Getting more from image description.

By Khomus, 15 October, 2026

Forum
macOS and Mac Apps

Hi all.

Pre-AI, if I opened an image of text in Preview, and hit VO-shift-l, it would try torecognize text. I just tried this with an image using the new AI stuff on the Mac, and here's what I get.

A vintage record album back cover featuring the title "THE MUSIC OF PRIMITIVE MAN" in large, bold, black serif lettering at the top center. Two stylized, dark brown, abstract bird-like figures with long, curved necks flank the title text, while two smaller, similar bird figures sit at the bottom corners. The text below the title introduces the album as a collection of more than fifty examples of chants and instruments, credited to "JOSEPH PROSTAKOFF, Musical Program Director." The lower two-thirds of the page contains three columns of small, dense black text listing the track titles and descriptions for "SIDE ONE" and "SIDE TWO," printed on a slightly aged, off-white paper background with a thin red border along the right edge.

I mean, that's great, I know there are wacky pictures now. What I *don't* know is the text, the far more useful portion of the image. How do I get that?

Hitting VO-shift-r and picking "ask about image" and asking for the text in the image just tells me it's album liner notes with two columns of text. Again, fascinating stuff, but almost entirely useless. I'm already warea that it's album liner notes. I'd like to read said notes.

Options

Comments

Let's see if it actually takes my comment this time, it apparently lost the one I left last night.

"What do the Liner notes say"?

"The liner notes are titled "THE MUSIC OF PRIMITIVE MAN" and include an introductory paragraph by Musical Program Director Joseph Prostakoff, followed by a numbered tracklist for Side One and Side Two. The introduction explains that the record presents more than fifty examples of chants, instruments, and songs of work, war, love, and celebration recorded on site in remote corners of the world. The tracklist details 24 selections on Side One and 21 selections on Side Two, providing the cultural origin and a brief description of the music for each track."

"Provide the text of the liner notes".

Sorry, something went wrong. Please try again.

"What is the text of the liner notes"?

"The liner notes are titled "THE MUSIC OF PRIMITIVE MAN" and begin with an introductory paragraph by Joseph Prostakoff, Musical Program Director. The text is divided into two sides, with Side One listing tracks 1 through 20 and Side Two listing tracks 1 through 21. Each track entry includes the title and a brief descriptive paragraph about the cultural origin and instrumentation of the music."

"What are the descriptive paragraphs"?

"The descriptive paragraphs are "1. AINU FOLK SONG", "2. PYGMY SONG", "3. GUANAVIDJE CHANT", "4. DUSUN XYLOPHONE DUET", "5. TONGA NOSE FLUTE TRIO", "6. GUAHARIBO CHIEF'S SONG", "7. FIJI ISLANDS DANCE", "8. NEW GUINEA JEW'S-HARP SOLO", "9. FINNISH LAPLAND JOIK", "10. NEW GUINEA BAMBOO FLUTES", "11. ESKIMO TRAVELING SONG", "12. ESKIMO DANCE", "13. SOUTH AMERICAN INDIANS", "14. MOTILΓ“N INDIANS, SOUTH AMERICA", "15. ABELAM WARNING", "16. ABELAM YAM DISPLAY SONG", "17. AKAWAI CASSAVA GRATING SONG", "18. AKAWAI TRIBE, GUYANA", "19. CUNA SHAMAN'S CHANT", "20. CUNA PUBERTY FIESTA", "21. MOTILΓ“N INDIANS, SOUTH AMERICA", "22. DUSUN GONGS", "23. KALAHARI LOVE SONG", "1. BAJAU LAUT DANCE", "2. BAJAU LAUT SONG", "3. MAKUNA CHANT", "4. TONGA CHORUS", "5. DYAK GAMBUS DUET", "6. DYAK MEN'S DANCE", "7. AINU PRAYER SONG", "8. TATUYA WAR DANCE", "9. WALBIRI BLOODING CEREMONY", "10. KALAHARI MUSICAL BOW SOLO", "11. CUNA PUBERTY CELEBRATION", "12. FIJI ISLANDS DANCE", "13. DUSUN BAMBOO STICK ORCHESTRA", "14. JEBERO PANPIPES AND DRUMS", "15. KOGI FLUTES AND MARACAS", "16. NOANAMA SONG", "17. SEMELAI TRIBE, MALAYA", "18. DUSUN NOSE FLUTE SOLO", "19. DUSUN SUNDATANG SOLO", "20. MANTERI MAGIC CEREMONY", and "21. FIJI ISLAND DANCE AND CHANT"."

So at least now I have the list of tracks. I'll have to get my wife to take a look at it and see if there's anything beyond that. Maybe it's hallucinating actual descriptive text.

BTW, is there an easier way to get/save the text? So far I've been using VO-shift-c, copy last read, but that can be problematic for longer text. If I can get all of the text from this image I'd probably want to save it somewhere for reference.

Did you ask Siri to copy and paste it on where you going to use? Heard that it can do so, not sure if is right but curious if it works.

By Holger Fiallo on Thursday, October 15, 2026 - 18:54

As soon you ask for the description, ask to copy the text and if you have open the message ask to paste it. Not sure if it will work, it should. Let me know. Good luck.

By Seanoevil on Thursday, October 15, 2026 - 23:08

Hi,
If the information that you want is not in the initial image description, try asking SIRI AI.
With the Preview Document still in VO focus, invoke SIRI AI and ask to transcribe the notes.
For example, I have an image of an identity card. Image description gives me a holistic overview of the card, noting the photo, some key dates and that there is an Australian Address. When I ask SIRI AI for the exact address, I get the specific information I need.
Hope this helps.

By Khomus on Thursday, October 15, 2026 - 23:45

In reply to by Seanoevil

How am I asking? I thought that's what VO-shift-r was for? Also, Siri doesn't work with my microphone. I'm on a Mac Mini. So I don't know if that would make it less useful or not. Like I said, I don't really use SIri on the Mac at all. So some basic steps would be appreciated. I'm a total noob!

The only thing I use SIri for is to restart my phone or set timers on my watch. The one time I tried it for anything else, it misheard me, no idea how, and called 911. So I had to explain to the person that it was a total accident and give the address of the place I was at so they at least knew who called.

By Khomus on Friday, October 16, 2026 - 01:53

I sent the image to my wife, and apparently on her iPad she can just select the text in an image and copy and paste it. Can we do that on the Mac?