote in message
> news:7JIie.4052$uR4.956@newsread2.news.atl.earthlink.net...
> > Hmmm . . .
> >
> > That's like asking "what flavor is a horse? ",  or "what color is the
> > music?
> > "
> >
> > The two don't go together.
> >
> > Steve
>
> But his problem of what resolution to scan material for inclusion in a
video
> project is no closer to a solution than it was before your comment.  I'm
no
> video editor.  I write, direct, and produce.  However, my understanding is
> that material scanned or saved at 72 dpi looks good on NTSC video, when
used
> directly.  If one wants to do a track-motion effect to zoom in on a
> particular part of a graphic image, then the resolution has to be
> proportionately higher so that the resultant expanded image is at least 72
> dpi.  Example:  an area on a graphic image will be enlarged 4 times on the
> video screen; then, the resolution of the original graphic should be at
> least 4 times 72, or approx 300 dpi.  No doubt others will offer more
> accurate and useful information.
>
> Steve King
>
>

"Jukka Aho" <jukka.***@iki.fi> wrote in message
news:lzMie.446$Ao6.408@reader1.news.jippii.net...
> Henry Padilla wrote:
>
>> I'm scanning in a book to read on video (the words will change color
>> for the kids to read along and stuff).  But I was wondering, other
>> than "as high as you can get" what dpi should I scan these at?
>>
>> My scanner goes to 2400dpi so "as high as you can" will make a
>> ~800Meg gif. That doesn't help.
>
> 1) Measure the height of your target TV screen (yes, with a tape measure.)
> For the purposes of this discussion, let's suppose it is a relatively
> small set and the height of the screen is approximately 12".
>
> 2) Watch tv shows on that screen - especially the kind of channels that
> often have titles, scrolling tickers, whatever. A news channel, such as
> CNN, would be a good choice here. Make a mental note of how all the
> different font sizes appear on the screen and how legible they are when
> compared to each other.
>
> 3) Choose an on-screen font with a size that you believe would be good
> enough for displaying the text scanned from your book. Measure its height
> with the tape measure. For the purposes of this discussion, let's suppose
> you want the height to be 1" (on that particular screen.)
>
> 4) Now it's time to do some calculations. The typical resolution for NTSC
> DVDs is 720×480. However, not all of these pixels will usually be visible
> on a tv screen - because of overscan (http://scanline.ca/overscan/). We
> will have to compensate for that in our calculations.
>
> We will assume that even though you measured the visible screen height as
> being 12", in reality, some 10 % of the actual video signal will be cut
> off at the edges on that screen (5 % above and 5 % below.)
>
> This means that the physical height of the "virtual canvas" we have at our
> disposal when editing and compositing is something like the "h" in the
> following equation:
>
> h * 0.9 = 12"
>
> In other words,
>
> h = 12" / 0.9
>
> which is approximately 13.3".
>
> 5) OK, the DVD image consists of 480 pixels in the vertical direction. The
> physical height of this "canvas", when projected to our tv screen (this
> particular one), will be 13.3" (although the edges of that virtual image
> will actually be cropped so that the visible image is still 12".) Now, we
> can calculate the "video DPI" for this particular screen (although it is
> almost blasphemy to talk about such things!). The dpi value is 480 / 13.3"
> =~ 36.
>
> 6) Earlier during the chain of this analysis, we decided that the
> preferred text size we would want to use on this particular screen is 1"
> (which, as calculated above, equals to the height of 36 DVD pixels - in
> this case.)
>
> 7) Now, we should set our scanning resolution so that the height of a
> single text line in the book, when scanned, will _also_ equal 36 pixels.
> Measure the physical height of a single text line in the book.
>
> For the purposes of this discussion, we will assume that the font used in
> the book is quite big (it's a children's book, after all) and the height
> of the letters is approximately 1/3". Let's calculate the needed dpi value
> from that:
>
> 1/3" should equal to 36 pixels
> 3 * 1/3" = 1"
> => 1" equals to 3 * 36 pixels = 108 pixels
>
> The conclusion: we should scan at 108 dpi. This will give us the height of
> 36 pixels per a single text line, which will equal to a physical height of
> 1" on our tv set, when using the NTSC DVD resolution (this font size was
> deemed to be legible on that set, and it is what we wanted to have in the
> first place.)
>
> In practice, we should probably scan using a bit higher resolution, since
> this will allow for zooming in, if needed. Scanning with a bit higher
> resolution will also compensate for the fact that the the NTSC DV/DVD
> format has more resolution in the horizontal direction than in the
> vertical direction. The DPI value could be multiplied by (say) 1.2 to be
> on the safe side - which would give us ~130 dpi in this example. You will
> then have to downscale the images a bit in the compositing app a bit to
> get to the desired size. Measure the height of the letters from the final,
> rendered frames again if you're not satisfied with the result, and tweak
> settings accordingly.
>
> * * *
>
> You can skip the steps 3 - 6 if you have a video capture card or if your
> camcorder can record analogue signals. In that case, simply capture the tv
> channel you're viewing and measure the height of the preferred fonts in a
> graphics app - directly as pixels.
>
> (I hope I didn't make any coarse mistakes in the above calculations. This
> is not exact science, but by doing calculations as in the above, you'll at
> least get some ballpark dpi figures, which should prevent you from
> scanning using an unnecessarily high resolution.)
>
> --
> znark
>

> Seattle Eric wrote:
>
>> Damn, that seemed pretty silly, since it basically is geared towards
>> ONE screen size.
>
> The OP was going to use this information for creating a DVD of him reading a
> fairy tale to his child - displaying scanned pages of the original book as
> some sort of synchronized, animated slideshow accompanying the narration. My
> presupposition is that he would know on which tv set this custom-made DVD is
> most likely to be watched in his home.
>
> What I didn't explicitly mention is that once you go through the trouble of
> assessing and deciding on a readable text size for one particular tv screen
> (while taking into account the limitations of the NTSC resolution), the same
> pixel size will most likely work reasonably well on _all_ NTSC tv sets, even
> though the physical size of the screen (and, subsequently, the size of the
> text) will vary.

>
>Trust me, the OCR is going to work a treat, and it can be highly
>automated without a lot of work, especially if the original poster also
>has AfterEffects with his Adobe Premiere setup... and everybody pretty
>much should have the one if they're going to bother having the other.
>There are canned plug-ins for AE that will knock this project out slick
>as snot. Soundkeys, for example, will automatically pop up each word in
>the sentence as it 'hears" the plosives of the spoken words in the
>audio track. Well, I'm overstating it a little, but still... ;-)

> I don't mean to mess up you guys' fine arguments here, but seems to me
> the better way to go is to use the scanner on an "OCR" (optical
> character recognition) setting, assuming you have OCR software, which
> often comes bundled free with the scanner and is used for things like
> digitizing recipes and articles, etc...
>
>   yes, well, anyhow, OCR the kid's book, then you have a clean text
> file you can drop into your NLE system's character generator and you
> have ultimate control of the font size, placement, etc. and you can
> super this over the illustrations of junior's book, and of course if
> you want to get fiddly, you can keyframe the words to change color or
> type themselves onto the screen as the text is read aloud, whataver you
> want to do...  DPI is irrelevant to the discussion then, and you will
> have the best resolution too. 
>
>
> /   as you were...

>"nobody special" <msu1049***@aol.com> wrote in message
>news:1116469516.774319.111640@g47g2000cwa.googlegroups.com...
>>I don't mean to mess up you guys' fine arguments here, but seems to me
>> the better way to go is to use the scanner on an "OCR"
>
>Huh!  That's...  I didn't... Umm... yeah.
>
>I was thinking the same as Gene ("Well, what about the pictures?") but then
>you got to the end and I have to say... huh.
>
>I picked 150dpi to scan cause then the picture comes out looking really
>good.
>I'm gonna spend 30 -45 mins. in Photoshop "erasing" the text from the
>pictures (or, maybe, moving them to a different layer).
>Put everything in Premiere.  I do want to change the color of the letters so
>I have a decision to make:  Play around with Premiere's titler and see what
>can be done (limited but easy), or mess around in AE and see what can be
>done (harder but very cool).
>
>Thanks for the help guys.  I wish I could show you how it turns out.  How
>big a sample would be too much for fair use?  (I smell a different topic.)

On Thu, 19 May 2005 02:09:06 +0300, "Jukka Aho" <jukka.***@iki.fi>
wrote:

>Seattle Eric wrote:
>
>> Damn, that seemed pretty silly, since it basically is geared towards
>> ONE screen size.
>
>The OP was going to use this information for creating a DVD of him
>reading a fairy tale to his child - displaying scanned pages of the
>original book as some sort of synchronized, animated slideshow
>accompanying the narration. My presupposition is that he would know on
>which tv set this custom-made DVD is most likely to be watched in his
>home.
>
>What I didn't explicitly mention is that once you go through the trouble
>of assessing and deciding on a readable text size for one particular tv
>screen (while taking into account the limitations of the NTSC
>resolution), the same pixel size will most likely work reasonably well
>on _all_ NTSC tv sets, even though the physical size of the screen (and,
>subsequently, the size of the text) will vary.

"Henry Padilla" <padil***@hotmail.com> wrote in message
news:KMydnXvpZbu2FhffRVn-qg@comcast.com...
> I'm scanning in a book to read on video (the words will change color for
> the kids to read along and stuff).  But I was wondering, other than "as
> high as you can get" what dpi should I scan these at?
>
> My scanner goes to 2400dpi so "as high as you can" will make a ~800Meg
> gif. That doesn't help.
>
> Tom P.
>
>
>
>