Metadata Support

[Mike Schell]

We are now discussing the addition of user definable metatdata to the Flash XDR and nanoFlash products. This metadata could include some of the following information: Take Number, Good/Reject, Client, Date, Camera Type, Location, DP, Shutter Speed, Video Format, Bit-Rate, etc. Some of the data would be entered by the user and some would be automatically extracted from the recorder (such as the video format and bit-rate).

We are still debating the best method to enter this data. One possible approach is to create the files on a PC/MAC and upload to the recorder via a CF card. You could define a key word, such as Location and then enter several possible choices (New York, London, Paris). Then in the field, you could select the appropriate choice from a pull-down menu.

After a shoot, all the QT or MXF files (with embedded metadata) could be downloaded to a common file folder on a PC/MAC. Then using our (yet to be developed) metadata extractor, you could create one XML file which has the extracted metadata from all the QT/MXF files. The XML file could then be loaded into Excel for analysis.

The real benefit of metadata can then be realized with yet another software tool we are considering, which allows the user to find all files with a specific attribute, (such as location = London) and copy these files to a new folder.

Anyway, I welcome any thoughts or recommendations, such as whether anyone would actually make use of metadata?

Thanks-

[Mark Job]

Originally Posted by Mike Schell

We are now discussing the addition of user definable metatdata to the Flash XDR and nanoFlash products. This metadata could include some of the following information: Take Number, Good/Reject, Client, Date, Camera Type, Location, DP, Shutter Speed, Video Format, Bit-Rate, etc. Some of the data would be entered by the user and some would be automatically extracted from the recorder (such as the video format and bit-rate).

We are still debating the best method to enter this data. One possible approach is to create the files on a PC/MAC and upload to the recorder via a CF card. You could define a key word, such as Location and then enter several possible choices (New York, London, Paris). Then in the field, you could select the appropriate choice from a pull-down menu.

After a shoot, all the QT or MXF files (with embedded metadata) could be downloaded to a common file folder on a PC/MAC. Then using our (yet to be developed) metadata extractor, you could create one XML file which has the extracted metadata from all the QT/MXF files. The XML file could then be loaded into Excel for analysis.

The real benefit of metadata can then be realized with yet another software tool we are considering, which allows the user to find all files with a specific attribute, (such as location = London) and copy these files to a new folder.

Anyway, I welcome any thoughts or recommendations, such as whether anyone would actually make use of metadata?

Thanks-

....Even better, Mike, this XML Meta data can be used to speak to an NLE, such as Avid Media Composer, or Final Cut Pro. The user programmed meta data can be carried over into the NLE. This data could make the NLE automatically create bins and auto capture-auto load only certain clips into certain bins as the NLE reads the embedded meta data on the CF card. Think about it. This is darn easy to program too. BTW, Avid has done the most research on the use of XML and Meta data. We use XML formatted Meta data right on the timeline to manually add chapter points and thumnail images for button creation from the clips on the timeline which gets sent over to the DVD authoring program. Just an idea. Don't shoot me for it !

[Hank Coffin]

Have you looked at Adobe's XMP public metadata SDK?

Adobe XMP: Adding intelligence to media

There's a push for interoperability among vendors.

[John Richard]

It's an EXCELLENT time saving idea! Huge improvement over having to keep a written log and/or searching thru the current file naming scheme.

[Mike Schell]

Thanks for the feedback and recommendations. We'll take a look at the initiatives from Adobe and Avid. Naturally, we would like to be as compatible as possible.

On the hardware side, we are considering a simple USB to RS485 converter which will plug into your laptop/desktop or sub-notebook (Dell has one for $249). This converter would have USB on one side and a telephone jack on the other side.

We would then create a simple connector which plugs into the 9-pin port on the XDR (or the 6-pin Hirose on the nano) and has two standard telephone jacks. You can then use standard telephone cable to connect to the USB converter and to daisy chain to other XDR/nano boxes.

Using this simple connection, we can then send metadata and/or video assist commands to one one more XDR/nanos from this simple laptop/notebook. The metadata in the all the recorders can then be simultaneously updated. Or, we can send one command to all the XDR/nanos in the chain to simultaneously play back a given clip.

Side-note: RS485 is a simple 2-wire half-duplex communication protocol capable of bit rates up to 10Mbps. In our implementation we would likely limit the speed to 115Kbps, which would allow cable runs in excess of 1000 feet (300 meters).

[Mark Job]

Originally Posted by Mike Schell

Thanks for the feedback and recommendations. We'll take a look at the initiatives from Adobe and Avid. Naturally, we would like to be as compatible as possible.

On the hardware side, we are considering a simple USB to RS485 converter which will plug into your laptop/desktop or sub-notebook (Dell has one for $249). This converter would have USB on one side and a telephone jack on the other side.

We would then create a simple connector which plugs into the 9-pin port on the XDR (or the 6-pin Hirose on the nano) and has two standard telephone jacks. You can then use standard telephone cable to connect to the USB converter and to daisy chain to other XDR/nano boxes.

Using this simple connection, we can then send metadata and/or video assist commands to one one more XDR/nanos from this simple laptop/notebook. The metadata in the all the recorders can then be simultaneously updated. Or, we can send one command to all the XDR/nanos in the chain to simultaneously play back a given clip.

Side-note: RS485 is a simple 2-wire half-duplex communication protocol capable of bit rates up to 10Mbps. In our implementation we would likely limit the speed to 115Kbps, which would allow cable runs in excess of 1000 feet (300 meters).

....Cable runs of up to a 1,000 feet, multiple converters...Hmmmm, well OK. This is just another outside thought, but since you're discussing USB, then I see a possibly more elegant solution (But I don't know if you guys can still do this - Maybe with the Nano ????). Put either a usb 3.0 plug in the Nano, or a small wireless draft "N" Wifi Radio receiver inside the Nano and let the laptops simply broadcast the data in full duplex over 108 Mbps data rate signal into the Nano. For the XDR, a wireless USB 2.0 dongle to telephone thingy. I'm not a great fan of wires.

[Mike Schell]

Originally Posted by Mark Andrew Job

....Cable runs of up to a 1,000 feet, multiple converters...Hmmmm, well OK. This is just another outside thought, but since you're discussing USB, then I see a possibly more elegant solution (But I don't know if you guys can still do this - Maybe with the Nano ????). Put either a usb 3.0 plug in the Nano, or a small wireless draft "N" Wifi Radio receiver inside the Nano and let the laptops simply broadcast the data in full duplex over 108 Mbps data rate signal into the Nano. For the XDR, a wireless USB 2.0 dongle to telephone thingy. I'm not a great fan of wires.

Hi Mark-
Sounds like a very cool approach, but I am very uneasy about multiple wireless devices on a production set, especially when you have 5 or more nano/XDR units to control. I know the wired approach will work and give solid dependable results (once it's fully debugged). It's also much easier to design and much much faster to market.

Best-

[Mark Job]

Originally Posted by Mike Schell

Hi Mark-
Sounds like a very cool approach, but I am very uneasy about multiple wireless devices on a production set, especially when you have 5 or more nano/XDR units to control. I know the wired approach will work and give solid dependable results (once it's fully debugged). It's also much easier to design and much much faster to market.

Best-

....With a possible WiFi solution you can simultaneously broadcast the bits to any number of XDR & Nano box combos you like without fear of cross interfearance on audio channels. I'm in production constantly on our own Tv series, and I can tell you wires are shunned. Our audio is wireless and when actors are running down streets in a blizzard with ice all over the streets we don't want to be connected to much of anything. Inside locations are more feasible for the wired approach, but who wants to if we can get away from it. Continuity and logging on set is done on a laptop, so imagine if we could get the broadcast of Meta data going in *both* directions. Have a Nano automaticall embed Day, Date, Shot, so each time the camera trigger is pressed it sends that XML Clip Meta Data over to the laptop program which auto updates and when the director calls a take, "Good," "Printable," "Safety," then this goes over in the opposide direction, from the laptop, to the box and embeds into the clip.

[Mike Schell]

Hi Mark-
I am in total agreement that wireless would be ideal, but it's simple outside the time frame to get this product completed and shipped. Yes, it is something we can consider for the future, but first we have to complete the basic metadata support code and provide a method for users to enter the data and a program to extract the metadata from the QT and MXF files. We also have plans for a program to find all files based on a certain keyword parameter, such as Client = HBO. All this code takes time to develop and debug.

Cheers-

[Mark Job]

Originally Posted by Mike Schell

Hi Mark-
I am in total agreement that wireless would be ideal, but it's simple outside the time frame to get this product completed and shipped. Yes, it is something we can consider for the future, but first we have to complete the basic metadata support code and provide a method for users to enter the data and a program to extract the metadata from the QT and MXF files. We also have plans for a program to find all files based on a certain keyword parameter, such as Client = HBO. All this code takes time to develop and debug.

Cheers-

,,,,Hi Mike. It's too bad you're locked into a time frame of manufacture which does not allow for a higher level of flexibility to add things at a later date. I understand you folks might be too far along in the manufacturing process on the Nano to make any changes at this point.

[Mike Schell]

Originally Posted by Mark Andrew Job

,,,,Hi Mike. It's too bad you're locked into a time frame of manufacture which does not allow for a higher level of flexibility to add things at a later date. I understand you folks might be too far along in the manufacturing process on the Nano to make any changes at this point.

Hi Mark-
Thanks for the recommendations, but we need to get the nanoFlash shipping ASAP. The nanoFlash does not fulfill everyone's wish list, but we're confident that many videographers/cinematographers/broadcasters will find the nano to be a very useful tool. It is still the smallest professional HD/SD recorder available and the quality of the 100 Mbps 4:2:2 footage continues to amaze most everyone.

[Mark Job]

Originally Posted by Mike Schell

Hi Mark-
Thanks for the recommendations, but we need to get the nanoFlash shipping ASAP. The nanoFlash does not fulfill everyone's wish list, but we're confident that many videographers/cinematographers/broadcasters will find the nano to be a very useful tool. It is still the smallest professional HD/SD recorder available and the quality of the 100 Mbps 4:2:2 footage continues to amaze most everyone.

....Until our SD Solid State Recorder comes out ;-)

[Mike Schell]

Originally Posted by Mark Andrew Job

....Until our SD Solid State Recorder comes out ;-)

Hi Mark-
Best of luck on your development!

Cheers-