Clip Naming Scheme

[Nir Shelter]

I would really like to see an option to choose a naming scheme for the clips. To the best of my knowledge this is not possible currently build 18 maybe....?

I.e time of day based adding on a set of digits derived from the currant time of day to the end of a clip name. Or a successive count adding a digit to the end of a clip name and then adding 1 to that number, starting over when reset by the user. Or a unique ID for each clip based on camera serial number, metadata or project name.

Reason being that if you format one mag using a specific slate and reel number and then latter during the day format another mag using the same slate and roll number the clips will have identical names. For off line and on line work flows this could be a problem.

Can this be done or can this be made possible for future builds?

 

[Cail Young]

Reason being that if you format one mag using a specific slate and reel number and then latter during the day format another mag using the same slate and roll number the clips will have identical names. For off line and on line work flows this could be a problem.

Since build 16 two random characters are added at the end of each clip name and also to each volume name - so you should never* have any clips with the same name.

*well, hardly ever.

 

[Stuart English]

Reel Number sequencing

Reel Number sequencing

if you format one mag using a specific slate and reel number and then latter during the day format another mag using the same slate and roll number the clips will have identical names.

Actually, because each magazine I.D and clip file name include a 2-digit random number, this is almost impossible.

e.g you might have A004_C001_1005A6.R3D but the "duplicate" then would for example be A004_C001_10059F.R3D

The random elements A6 and 9F in this example means the file names are unique.

 

[Nir Shelter]

Thats very good but wouldn't a random selection have an all be it teenee weenee chance but one none the less of being repeated?

 

[kivceshnamun]

Thats very good but wouldn't a random selection have an all be it teenee weenee chance but one none the less of being repeated?

With 10 numbers and 26 letters, it gives you (36!) permutations...or about 371,993,326,789,901,217,468,000,000,000,000,000,000,000 different possibilities

 

[Jörgen Persson]

With 10 numbers and 26 letters, it gives you (36!) permutations...or about 371,993,326,789,901,217,468,000,000,000,000,000,000,000 different possibilities

If the first clip has 0101AA the possibilities for a second clip with the same camera, clipnumber and date is:

One chance in 36 to begin with the letter A
One chance in 36 to end with the letter A.

Combined it's 1/36 x 1/36 = 1/1296 = one chance in 1296.

Which still is a lot but nowhere near the 371 gazillion you mention above

 

[Stuart English]

File Names

File Names

Combined it's 1/36 x 1/36 = 1/1296 = one chance in 1296.

Which is much better than most other file naming schemes ......

At the end of the day there will always be some requirement not to mess reel naming up on the operator's part

(or for that matter not to over-record a video tape / loose a roll of film) :red_bandana:

 

[Nir Shelter]

How ever big or small a chance it is it still exists. The P2 cameras generate a unique clip I.D derived from the camera serial number, clip number, time of day and dat and I think a few other variables. Some of the variables stay the same i.e camera serial but others are in constant change i.e time of day/date. I just think that relying on chance in any capacity is asking for trouble. I have to agree with Stuart it is good to avoid this with correct reel and slate I.D, but better to combine this with a fool proof numbering scheme though.

 

[Stuart English]

Clip Names

Clip Names

How ever big or small a chance it is it still exists. The P2 cameras generate a unique clip I.D derived from the camera serial number, clip number, time of day and date and I think a few other variables. Some of the variables stay the same i.e camera serial but others are in constant change i.e time of day/date. I just think that relying on chance in any capacity is asking for trouble. I have to agree with Stuart it is good to avoid this with correct reel and slate I.D, but better to combine this with a fool proof numbering scheme though.

Thanks for that, I think he original comment was about the visible clip name though.

So a more complete statement would be: in addition to the visible name the metadata contains all of the above.....

.. the issue is what do existing applications look for / display, and the answer to that tends to be 8 or 16 characters.

Hence the file name format chosen A001_C001_0810** which can be collapsed to A001C001 if needs be.

 

[Shawn Nelson]

Those random numbers saved my butt. On Build 16 the camera spontaneously decided to make two mags A053 (back-to-back). It's still a pain to deal with, but those random digits at least made it technically work for me.

 

[Emery Wells]

While duplicate filenames are definitely a major pain we dealt with this on several occasions before B16 introduced the random 2 digit alpha numeric naming scheme. The good news is that if you are in time of day TC you will still have unique TC in which case it will be fairly easy to deal with. The chances that you end up with one duplicate file name is slim. The chances that you end up with an entire reel of duplicates are nearly impossible. Id say this is a non issue for me at this point.

 

[Nigel Hadley]

Actually, because each magazine I.D and clip file name include a 2-digit random number, this is almost impossible.

e.g you might have A004_C001_1005A6.R3D but the "duplicate" then would for example be A004_C001_10059F.R3D

The random elements A6 and 9F in this example means the file names are unique.

Hello,

Here at Digital Vision we have just launched our first version of code with Native R3D support. This has been using a conform method (based on build 15) that looks like this;

A009_C001_080812

All of these elements exist as meta-data within the file so this was a very safe conform process as there was no chance that the "filename" could have been corrupted or changed by someone/something.

Now suddenly we are faced with EDL's like this;
A043_C006_09049D

Where the 9D are random characters? The problem I perceive with this is that you are now vulnerable to the filename itself being altered - whereas before the meta-data was protected.

I looked for the random characters in the meta-data but they don't appear to be there? Is that true?

Regards

Nigel

 

[Stuart English]

Reel numbers

Reel numbers

The reason the "year" values were dropped in favor of a 2 random digits - such as 9D in the above example - is the duplicate file name problem rears its ugly head upstream of the edit bay.

For example I have two camera on a production. I say label one camera A and one B. Its a multi-camera shoot (perhaps stereo) We use Time of Day timecode.

Now that means each camera with generate a magazine with this sequence

A001_C001_100808.R3D A001_C002_100808 A001_C001_100808.R3D
and
B001_C001_100808.R3D B001_C002_100808 B001_C001_100808.R3D

Forget to change the "B" camera I.D to B though and you will have -

A001_C001_100808.R3D A001_C002_100808 A001_C001_100808.R3D
and
A001_C001_100808.R3D A001_C002_100808 A001_C001_100808.R3D

So someone is now backing this media up to a RAID and the computer says
"hey a file name A001_C001_100808.R3D already exists - overwrite " O.K says the data wrangler and a clip, or even a reel is lost.

So now to make the system more fool proof we randomize the last two digits of the file names.

A001_C001_1008A6.R3D A001_C002_1008A6 A001_C001_1008A6.R3D
and
B001_C001_1008FD.R3D B001_C002_1008FD B001_C001_1008FD.R3D

Forget to change the "B" camera I.D to B though and you will have -

A001_C001_1008A6.R3D A001_C002_1008A6 A001_C001_1008A6.R3D
and
A001_C001_1008FD.R3D A001_C002_1008FD A001_C001_1008FD.R3D

Now the above can't happen.

 

[Nir Shelter]

You shouldn't be getting the same name on the "same" card seeing as each usage of a card should be unique. When ever you re insert a CF back into the camera force a new reel number on it before formatting, those ids you mentioned wont be the same as the reel number will successively increase by by one with each reel.

I appreciate the random digits at the end but still would like to be able to choose a naming scheme per job to better suit the post work flow. Preferably even have the... I know it is insane but have the option to choose a naming scheme for clips and choose whether or not the clip name in its entirety or not will be recorded in the metadata. This would be a great thing for data base purposes when shooting large amounts of data for a feature.