Databases (possibly featuring tags)

I've been given the delightful task of taking all of the tapes I've shot this past year (lots) and creating a new database for them for logging/easy retrieval. There's currently an archive system in place which allows users to log in remotely and either insert new tape information, edit old information or just look up what is on each tape.

The current options are divided into two types of potential input - it can either be a text bar (text doesn't have to be separated by comma etc) OR a drop down menu with only a few limited options and no way of adding new ones (the guy who built the page left years ago apparently).

In terms of search functions you can type into a search bar and there is a tick box for an Exact Match which is hugely ineffective simply due to End User Idiocy and the poorly detailed way in which the tapes get logged.

The plan is ultimately to add the entire library of many thousands of tapes (thankfully this won't fall down to me!) to the new system as well as the footage I've shot myself this year which is acting as kind of a test run to see how effective it could be.

What I'd like though is to know of any software that allows me to use tags in a similar way to the LFGSS search options. Say for example I had a tape which was called "John Smith Interview" featured a man called John Smith being interviewed in... Chelsea Harbour so there were set-ups of the interviewee, GVs of the harbour, then the interview. I'd like a way of "Tagging" what was on the tape so that if someone needed to search for "boats" or "yachts" or... "seagulls"(!) it would flag this tape as featuring that footage (in the GVs of the harbour). Currently the search feature wouldn't return any of those results and the only way you would know the tape existed is if you typed in "John Smith" which as you can imagine isn't a very effective way of finding footage..

So, software suggestions? Things to look out for? Can the end database be accessed online? I've never done anything like this before so any pointers much appreciated...

Go bike forums!

This is classic application developer stuff.

The answer is a clear Yes. That's exactly what databases are there for. You need to add information to your records and that can be done relatively easily by a skilled developer. Get yourself someone who knows SQL (database language), the database software you are using (most likely MySQL) and the programming language that currently creates HTML from the database (could be loads). Also, your developer should know about usability issues: geeks usually are very good at making things fast and consistent (which is important for a database) but fail to take into account that not everybody wants to tap into the full power of the software, they just wanna find what they are looking for.

On the other hand, if you don't want to go down the custom route I bet there are loads of software packages that let you to make an inventory of your footage. In this day and age they should have a web interface.

This is the infrastructure covered. You won't be able to automate the tagging though. Some poor sod will have to sit down, view the footage and add tags. I don't envy him/her.

School leavers / Graduates. Don't know they've been born, them kids.

And they're cheap. Or free, if you call it "work experience".

I like your style.

Cheers both of you!

A database for videos with metadata describing the video?

I've done this, it's not easy... I used several tools, the key one was Virage Videologger for capturing the digital video from tape or a broadcast stream and then merging it with an XmlStream that described the video and then encoding the video for the archival system whilst sending the XmlStream and the related timepoints for the videos to an Oracle database.

That system wasn't cheap, but it worked fantastically well and powered nearly all of the premier league football websites as well as a few porn ones (video search forté). For example, it was possible to ask the system to return all goals scored by a specific player against a certain goal keeper in a given season. It was also possible to ask it to return all redheads performing fellatio with greek men... hence it ended up being sold for porn too, oh well.

Anyhow, you should outline the scope of what you want a bit more and let us know the kind of budget you have to play with.

i.e. Are you just wanting a catalogue of all tapes and nothing more granular? Or are you wanting to be able to say "first part of this tape is blah, second part is boo"? Or would you like to go further and use voice recognition to create a textual index of all audio and be able to search for phrases spoken and retrieve the timestamp of that phrase in a given tape?

The scope of what you want determines cost and complexity... it can be extremely simple and cheap (a catalogue) or incredibly expensive and complex (something along the lines of what I described above).

The premier league project sounds ridiculous! Certainly won't need anything so complex - essentially they're after a catalogue but it needs to be one that actually works effectively unlike the current system.

Unfortunately our library isn't digitised (yet, though apparently it's being given serious consideration) so I presume that rules out any requirements for tools like Virage etc.

I've attached a couple of screens which show how utterly useless it is if a tape has been badly logged (which invariably is the case).

Some of the search options don't even get used (They're many years old).

All it really needs is: Tape Number, Tape Name, Category, A paragraph or so as a description and "tags" with potential for adding attachments i.e. audio transcriptions and shot lists (though if there were space in the "paragraph of description field" then the transcripts and shot lists could go straight in but it's a lot of work that I don't really want to be getting too involved in bearing in mind I was technically hired as camera/production...

It would also I suppose be useful if there were potential for the database to be easily adapted so that if in the next year or so they get the go-ahead to digitise (at a cost of around £80,000 -£100,000 we have so far been quoted) then we didn't have to come up with a whole new system just 12 months after the latest one started...

2 Attachments

'0000-00-00'?

According to your DB you have tapes reaching back to the time of Christ.

yeah, we're pretty hard core.

If all you want is a gentle reworking of the current system it shouldn't be too difficult. If I understand correctly you wouldn't want to ditch the current database but simply add the tagging function and have a better user interface and some features like unified search where people have only one search field rather than millions. I would say that this would need a decent developer about 1-2 weeks.

I used to do these things when I was at uni (only just left though) and I think this could be perfect for a switched-on student or a freelancer.

Couldn't find an appropriate thread so thought I'd dredge this one and save the merge.

Each week an .XLS file is updated (quite simple - 4 or 5 rows, 56(weekly!) columns on a networked drive by up to three users. The file is basically a weekly list of all the stories that go out to broadcast.

When I'm out of the office I'd like to have access to this and rather than use some sort of VNC set up I'd like to synch that XLS or equivalent into a calender on outlook which I can in turn synch on my phone.

Is there any easy way to do that rather than re-enter all the data into Outlook?

PS - The above situation that warranted (or didn't) warrant the original thread looks like it may now finally be resolved this summer - if so, I'll post results as/when.

You can try-

select the range in Excel then edit-copy, edit-paste special - paste link- as Excel Spreadsheet into the body of Calendar item.

whenever the spreadsheet updates the link will refresh in the Calendar item, but depends if your phone sync reads the excel object.

Alternatively you could a macro written that would automatically email you a copy whenever the file was updated?

Use Google Docs and create a shared spreadsheet?

Use Google Docs and create a shared spreadsheet?

Works for me.

[URL="http://www.microsoft.com/downloads/details.aspx?familyid=C26EFA36-98E0-4EE9-A7C5-98D0592D8C52&displaylang=en"][/URL]http://www.microsoft.com/downloads/details.aspx?familyid=C26EFA36-98E0-4EE9-A7C5-98D0592D8C52&displaylang=en

What about this ?

Returning to the original post, MongoDB's new version would be my first choice.

7 years later, similar question.

I have a new client and we're delivering as much content as we can to them per month - they'll pay for as much as we can give. At minimum this will be 40 still images and 10 x 5-10 second videos. Realistically this could be hundreds of images/videos per month. They already an archive in the hundreds/ Over time this will quickly become an archive of thousands/tens of.

Client needs a way to quickly find images for posting that doesn't rely on existing knowledge of archive.

Currently the client has a (to my eye) fairly sophisticated Google Sheets doc with various 'rules' that define naming conventions but essentially it's a fuck off great spreadsheet with long cumbersome file names such as:

McDonaldsSpicyChickenWinterMeltSmallFriesCokeWoodenTable.jpg

Apart from a Google Sheet with 10 quazillion rows is there an obvious way to make it easy to find stills or video based on what's in the picture?

I.e.:

Social Media Assistant wants to post image of something spicy that features chicken but doesn't know what exactly. They could filter/tick a few boxes that whittles down a selection of results to a few useable examples? Categories could be:

Merchant (McDonalds), Item (Spicy Chicken Winter Melt), Sides (Small Fries), Drink (Coke), Setting (Wooden Table) - perhaps also Media (Still/Video/GIF etc).

I'm open to the idea of pitching hire of a developer to create something bespoke but is there anything straight out of the box that's immediately user friendly? They use Google Apps for Business so if it's an app or something that can be incorporated that'd be marvellous - the media is realistically going to be stored on Google Drive so would be good to have it all under one roof (the search results could return a link to the actual file for example rather than just a file name you have to take back to Google Apps and manually search).

Bump anyone? Will take to AQA shortly.

Google Cloud Vision API & google storage bucket should give you what you want. That's assuming the challenge is analysing and identifying your asset attributes. Disclaimer: I've not used it, though I evaluated big table earlier this year for a bespoke solution to a similar problem.

Have you looked at something like Canto's Cumulus?

A few further thoughts with my first coffee…

TAXONOMY - How are the media assets to be described.
This will open you up to a never ending world of indecision. Nail down as good a structure as early as you can and at least threaten to charge for subsequent alterations.

MEDIA LOCATION
Store the media files inside your database and the database size grows very quickly (speed/backup issues). Store only paths to the files and you have to be very careful about moving them around.

MEDIA FORMATS
Don’t choose anything too proprietary. For eventual display on-line you may need to repurpose source (archive quality) files.

DRIFT
Lock down what data can be entered as much as possible. Even free text fields should have guidance as to the scope of the language that can be used. Otherwise the accuracy of the records WILL drift over time.

D.A.M.
Digital Asset Management is the general term for this stuff. Certainly the package I referenced earlier is mature but I’ve no current idea of costs. As a FileMaker user for my own needs I would be inclined to tackle the taxonomy first in that, charging this issue as a separate consultancy. Once this is largely settled SQL tables can be set up on-line (which, incidentally FileMaker can also use locally) and the PHP/HTML started.

I wrote a very simple “getting started” article for a client recently, pm me if you would like a copy.

Ian

Taken to DM for thanks to Ian already but thank you again here - thanks also to @freddo - will investigate though suspect Google Cloud Platform is a bit less user friendly than something like Canto.

If anyone else thinks of anything feel free to chip in. I suspect no software will be able to analyse an image accurately enough and discern between a Quarter Pounder or a Double Cheeseburger and will instead just return results for 'burger' so we're looking at a metadata based solution with manual human inputs. We want some sort of storage and retrieval solution that with a few quick toggle-able options can return a smaller pool of results (ideally with thumbnails). Again if anyone knows an out of the box solution, great, if anyone knows a solution that's part of Google's app store, even better, if someone is a developer and reckons they could build something bespoke, that's great too.

An update on this - looked at some functions from Adobe Marketing Cloud but this was deemed too expensive by client (I wasn't aware of price so it's something they had looked at themselves).

Tried Canto but it's taken them nearly 10 days to respond to my request for info about pricing.
I've had several emails from them but all skirting the issue of pricing when that's all I care about.
It's like when I joined a gym last year - no pricing information on their website and even when I went to the place they spent about 20 minutes avoiding talking about costs. I don't care what it costs, I just want to know, fuck heads. Also with Canto and so many other companies like them - why have a Pricing tab/page on your website if all it does is re-direct you to a contact form?

So in summary I'm still looking and the Google Sheets doc is getting larger. Had a meeting with the client and their digital team yesterday who would definitely like something with thumbnail previews of assets if it helps narrow down.

@velocio you're at Cloudflare I know but anything jump to mind for you these days?

The two most obvious solutions (given that I don't know the competitor landscape for corporate video cataloguing software):

1) Free: You use a Google Form for submitting videos, and hard-code the drop-downs into that... it populates a Google Sheet, which you can create a basic form for searching or wire it up to Excel.

2) No more than £10k: You get a developer to make you a bespoke web app that has a well-tuned form to do the data entry, provides a simple listing screen, and stores a catalogue in a database with perhaps an API. This will take 1 month of dev work if you know what the form requirements, and search requirements look like.

If you only have 1 user, there are lots of personal video catalogues.

If you have multiple users then you're in the realm of corporate software.

Though I wonder... does Plex allow custom metadata? Hah, yes! https://support.plex.tv/hc/en-us/articles/201272763-Edit-Details

Hehe... I'd implement Plex!

Did I ever ask what your budget was? If it's below £10k... yeah, it's all DIY, but the Plex solution looks like a blast. Surely wouldn't be too hard.