Adding text label to photos (blurry fuzzy text)

"FUBARinSFO" wrote in message

Hi:

N00b here — apologies in advance.

I’m documenting a bunch of 16 x 20 black and white photographic prints. I’ve printed a postcard sized 4×6 description of each photo and placed it in the lower left-hand portion of the photo. I’ve taken a picture of this print with the description with a Canon 600SD 6 mp digital camera, at 1600×1200 resolution. The result is a .jpg image of about 700 KB.

The overall photo is fine, but the text in the label is very poor. The title line is Arial 20pt, barely readable, but the rest has to be blown up and even then is very poor. Pixel calcs are below. Clearly 75 – 80 ppi is far below the 300 dpi you get with a printer. What I want is a vector or hires label so that the viewer of the jpg can see the text description onscreen as clearly as he/she can read this text in this message.

My question is: how do I place a high-resolution label on this photograph? It doesn’t look like Photoshop Elements can do this. It appears to need a separate higher res or vector layer to be placed on top of the photo. Do I need to use the regular Photoshop or perhaps Adobe Illustrator to do this?

TIA

— Roy Zider

horizontal: 1600 pixels / 20" = 80 ppi
vertical: 1200 pixels / 16 " = 75 ppi

(1600 x 1200) / (20 x 16) = 6000 pixels/sq in = 77.5 ppi linear

Hi.

I don’t understand your decision NOT to use the full pixel count of the camera.

It is hardly surprising that you are having difficulty reading the text.

Your calculations are also a bit cock-eyed. The Ppi figure needs to be done linear. You can not have 2 different Ppi figures for the same image.

The camera would need to be mounted on a tripod, and the delayed action used to trigger the shutter, to eliminate camera shake. You would also need to be sure that the corners are not subject to any distortion, or out of focus.

You are trying to use a very compact P & S camera to do a job way beyond what it is intended for.

Roy G

Roy G:

Thank you for taking the time to post your thoughtful reply. Let me see if I can answer some of your questions.

1. I didn’t use the highest resolution of the camera (2816×2112) because the mode I used, Postcard mode, is the only one that imprints a date/time stamp on the photographic image. Using the highest res image means that I must sacrifice this timestamp, on this particular camera. Other, more capable cameras, probably allow you to imprint the images no matter what the resolution.

2. The calculations were provided so that you could assess the resolution range of the photos I was taking. Depending on which edge or the orientation of the photo, you would be able to get a better idea of the linear resolution. Overall it was about 77.5 ppi.

3. The camera was in fact mounted on a tripod, with delayed shutter (2 secs), natural lighting no flash. Sorry for not having mentioned that, which I sort of took for granted.

I am going to go back and try B&W mode (monochrome) and highest res, to see what the outer limits of this camera might be when photographing a text note tagged onto a photo.

But the problem here seems to boil down to a requirement for two different layers in the resulting file. The photo image layer — .jpg, say — and a text overlay layer. Notes can be attached to these files with PhotoShop Elements or even Adobe Photoshop Pro, but what I want is the ability to attach a Word-level note, with formatting. This may require Adobe Illustrator. If anyone has any experience with this, I would appreciate knowing.

Thanks again for your help.

— Roy Zider

FUBARinSFO wrote:

The overall photo is fine, but the text in the label is very poor. The title line is Arial 20pt, barely readable, but the rest has to be blown up and even then is very poor. Pixel calcs are below. Clearly 75 – 80 ppi is far below the 300 dpi you get with a printer. What I want is a vector or hires label so that the viewer of the jpg can see the text description onscreen as clearly as he/she can read this text in this message.

Are you using the superfine compression setting on the camera?

When transferring "designed for print" documents to a web viewable format, I have found that because of the lossy compression, JPEGs need to be extremely large to preserve text readability. Instead, I have been using GIFs, although they are only 8-bit. I’m afraid that for your photos, you are stuck with whatever compression the camera uses, so the index cards in the photo may be out the window, at least with that camera.

I don’t know how much information you are trying to pack onto those index cards, but here is an example of a small (596 x 512) image in which the text appears reasonably sharp:
http://bergdesign.net/images/HOT07.gif
I believe the document was saved from QuarkXpress as a pdf or eps and converted to gif in Photoshop.

Assuming post-processing is acceptable, you might try adding text to your JPEG in PSE and save as a GIF or lossless JPEG2000 (JP2). Also, IrfanView offers an "Insert Text into Selection" on the Edit menu that you may wish to try.

Fishface wrote:

Are you using the superfine compression setting on the camera?

Good catch — no, I’m not — just Fine. I am going to reshoot with SuperFine and see if that helps (it should, of course). But the reason I didn’t originally is because "Fine" is the only resolution available in this "postcard date imprint" mode, which is the only mode supporting a date/timestamp. So unless I abandon the date/timestamp (which I’m trying to retain), this is all I’ve got. Abandoning that, I can then go to SuperFine 2816 x 2112, the highest avail here (6 MP).

When transferring "designed for print" documents to a web viewable format, I have found that because of the lossy compression, JPEGs need to be extremely large to preserve text readability. Instead, I have been using GIFs, although they are only 8-bit. I’m afraid that for your photos, you are stuck with whatever compression the camera uses, so the index cards in the photo may be out the window, at least with that camera.

The camera supports Exif 2.2 JPEG file format. I don’t see anything about .gif format, so I guess that’s not an option.

I don’t know how much information you are trying to pack onto those index cards, but here is an example of a small (596 x 512) image in which the text appears reasonably sharp:
http://bergdesign.net/images/HOT07.gif
I believe the document was saved from QuarkXpress as a pdf or eps and converted to gif in Photoshop.

I checked out this image, and of course the text is readable, if not particularly clean. See my note below.

Assuming post-processing is acceptable, you might try adding text to your JPEG in PSE and save as a GIF or lossless JPEG2000 (JP2). Also, IrfanView offers an "Insert Text into Selection" on the Edit menu that you may wish to try.

There is evidently a RAW (lossless) mode, since the Advanced Guide shows it as an icon that may be displayed on an image, but at the moment I can’t find out how to set it.

I have IrfanView, and will check that out.

Finally: What I know will work is to bring the image into Microsoft Word and just use the text and tables in Word to manage the text, rather than using an image editor to import the text. This latter effort seems to be asking the products (PSE, PS, AI, etc) to commit a crime against nature.

Thanks for your suggestions and comments, Fishface.

— Roy Zider

FUBARinSFO wrote:

Hi:

N00b here — apologies in advance.

I’m documenting a bunch of 16 x 20 black and white photographic prints. I’ve printed a postcard sized 4×6 description of each photo and placed it in the lower left-hand portion of the photo. I’ve taken a picture of this print with the description with a Canon 600SD 6 mp digital camera, at 1600×1200 resolution. The result is a .jpg image of about 700 KB.

The overall photo is fine, but the text in the label is very poor. The title line is Arial 20pt, barely readable, but the rest has to be blown up and even then is very poor. Pixel calcs are below. Clearly 75 – 80 ppi is far below the 300 dpi you get with a printer. What I want is a vector or hires label so that the viewer of the jpg can see the text description onscreen as clearly as he/she can read this text in this message.

My question is: how do I place a high-resolution label on this photograph? It doesn’t look like Photoshop Elements can do this. It appears to need a separate higher res or vector layer to be placed on top of the photo. Do I need to use the regular Photoshop or perhaps Adobe Illustrator to do this?

TIA

— Roy Zider

horizontal: 1600 pixels / 20" = 80 ppi
vertical: 1200 pixels / 16 " = 75 ppi

(1600 x 1200) / (20 x 16) = 6000 pixels/sq in = 77.5 ppi linear

While I would tend to agree with the others, your primary problem is the inability for the file size to store enough data. However I would suggest that part of the problem is likely due to the location of the text. Generally lenses are sharpest on the center area of the image and you image is at a corner. In addition since it is a flat subject and most lenses have a curved focus plane, that would be adding insult to injury. A true macro (as opposed to a macro mode) lens tends to greatly reduce both of those issues.

—
Joseph Meehan

Dia ‘s Muire duit

Joseph Meehan wrote:

While I would tend to agree with the others, your primary problem is the inability for the file size to store enough data. However I would suggest that part of the problem is likely due to the location of the text. Generally lenses are sharpest on the center area of the image and you image is at a corner. In addition since it is a flat subject and most lenses have a curved focus plane, that would be adding insult to injury. A true macro (as opposed to a macro mode) lens tends to greatly reduce both of those issues.

Yes, it seems that photographing a postcard-sized block of text is a loser. I’ve gone to SuperFine, highest resolution, B&W mode and I can read it, after zooming. But it looks like what I really need is a text layer that I can merge with the image and still retain text readability and scaling. Flattening it to a bitmap image loses its legibility. I’ll probably just cave in and import the image into Word, rather than the Word text and table into the image editor.

— Roy

FUBARinSFO wrote:

The camera supports Exif 2.2 JPEG file format. I don’t see anything about .gif format, so I guess that’s not an option.

I’ve never seen it offered in a camera, but RAW and TIFF would be lossless, I believe. The JPEG compression is not kind to sharp edges.

I checked out this image, and of course the text is readable, if not particularly clean. See my note below.

I believe it was anti-alised. It softens the edges a bit, but being composed of pixels, the alternative was jaggies. There’s always a trade-off, it seems.

Finally: What I know will work is to bring the image into Microsoft Word and just use the text and tables in Word to manage the text, rather than using an image editor to import the text. This latter effort seems to be asking the products (PSE, PS, AI, etc) to commit a crime against nature.

The EXIF information does contain the date. I still have no idea why you need to do this, how much text you’re inserting, how often you need to do this, nor how you are viewing them, so I’m just stabbing around in the dark. Doing this often and/or continually would prescribe an automated solution.

Another thought is to use the PDF format. I believe it offers vector and raster in combination. You could even compose it in Word, if that is convenient. I found this "Print2PDF" on Sourceforge the other day, and though not perfect, certainly it’s worth the asking price of free: http://sourceforge.net/projects/print2pdf/
It installs as a printer which you can select when printing and it creates a PDF instead.

Fishface wrote:

The EXIF information does contain the date. I still have no idea why you need to do this, how much text you’re inserting, how often you need to do this, nor how you are viewing them, so I’m just stabbing around in the dark. Doing this often and/or continually would prescribe an automated solution.

Having the date/timestamp embedded in the graphic image establishes the uneditable date for purposes of documentation and chain of custody.

Another thought is to use the PDF format. I believe it offers vector and raster in combination. You could even compose it in Word, if that is convenient. I found this "Print2PDF" on Sourceforge the other day, and though not perfect, certainly it’s worth the asking price of free: http://sourceforge.net/projects/print2pdf/
It installs as a printer which you can select when printing and it creates a PDF instead.

I may emit the result as PDF, which of course you can do with Adobe Acrobat Pro. The composition I’m using now is Word with an embedded image. Emiting it to a pdf allows anybody to read it, and as you point out preserves the readability of the table data. The text in the resulting image is unreadable (loss of pixels) but then it’s all clearly reproduced in the text from Word itself. None of this is clean.

Thanks for your comments.

FUBARinSFO wrote:

Joseph Meehan wrote:

While I would tend to agree with the others, your primary problem is the inability for the file size to store enough data. However I would suggest that part of the problem is likely due to the location of the text. Generally lenses are sharpest on the center area of the image and you image is at a corner. In addition since it is a flat subject and most lenses have a curved focus plane, that would be adding insult to injury. A true macro (as opposed to a macro mode) lens tends to greatly reduce both of those issues.

Yes, it seems that photographing a postcard-sized block of text is a loser. I’ve gone to SuperFine, highest resolution, B&W mode and I can read it, after zooming. But it looks like what I really need is a text layer that I can merge with the image and still retain text readability and scaling. Flattening it to a bitmap image loses its legibility. I’ll probably just cave in and import the image into Word, rather than the Word text and table into the image editor.
— Roy

With elements you should be able to enlarge the canvas and put text into the new area.

—
Joseph Meehan

Dia ‘s Muire duit

Joseph Meehan wrote:

With elements you should be able to enlarge the canvas and put text into the new area.

Yes, but it’s terribly awkward, manual, and as I’ve just been experimenting with it just now, really isn’t close to do anything like documenting a picture. It does make a nice label in 36 pt type across a b&w photo, though, in green.

Thanks anyway. I’m going with an image embedded in a Word document (image into text) rather than the other way around.

— Roy