(Archived) Evernote as a Media Manager...

[gazumped 12,021]

There's a query at the end of all this, if you have the patience to read on..

My latest brainwave has been to try out Evernote as a media manager. Like Food, but with Music and Film.

To explain: I have, as part of my archive of a misspent life, a copiously random collection of acetate (I think), vinyl, cassette tape, reel-to-reel, VHS, CD and DVDs, not to mention MP3/4, AVI, MOV and miscellaneous other file formats.

I also have a few 'media manager' databases that list some of the above with a bewildering number of extra fields for genre, producer, inside leg, and other terms relevant to their specialist content. I haven't found one yet that deals with all media types, and there are glaring gaps in some records where the data is not available, or I've been too lazy to type it in.

Despite considerable innovation on the part of my various software providers, there's no easy way to enter a new item either - you fill in all the details in the database, or you select a match (if there's one available) from the thousands of language and media options that apply for that media.

What I want to do is have one place where I can think "I want to see a Spielberg Film.." or "What Django tracks can I play.." and check out what I have available.

The 'easy entry' part comes with the DVD cover or the sleeve notes - scan (or photograph, for LP's and boxes) the documentation that was wrapped around the media, and you necessarily have all the data that are available for that specific item.

The limit on note sizes means I won't be storing any movies or DVDs, but I can at least add a code or confirm the name under which the item is filed, so I can walk right to it.

I'm just in the testing phase at the moment, and I ran into one small glitch - a recent addition to the 33 items in there now was Tom Cruise in Mission Impractical 4 (sorry - not my choice of film!) which is a 3MB JPG of the DVD cover. There's lots of design in the cover, and so far I can't 'find' that item by any normal search - 'mission' or 'cruise' etc.

(If you're still with me, thanks; you've arrived at the punchline:)

So - query for any peeps who Know About This Stuff - is there any limit on size for an OCR'd JPG? Any known limits for text rotation or colouring? TC's name forinstance is in white text on a dark brown/black background.

(If Evernote doesn't yet parse pics of your typical two-page and a spine DVD cover (with explosion graphics) please can this be logged as a feature request!)

It could also be that the OCR gnomes haven't done their job yet, so I'll retry in a day or two..

If all this works I'll have the most convenient means to index my collection, and for most purposes I can probably box or sell lots of the items after digitising the cover and transferring the content to MP4. All of my stuff -including the items I've forgotten about- will be available to family and friends, and there will be even more free shelf space around the Casa del Gaz.

All of the above, of course, is subject to Finagle's Law of Copious Free Time - something about the faster you move, the less of it you seem to have - so I'll put this in my 'someday' folder for now...

[GrumpyMonkey 4,320]

Yikes. Well, I think this is well beyond me, but I would try to do a couple of things. A photograph of the front and back of Tom's DVD cover (doesn't the back side have the information in black and white?) might help the OCR gnomes along. In addition, you might want to consider something like Scanner Pro or JotNot that will "process" these images so that they are much smaller, and if you do it in grayscale or black and white, probably a whole lot easier to OCR. It would at least be worth a test to see if the OCR does better with smaller and/or processed images.

[gazumped 12,021]

Thanks for the input GM - I'm hoping to avoid having to do too much to the image, because there's a lot of stuff to catalogue and my ideal is going to be to run DVD and CD covers through a scanner, and add the rest via my mobile camera. I appreciate I can do stuff to lower the quality of scans and pictures, but it seems a shame to add extra processing steps - and also unnecessary if it's the designer-chaotic layout of music and video sleeves that's causing the problem. I'll hope one of the devs may be able to comment further.. pretty please?

[GrumpyMonkey 4,320]

Yeah. I figured any answer I gave would be too superficial. Perhaps you should simplify your questions for busy developers. I think it might be just one question

(1) Are there any limits on images (size, height, width, etc) for OCR? If not, why is this image (insert shared note link) not getting OCR'd?

[gazumped 12,021]

As usual life is not so simple. Evernote now seems to recognise 'Mission' but not 'Impossible' and 'Ghost' but not 'Protocol' (they're all separate words) and Tom's name just don't happen at all... I'm not surprised about the latter since (I now notice) the main credit on the front of the jacket is well spaced - C R U I S E - so OCR may not realise it is a word. But he's in the smaller print on the back too..

There must (obviously) be a limit to how far OCR can disentangle recognisable words from a designer jacket, and I think a DVD cover may be a step too far.. here's a link to the front cover on IMDB.

I also tried converting a scan to PDF and performing a local OCR, but had even less success.. Evernote do seem to have a better result, though not perfect. More trials and tribulations required, I fear...

OK - Dear Busy Dev:

1. What practical limits are there on image OCR please? (size, height, width, content/ design etc)
   
2. Should I be able to photograph and OCR a film DVD cover with useful results? Or is that just not feasible yet?
   

(I know that's two questions, but I'm just greedy.)

[Blake Burch 60]

I've never really had luck with OCR on any program. It's always a hit and a miss, especially if you throw in graphic design and funky typography.

For your purpose, why not just clip the top portion of IMDB like so? It gets the movie cover, title, genre, director, writers and actors. Plus, it will never have any errors when it comes to searching out what you want.

Apparently, at the moment, you cannot highlight the whole thing and create a hyperlink to a local file on your computer. I just crashed Evernote trying to accomplish this... Additional feature request?

[gazumped 12,021]

"why not just clip the top portion of IMDB" - Thanks for the thought Blake, and that's probably what I'll do; but the attraction of scanning or photographing covers and sleeve notes is that it includes things like cassettes, CDs and LPs that aren't IMDB-fodder, and in some cases can amount to a small book about the content. -Anyone seen Jeff Wayne's War of the Worlds CDs and LPs?

(Drat: now I'm going to have to play some music... Just talk quietly amongst yourselves..)

"No one would have believed in the last years of the nineteenth century that this world was being watched... "

[sthomas51004 2]

Couldn't you do a google search, or whatever search engine you prefer, and clip that information to evernote. For example http://en.wikipedia.org/wiki/Jeff_Wayne's_Musical_Version_of_The_War_of_the_Worlds -- go to this link. Searching for each media might take some time but so would the scanning and this way it would be searchable.

[GrumpyMonkey 4,320]

I think one "could" create a digital library of your media that way (searching around and finding stuff), but if it were me, and I were a collector, I'd want photographs of "my" stuff. It is not uncommon, even in this age of mass production, to find inconsistencies/idiosyncrasies in items. They somehow escaped QC, there was a change in production at some point, etc. At any rate, it is more about the principle (photographing and having that OCR'd) than it is about accomplishing this particular task, right?

[gazumped 12,021]

At any rate, it is more about the principle (photographing and having that OCR'd) than it is about accomplishing this particular task, right?

...although it would be nice to have it working too...

As a (sort of) further example I now have a working system of scanning receipts - for anything - before I bin them. I had a lot of old receipts in envelopes and boxes which are now history, and a bigger pile which I'd been too lazy to file into the envelopes and boxes. (They were split by month and year). All that is now an empty space on a shelf. I do have some receipts still - they're for bigger ticket items that I might want to complain about in the next 12 months. UK complaints customer service departments tend to focus on paper receipts, not pictures of them. We Brits must be untrustworthy or something..

Anyhoo - the system is, get a day's worth of receipts; scan and bin everything but the BT items; move on. And this worked very well for me yesterday when I wanted to buy a new rear windscreen wiper blade for the Jalopy. Big store, about 100ft of shelves with brightly coloured boxes.. What type of blade did I buy last time? Searched "Citroen" and "wiper" and came up with a year-old receipt for the last time, including the right box code. Ten seconds later, job done!

I'd like to be in a similar state for media - not saying I'll be replacing any of it soon (although there's always BlueRay copies..) - but getting all my current items logged into my external brain, and being able to add any new ones quickly and easily. In this case it'll confirm which films and tracks I actually own, and possibly stop me buying duplicates at some point. And, as I said originally, it'll help me list out for my family what I've got that they can borrow if they want.

(It now occurs that I'm going to have to log "loaned to..." somewheres, just in case my nearest and dearest get more absent minded than usual..)

[gazumped 12,021]

Just an addendum.. had a look at KustomNote which offers templates and layout, and found that it also supports "suggest as you type" options so a quick input of "mission" and "impossible" quickly got me to a cover picture of Ghost Protocol, some basics about the film, and links to some review sites. Looks like a fair amount of effort to get into it, but potentially lots of added value. There's also the other downside that KN seems to have access to all my notebooks and notes...