Anyone know a foolproof method for changing sprite files to 2 colours and square pixcels so Sleuth won't complain?
On 22 Feb, Chris Newman wrote in message
<5bf32b2585newslists@npost.uk>:
Anyone know a foolproof method for changing sprite files to 2 colours and square pixcels so Sleuth won't complain?
ChangeFSI? The current version in RISC OS 5 appears to do this easily.
In article <mpro.ss38ij00gcjvo0mg5.news@stevefryatt.org.uk>,
Steve Fryatt <news@stevefryatt.org.uk> wrote:
On 22 Feb, Chris Newman wrote in message
<5bf32b2585newslists@npost.uk>:
Anyone know a foolproof method for changing sprite files to 2 colours and >>> square pixcels so Sleuth won't complain?
ChangeFSI? The current version in RISC OS 5 appears to do this easily.
I found I have 1.69 (17-Dec-22) in resources.
I set the sprite output to monochrome 25 coors and that opened in
Sleuth 3 but when I tried to OCR it gave a fatal error warning.
Setting to 16 colours opened in Sleuth but as a big black square - no
text. I'll keep working on it.
In message <5bf3a4da5fnewslists@npost.uk> Chris Newman
<newslists@npost.uk> wrote:
In article <mpro.ss38ij00gcjvo0mg5.news@stevefryatt.org.uk>, Steve
Fryatt <news@stevefryatt.org.uk> wrote:
On 22 Feb, Chris Newman wrote in message
<5bf32b2585newslists@npost.uk>:
Anyone know a foolproof method for changing sprite files to 2
colours and square pixcels so Sleuth won't complain?
ChangeFSI? The current version in RISC OS 5 appears to do this
easily.
I found I have 1.69 (17-Dec-22) in resources.
For reference current version 1.70 (10 Jul 23)
I got it via Packman.
I set the sprite output to monochrome 25 coors and that opened in
Sleuth 3 but when I tried to OCR it gave a fatal error warning.
Setting to 16 colours opened in Sleuth but as a big black square - no
text. I'll keep working on it.
I am assuming you are running Aemulor to get Sleuth 3 working as it is
26 bit only application.
In article <mpro.ss38ij00gcjvo0mg5.news@stevefryatt.org.uk>,
Steve Fryatt <news@stevefryatt.org.uk> wrote:
On 22 Feb, Chris Newman wrote in message
<5bf32b2585newslists@npost.uk>:
Anyone know a foolproof method for changing sprite files to 2
colours and square pixcels so Sleuth won't complain?
ChangeFSI? The current version in RISC OS 5 appears to do this easily.
I found I have 1.69 (17-Dec-22) in resources.
I set the sprite output to monochrome 25 coors and that opened in
Sleuth 3 but when I tried to OCR it gave a fatal error warning.
Setting to 16 colours opened in Sleuth but as a big black square - no
text. I'll keep working on it.
In the end, I gave up and moved WinPC side, where I can
now scan and OCR documents of all sorts without hardly
any problems.
In article <2a28a8f35b.chris@mytardis>, Chris Hughes
<news13@noonehere.co.uk> wrote:
In message <5bf3a4da5fnewslists@npost.uk> Chris Newman
<newslists@npost.uk> wrote:
In article <mpro.ss38ij00gcjvo0mg5.news@stevefryatt.org.uk>, Steve
Fryatt <news@stevefryatt.org.uk> wrote:
On 22 Feb, Chris Newman wrote in message
<5bf32b2585newslists@npost.uk>:
Anyone know a foolproof method for changing sprite files to 2
colours and square pixcels so Sleuth won't complain?
I am assuming you are running Aemulor to get Sleuth 3 working as it
is 26 bit only application.
I'd forgotten that as it opens happily without Aemulor. I'll try again
with it later. Dashing out to gig now.
In article <5bf3aa4d8bdave@triffid.co.uk>, Dave
<dave@triffid.co.uk> wrote:
In the end, I gave up and moved WinPC side, where I can
now scan and OCR documents of all sorts without hardly
any problems.
Hear, hear, Dave.
I used ABBYY FineReader for OCRing Doctor Wimp's Surgery.
The only problems it gave me were related to things like
!Run coming out as ! Run and one or two other items
relating to RISC OS naming conventions.
That was all sorted out under RISC OS rather than trying to
teach FineReader. It would have been even easier if I'd
been aware of StrongED's ability to do Search and Replace
on multiple files all at the same time.
John
In article <mpro.ss38ij00gcjvo0mg5.news@stevefryatt.org.uk>,
Steve Fryatt <news@stevefryatt.org.uk> wrote:
On 22 Feb, Chris Newman wrote in message
<5bf32b2585newslists@npost.uk>:
Anyone know a foolproof method for changing sprite files to 2 colours and >>> square pixcels so Sleuth won't complain?
ChangeFSI? The current version in RISC OS 5 appears to do this easily.
I found I have 1.69 (17-Dec-22) in resources.
I set the sprite output to monochrome 25 coors and that opened in
Sleuth 3 but when I tried to OCR it gave a fatal error warning.
Setting to 16 colours opened in Sleuth but as a big black square - no
text. I'll keep working on it.
Indeed John, Back in in the days when I was running my business I did
try with Sleuth 3... Goodness me, did we (Fay and I) try hard, but as
I've previously noted it was so problematic and unreliable, just not
usable in a business situation.
In article <5bf41b6693dave@triffid.co.uk>, Dave <dave@triffid.co.uk>
wrote: [Snip a lot]
Indeed John, Back in in the days when I was running my business I did
try with Sleuth 3... Goodness me, did we (Fay and I) try hard, but as
I've previously noted it was so problematic and unreliable, just not
usable in a business situation.
Scan in 8bit Grayscale at 100, 150, 200, 300, 400 dpi.
Take the images (Not Jpegs as I want to keep the images clean) to the
RISC OS side. (BMP or PNG)
Use DPlingScan to reduce to 4bpp then Save as a Spritefile.
100, 150, 200dpi completely useless.
300dpi OCR'd very well with just a few things out of place.
400dpi OCR'd completely okay.
I have a file which consists of many, many PDF pages.
They contain the words to approx 2075 jazz tunes.
I wanted one so I used !Snapper to tack a screen shot.
I also tried outputting a whole page fro !PDF then cropping what I didn't need. Also tried the same from MuView.
'Twould seem Sleuth doesn't want to lay with sprites taken like this.
It seems
On 26 Feb 2025 as I do recall,
Chris Newman wrote:
[snip]
I have a file which consists of many, many PDF pages.
They contain the words to approx 2075 jazz tunes.
I wanted one so I used !Snapper to tack a screen shot.
I also tried outputting a whole page fro !PDF then cropping what I didn't need. Also tried the same from MuView.
'Twould seem Sleuth doesn't want to lay with sprites taken like this.
It seems
OCR won't work with sprites at screen resolution (which is only about 90x90dpi).
On 26 Feb 2025 as I do recall,
Harriet Bazley wrote:
On 26 Feb 2025 as I do recall,However, depending on how the PDF file was created, i.e. (whether it
Chris Newman wrote:
[snip]
I have a file which consists of many, many PDF pages.
They contain the words to approx 2075 jazz tunes.
I wanted one so I used !Snapper to tack a screen shot.
I also tried outputting a whole page fro !PDF then cropping what I didn't >>> need. Also tried the same from MuView.
'Twould seem Sleuth doesn't want to lay with sprites taken like this.
It seems
OCR won't work with sprites at screen resolution (which is only about
90x90dpi).
holds the actual text or whether it is just somebody's wrapper for a collection of fuzzy JPEGS) it might be possible to get the page output
at a much higher resolution - you can fake this in !PDF by zooming the
view to, say, 400% and then using the Save as Sprite option. It still
won't be in monochrome, so you will have to tick the monochrome 2-colour option in ChangeFSI and resave as a Mode 25 sprite.
Sleuth may then be able to read it.
Have you tried to open the PDF file with !Gview? with the choice menu we
can produce a sprite file (2_colour monochrome) Mode 18 and can choose a
300% scale for example. (It works very well with Sharpeye)
In article <5bf5356db2dave@triffid.co.uk>, Dave <dave@triffid.co.uk>
wrote:
In article <5bf41b6693dave@triffid.co.uk>, Dave <dave@triffid.co.uk>
wrote: [Snip a lot]
Indeed John, Back in in the days when I was running my business I did
try with Sleuth 3... Goodness me, did we (Fay and I) try hard, but as
I've previously noted it was so problematic and unreliable, just not
usable in a business situation.
<snip>
Scan in 8bit Grayscale at 100, 150, 200, 300, 400 dpi.
Take the images (Not Jpegs as I want to keep the images clean) to the
RISC OS side. (BMP or PNG)
Use DPlingScan to reduce to 4bpp then Save as a Spritefile.
100, 150, 200dpi completely useless.
300dpi OCR'd very well with just a few things out of place.
400dpi OCR'd completely okay.
I usually get good results doing things that way (More or less).
I was trying to do something slightly different.
I have a file which consists of many, many PDF pages.
They contain the words to approx 2075 jazz tunes.
I wanted one so I used !Snapper to tack a screen shot.
I also tried outputting a whole page fro !PDF then cropping what I didn't >need. Also tried the same from MuView.
'Twould seem Sleuth doesn't want to lay with sprites taken like this.
It seems
On 26 Feb 2025 as I do recall,
Jean-Michel wrote:
[snip]
The above information is that of! Paint after saving the sprite.Have you tried to open the PDF file with !Gview? with the choice menu we
can produce a sprite file (2_colour monochrome) Mode 18 and can choose a
300% scale for example. (It works very well with Sharpeye)
I tried saving a 2 colour monochrome file at 300% scale from GView, butThis is the message that appears when the file is in 32 -bit color mode, Sharpeye does not want it, but manages very well with an 8 -bit file.
it crashed Paint when I double-clicked on it, and gave an Xsprbadbpp
error when I dragged it into SharpEye. :-(
(ChangeFSI can display it, and reports it as "RISC OS sprite, 744 byOk,
1052 pixels, 32 bits per pixel" - DPlngScan reports it as 'Type 6.
Palette entries 0")
On 26 Feb 2025 as I do recall, Harriet Bazley wrote:
On 26 Feb 2025 as I do recall, Chris Newman wrote:
[snip]
I have a file which consists of many, many PDF pages. They contain
the words to approx 2075 jazz tunes. I wanted one so I used
!Snapper to tack a screen shot. I also tried outputting a whole
page fro !PDF then cropping what I didn't need. Also tried the same
from MuView. 'Twould seem Sleuth doesn't want to lay with sprites
taken like this. It seems
OCR won't work with sprites at screen resolution (which is only about 90x90dpi).
However, depending on how the PDF file was created, i.e. (whether it
holds the actual text or whether it is just somebody's wrapper for a collection of fuzzy JPEGS) it might be possible to get the page output
at a much higher resolution - you can fake this in !PDF by zooming the
view to, say, 400% and then using the Save as Sprite option. It still
won't be in monochrome, so you will have to tick the monochrome
2-colour option in ChangeFSI and resave as a Mode 25 sprite.
The original book was printed from a home computer. That has then been
taken apart re runthrough an automatic feed sacanner to produce a file of
397 PDF pages (171Mb). Each page has several songs distributed over two columns.
By uing !PDF and zooming the view up mightinly I was able to take a
screen shot of just he song I wanted.
Running that through ChngeFSI at monochrome resulted in a block of black instead of text unless I set the output to 256 colours. Then Sleuth3
worked.
The output wasn't brilliant due to the imitations of te original.
On 27 Feb 2025 as I do recall,
Chris Newman wrote:
[snip]
The original book was printed from a home computer. That has then been taken apart re runthrough an automatic feed sacanner to produce a file of 397 PDF pages (171Mb). Each page has several songs distributed over two columns.
By uing !PDF and zooming the view up mightinly I was able to take a
screen shot of just he song I wanted.
Running that through ChngeFSI at monochrome resulted in a block of black instead of text unless I set the output to 256 colours. Then Sleuth3 worked.
The output wasn't brilliant due to the imitations of te original.
Yes, scanned text is unlikely to work well (especially as those scanned
PDF compilations tend to contain images in the form of compressed lossy
JPEGS -- fine for the human eye but not too good for magnification to
the sort of resolution required for OCR).
Transcribing a single song manually off the screen might have been
quicker!
In article <bcbb56f55b.harriet@bazleyfamily.co.uk>, Harriet Bazley ><harriet@bazleyfamily.co.uk> wrote:
On 26 Feb 2025 as I do recall, Harriet Bazley wrote:
On 26 Feb 2025 as I do recall, Chris Newman wrote:However, depending on how the PDF file was created, i.e. (whether it
[snip]
I have a file which consists of many, many PDF pages. They contain
the words to approx 2075 jazz tunes. I wanted one so I used
!Snapper to tack a screen shot. I also tried outputting a whole
page fro !PDF then cropping what I didn't need. Also tried the same
from MuView. 'Twould seem Sleuth doesn't want to lay with sprites
taken like this. It seems
OCR won't work with sprites at screen resolution (which is only about
90x90dpi).
holds the actual text or whether it is just somebody's wrapper for a
collection of fuzzy JPEGS) it might be possible to get the page output
at a much higher resolution - you can fake this in !PDF by zooming the
view to, say, 400% and then using the Save as Sprite option. It still
won't be in monochrome, so you will have to tick the monochrome
2-colour option in ChangeFSI and resave as a Mode 25 sprite.
The original book was printed from a home computer. That has then been
taken apart re runthrough an automatic feed sacanner to produce a file of
397 PDF pages (171Mb). Each page has several songs distributed over two >columns.
By uing !PDF and zooming the view up mightinly I was able to take a
screen shot of just he song I wanted.
Running that through ChngeFSI at monochrome resulted in a block of black >instead of text unless I set the output to 256 colours. Then Sleuth3
worked.
The output wasn't brilliant due to the imitations of te original.
I couldn't find a way of sending a single page to CView and running the
whole thing caused GView to stop with a window of multiple error messages.
On 26 Feb 2025 as I do recall,
Jean-Michel wrote:
I tried saving a 2 colour monochrome file at 300% scale from GView, but
it crashed Paint when I double-clicked on it, and gave an Xsprbadbpp
error when I dragged it into SharpEye. :-(
(ChangeFSI can display it, and reports it as "RISC OS sprite, 744 by
1052 pixels, 32 bits per pixel" - DPlngScan reports it as 'Type 6.
Palette entries 0")
I've got a paper copy of the book somewhere. As necessary, I'll scan
pages to get an appropriate sprite file of the bit I want.
I just thought getting them out of the PDF vesion might be simpler. Silly
me!
On 26/02/2025 22:04, Harriet Bazley wrote:
On 26 Feb 2025 as I do recall,
Jean-Michel wrote:
I tried saving a 2 colour monochrome file at 300% scale from GView, but
it crashed Paint when I double-clicked on it, and gave an Xsprbadbpp
error when I dragged it into SharpEye. :-(
(ChangeFSI can display it, and reports it as "RISC OS sprite, 744 by
1052 pixels, 32 bits per pixel" - DPlngScan reports it as 'Type 6.
Palette entries 0")
That isn't a monochrome sprite, they are 1 bit per pixel.
On 28 Feb 2025 as I do recall,
druck wrote:
On 26/02/2025 22:04, Harriet Bazley wrote:That's what my copy of GView (v1.52) apparently generates when 'Colour
On 26 Feb 2025 as I do recall,
Jean-Michel wrote:
I tried saving a 2 colour monochrome file at 300% scale from GView, but
it crashed Paint when I double-clicked on it, and gave an Xsprbadbpp
error when I dragged it into SharpEye. :-(
(ChangeFSI can display it, and reports it as "RISC OS sprite, 744 by
1052 pixels, 32 bits per pixel" - DPlngScan reports it as 'Type 6.
Palette entries 0")
That isn't a monochrome sprite, they are 1 bit per pixel.
depth' is set to 'monochrome'.... :-(
And the visual output in the window certainly appears black and white
with no greyscale, so I don't know what's going on. Switching the radio
icon to '16 colours' instead of 'monochrome' saves a sprite that still crashes !Paint 2.55, but which ChangeFSI reports as being 4 bits per
pixel - since when has Paint been unable to display a 16-colour sprite?
In fact on experiment *all* sprites saved out from GView's 'Save' option currently appear to be crashing Paint, whatever colour depth or scaling factor is set - and is apparently being obeyed in the visual display -
and whether or not the 'Use GhostScript sprite device' option at the
bottom of the choices is ticked. So I really don't know what is going
on!
The filesize of the generated spritefile varies plausibly according
to the colour depth requested, and they all load correctly if dragged
into ChangeFSI but give an "Internal error: abort on data transfer"
error if I double-click on them. Maybe I need to reboot....
In article <5bf5fe3e31newslists@npost.uk>,
Chris Newman <newslists@npost.uk> wrote:
I've got a paper copy of the book somewhere. As necessary, I'll scan
pages to get an appropriate sprite file of the bit I want.
I just thought getting them out of the PDF vesion might be simpler. Silly
me!
Unless you have something like Acrobat, the best way is to print the page
you want, then rescan to B&W JPEG, 600dpi and use that as your source for OCR. When using !Sharpeye I set !Changefsi to Monochrome, 2 colours Mode25
In fact on experiment *all* sprites saved out from GView's 'Save' option currently appear to be crashing Paint, whatever colour depth or scaling factor is set - and is apparently being obeyed in the visual display -
and whether or not the 'Use GhostScript sprite device' option at the
bottom of the choices is ticked. So I really don't know what is going
on!
The filesize of the generated spritefile varies plausibly according
to the colour depth requested, and they all load correctly if dragged
into ChangeFSI but give an "Internal error: abort on data transfer"
error if I double-click on them. Maybe I need to reboot....
In message <711937f65b.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
In fact on experiment *all* sprites saved out from GView's 'Save' option currently appear to be crashing Paint, whatever colour depth or scaling factor is set - and is apparently being obeyed in the visual display -
and whether or not the 'Use GhostScript sprite device' option at the
bottom of the choices is ticked. So I really don't know what is going
on!
Can you specify the version of the Ghostscript you use. I have no problem with version 8.54.
I had installed the more recent version but I think it generated the
problem of monochrome sprite. Too bad because it's a VFP version!
On 28 Feb 2025 as I do recall,
druck wrote:
On 26/02/2025 22:04, Harriet Bazley wrote:That's what my copy of GView (v1.52) apparently generates when 'Colour
On 26 Feb 2025 as I do recall,
Jean-Michel wrote:
I tried saving a 2 colour monochrome file at 300% scale from GView, but
it crashed Paint when I double-clicked on it, and gave an Xsprbadbpp
error when I dragged it into SharpEye. :-(
(ChangeFSI can display it, and reports it as "RISC OS sprite, 744 by
1052 pixels, 32 bits per pixel" - DPlngScan reports it as 'Type 6.
Palette entries 0")
That isn't a monochrome sprite, they are 1 bit per pixel.
depth' is set to 'monochrome'.... :-(
In message <5bf5f2ac6enewslists@npost.uk> Chris Newman
<newslists@npost.uk> wrote:
In article <bcbb56f55b.harriet@bazleyfamily.co.uk>, Harriet Bazley ><harriet@bazleyfamily.co.uk> wrote:
On 26 Feb 2025 as I do recall, Harriet Bazley wrote:
On 26 Feb 2025 as I do recall, Chris Newman wrote:
[snip]
I have a file which consists of many, many PDF pages. They
contain the words to approx 2075 jazz tunes. I wanted one so I
used !Snapper to tack a screen shot. I also tried outputting a
whole page fro !PDF then cropping what I didn't need. Also tried
the same from MuView. 'Twould seem Sleuth doesn't want to lay
with sprites taken like this. It seems
With PDFutils you can split the pdf up into single pages, export the
pdf to jpegs, pngs, tiffs plus other graphic formats.
<https://www.riscos.info/packages/DocumentDetails.html#PDFUtils>
Plus my KPDFutil provides a front end so their is no need to enter the commands:
<http://kevsoft.co.uk/#KPDFutil>
In message <5bf635f458Spambin@argonet.co.uk> Stuart
<Spambin@argonet.co.uk> wrote:
In article <5bf5fe3e31newslists@npost.uk>, Chris Newman
<newslists@npost.uk> wrote:
I've got a paper copy of the book somewhere. As necessary, I'll scan
pages to get an appropriate sprite file of the bit I want.
I just thought getting them out of the PDF vesion might be simpler.
Silly me!
Unless you have something like Acrobat, the best way is to print the
page you want, then rescan to B&W JPEG, 600dpi and use that as your
source for OCR. When using !Sharpeye I set !Changefsi to Monochrome,
2 colours Mode25
You can do it directly with !Gview + !Ghostscript monochrome + 300% (or
more) just try! It works very well with Sharpeye .... not many errors
and Rhapsody likes
:-)
When PDF files contain images, the result is not good. !DPLngScan has a function that helps a lot: histogram, To test, Important: change to
monochrom and the size of the document before (300%)
On 28 Feb 2025 as I do recall, Jean-Michel wrote:
In message <711937f65b.harriet@bazleyfamily.co.uk> Harriet Bazley
<harriet@bazleyfamily.co.uk> wrote:
[snip]
In fact on experiment *all* sprites saved out from GView's 'Save'
option currently appear to be crashing Paint, whatever colour depth
or scaling factor is set - and is apparently being obeyed in the
visual display - and whether or not the 'Use GhostScript sprite
device' option at the bottom of the choices is ticked. So I really
don't know what is going on!
Can you specify the version of the Ghostscript you use. I have no
problem with version 8.54. I had installed the more recent version
but I think it generated the problem of monochrome sprite. Too bad
because it's a VFP version!
That would explain it - I upgraded my copy of Ghostscript to 9.26....
(But surely if GhostScript's sprite output is bugged then someone would
have noticed by now?)
In article <23143df65b.jmb@jmc.bruck.orange.fr>, Jean-Michel <jmc.bruck@orange.fr> wrote:
!DPLngScan has a function that helps a lot: histogram, To test,
Important: change to monochrom and the size of the document before
(300%)
never used histogram.I must explore that.
On 28 Feb 2025 as I do recall,
Chris Newman wrote:
In article <23143df65b.jmb@jmc.bruck.orange.fr>, Jean-Michel
<jmc.bruck@orange.fr> wrote:
[snip]
How do you use a histogram? (I always thought it was just a form of!DPLngScan has a function that helps a lot: histogram, To test,
Important: change to monochrom and the size of the document before
(300%)
never used histogram.I must explore that.
graph displaying the maximum number of pixels in each colour or
something - it never seemed like very useful information....)
In article <bd2849f65b.harriet@bazleyfamily.co.uk>, Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
On 28 Feb 2025 as I do recall, Jean-Michel wrote:
In message <711937f65b.harriet@bazleyfamily.co.uk> Harriet Bazley
<harriet@bazleyfamily.co.uk> wrote:
[snip]
The problem is not new, I quickly returned to Martin's version.In fact on experiment *all* sprites saved out from GView's 'Save'
option currently appear to be crashing Paint, whatever colour depth
or scaling factor is set - and is apparently being obeyed in the
visual display - and whether or not the 'Use GhostScript sprite
device' option at the bottom of the choices is ticked. So I really
don't know what is going on!
Can you specify the version of the Ghostscript you use. I have no
problem with version 8.54. I had installed the more recent version
but I think it generated the problem of monochrome sprite. Too bad
because it's a VFP version!
That would explain it - I upgraded my copy of Ghostscript to 9.26....
(But surely if GhostScript's sprite output is bugged then someone would
have noticed by now?)
Oh dear! Have I opened a can of worms?
First, I Convert the Drawing to greyscale (256 levels), and increase size
by 300%.
Then, I open with menu the Histogram window. ( R=G=B)
You can see the density of the values (between 0 and 255), well otfen they are almost all very close which will prevent the OCR from working.
To have a good result it is necessary that this density is better
distributed between 0 and 256, which is 2 arrows of the histogram tool
makes it possible to make
UNDO allows you to test until you get a well contrasting image for the elements to be analyzed.
For example for Sharpeye the lines are essential ...
So you have to test, but you can use this process with photographs (jpeg, greyscale if possible)
This requires a little training, but worth it!
On 28 Feb 2025 as I do recall,
Jean-Michel wrote:
I understand...First, I Convert the Drawing to greyscale (256 levels), and increase size
by 300%.
Then, I open with menu the Histogram window. ( R=G=B)
You can see the density of the values (between 0 and 255), well otfen they >> are almost all very close which will prevent the OCR from working.
Hmm - if the image is monochrome or greyscale then I would assume that
the R G and B values are expected to be all more or less identical?
I have here an example of a sprite file which I managed to save out from
a public-domain string quartet score (downloaded from the IMSLP) which I
was attempting at one point to transfer into Rhapsody before I gave up. Mainly because it contains sixteen pages, each one of which
needed to be transferred separately and required corrections - I got
bored :-(
The histogram for this black and white image has a single vertical lineYes,
on the far right-hand side which is in exactly the same position for
Red, Green and Blue.
You can get closer to the vertical line, there is no information in theTo have a good result it is necessary that this density is better
distributed between 0 and 256, which is 2 arrows of the histogram tool
makes it possible to make
There are two arrows underneath the histogram, one about a quarter of
the way across and another about three-quarters of the way across.
Note the best result is when the sprite is about 2500x3300.UNDO allows you to test until you get a well contrasting image for the
elements to be analyzed.
For example for Sharpeye the lines are essential ...
In message <dedc71f65b.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
The histogram for this black and white image has a single vertical lineYes,
on the far right-hand side which is in exactly the same position for
Red, Green and Blue.
Move the arrows so as to frame (encadrer) the vertical line, the density spectrum does not move, on the other hand the left arrow will correspond
to black and that of the right to the white, the contrast is much more important.
To have a good result it is necessary that this density is better
distributed between 0 and 256, which is 2 arrows of the histogram tool
makes it possible to make
There are two arrows underneath the histogram, one about a quarter ofYou can get closer to the vertical line, there is no information in the
the way across and another about three-quarters of the way across.
white parts of the histogram.
Yes, but the details in this case are the gray levels, for the image youOh, I see (I think) - the idea is to discard the 'white' areas in orderYou can get closer to the vertical line, there is no information in theTo have a good result it is necessary that this density is better
distributed between 0 and 256, which is 2 arrows of the histogram tool >>>> makes it possible to make
There are two arrows underneath the histogram, one about a quarter of
the way across and another about three-quarters of the way across.
white parts of the histogram.
to focus more closely on the existing range of variation within the
small 'black' area. Like cropping a digital image to give the effect of zooming in on the detail in part of the picture.
(I don't get any visible result by clicking 'Expand' on my images,Indeed if the Sprite contains only pixels "0" and "1", the histogram will
though, probably because they were already exported from PDF in a
2-colour sprite mode; the histogram remains as a single thin vertical
line, but with two arrows now located close on either side of it....)
In message <35fb3d025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
[snip]Oh, I see (I think) - the idea is to discard the 'white' areas in orderYou can get closer to the vertical line, there is no information in theTo have a good result it is necessary that this density is better
distributed between 0 and 256, which is 2 arrows of the histogram tool >>>> makes it possible to make
There are two arrows underneath the histogram, one about a quarter of
the way across and another about three-quarters of the way across.
white parts of the histogram.
to focus more closely on the existing range of variation within the
small 'black' area.
(I don't get any visible result by clicking 'Expand' on my images,Indeed if the Sprite contains only pixels "0" and "1", the histogram will
though, probably because they were already exported from PDF in a
2-colour sprite mode; the histogram remains as a single thin vertical line, but with two arrows now located close on either side of it....)
not be useful.
I guess you export from a file created by Gview?
The histogram does not change, but the image yes, the UNDO button allows
you to make tests and refine the adjustments of the cursors.
Freescore scores, often are only scan, not very good.
On 24 Mar 2025 as I do recall,Ok
Jean-Michel wrote:
In message <35fb3d025c.harriet@bazleyfamily.co.uk>[snip]
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
Oh, I see (I think) - the idea is to discard the 'white' areas in orderYou can get closer to the vertical line, there is no information in the >>>> white parts of the histogram.To have a good result it is necessary that this density is better
distributed between 0 and 256, which is 2 arrows of the histogram tool >>>>>> makes it possible to make
There are two arrows underneath the histogram, one about a quarter of >>>>> the way across and another about three-quarters of the way across.
to focus more closely on the existing range of variation within the
small 'black' area.
(I don't get any visible result by clicking 'Expand' on my images,Indeed if the Sprite contains only pixels "0" and "1", the histogram will
though, probably because they were already exported from PDF in a
2-colour sprite mode; the histogram remains as a single thin vertical
line, but with two arrows now located close on either side of it....)
not be useful.
I guess you export from a file created by Gview?
I think it was from Gview, but it might have been from !PDF or !PDFTest;
Sorry but I can't find the file, if you can send it back to me.The histogram does not change, but the image yes, the UNDO button allows
you to make tests and refine the adjustments of the cursors.
I tried again by exporting a page from the same PDF file (the Pleyel
quartet I sent to you), which is definitely not in monochrome since it contains red bar numbers.
But I still don't see any change to the image, even to the red sections.
However, if I go to the *first* page of the score, which contains a
scanned image rather than the notes drawn in an embedded music font,
then moving the arrows on the histogram and clicking on 'Expand' does
affect the portion of the page with the image on it.
Freescore scores, often are only scan, not very good.
Clearly this particular score is too high-quality a PDF file (probably computer-typeset rather than scanned in) to be affected by histogram manipulation. :-)
In message <8062ac025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
I tried again by exporting a page from the same PDF file (the Pleyel quartet I sent to you), which is definitely not in monochrome since it contains red bar numbers.
But I still don't see any change to the image, even to the red sections.
Sorry but I can't find the file, if you can send it back to me.
I found a file at ISMLP (Pleyel Quartet) ...
and also at MuseScore String Quartet B303.
On 24 Mar 2025 as I do recall,
Jean-Michel wrote:
In message <8062ac025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
[snip]
Still not found ...I tried again by exporting a page from the same PDF file (the Pleyel
quartet I sent to you), which is definitely not in monochrome since it
contains red bar numbers.
But I still don't see any change to the image, even to the red sections.
Sorry but I can't find the file, if you can send it back to me.
Well, as I said in my e-mail, I did download it quite a long time ago
now, so it may not be very representative of what is currently on offer
from the ISMLP :-(
[snip]There are several string quartets in A major, and I can't try these
I found a file at ISMLP (Pleyel Quartet) ...
and also at MuseScore String Quartet B303.
IMSLP757242-PMLP53080-partitur/pdf - Opus 74, the string quartet in A
major.
I thought we had parts for a selection of Pleyel quartets - also Ignaz Lachner - but they seem to have gone missing somewhere, so I had
ambitions to get at least one example into Rhapsody so that it was
possible to print out the separate parts from the score.
In message <9043bf025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
On 24 Mar 2025 as I do recall,
Jean-Michel wrote:
In message <8062ac025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
[snip]
I tried again by exporting a page from the same PDF file (the Pleyel
quartet I sent to you), which is definitely not in monochrome since it >>> contains red bar numbers.
But I still don't see any change to the image, even to the red sections.
Sorry but I can't find the file, if you can send it back to me.
Well, as I said in my e-mail, I did download it quite a long time agoStill not found ...
now, so it may not be very representative of what is currently on offer from the ISMLP :-(
On 25 Mar 2025 as I do recall,
Jean-Michel wrote:
In message <9043bf025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
On 24 Mar 2025 as I do recall,
Jean-Michel wrote:
In message <8062ac025c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
[snip]
Still not found ...I tried again by exporting a page from the same PDF file (the Pleyel >>>>> quartet I sent to you), which is definitely not in monochrome since it >>>>> contains red bar numbers.Sorry but I can't find the file, if you can send it back to me.
But I still don't see any change to the image, even to the red sections. >>
Well, as I said in my e-mail, I did download it quite a long time ago
now, so it may not be very representative of what is currently on offer
from the ISMLP :-(
I have now e-mailed it to you twice, so I think something must be
blocking the file :-(
In message <448b16035c.harriet@bazleyfamily.co.uk>
Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:
I have now e-mailed it to you twice, so I think something must be
blocking the file :-(
You did well to warn me, your mail was considered spam by my ISP and therefore blocked. I just recovered it.
In the file, the composer is Ignaz Lachner, not Ignaz Pleyel?
I did a test with Sharpeye, the PDF file is almost perfect and therefore
the result too. With this file no need to use! DPlngScan.
Il y a quand même quelques erreurs < 1%
Note for Rhapsody4, do not forget to recreate the ties, they are not
properly built in Sharpeye.
Important I use !GView with ghostscript 8.54 !
In article <989a43035c.jmb@jmc.bruck.orange.fr>,
Jean-Michel <jmc.bruck@orange.fr> wrote:
Important I use !GView with ghostscript 8.54 !
Where do you get it from?
Alan
In message <5c03509b17ajg@argonet.co.uk>
Alan Griffin <ajg@argonet.co.uk> wrote:
In article <989a43035c.jmb@jmc.bruck.orange.fr>,
Jean-Michel <jmc.bruck@orange.fr> wrote:
Important I use !GView with ghostscript 8.54 !
Where do you get it from?
Alan
From Martin Würhtner site:
http://www.mw-software.com/software/freeware.html
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 546 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 45:41:04 |
| Calls: | 10,394 |
| Calls today: | 2 |
| Files: | 14,066 |
| Messages: | 6,417,268 |