TC TC TC, man I wish it was simple....

[pvanstry]

Allright, I consider myself knowledgeable when it comes to TC but since I am working a lot in situations where things I can't explain happens, I would like to comfirm some of my "certainty". So here is a list of what I believe to be true. Please add comment on the list as per there validity.

1- TC information is taking at the beginning of a recording and the value at that time is then stamped into the Metadata field ( TC start or something of the sort. ).

2- Speed of sound is decided by sample rate NOT by TC. TC is simply a clock that runs independently to the actual audio or image and is read and recorded at the moment the recording starts.

3- If two device have quality TC generators that have been jam synced together, jam syncing them, their TC clock should remain stable and in sync for a solid 5-6 hours.

4- regardless if TC frame rate, exactly one minute recorded is exactly one minute in length, even if one was to change the frame rate using something like Wave Agent.

5- If a camera is set to 23.98ndf a d an audio recorder to 24, if both have the same sampling rate audio wise ( 48khz ), only the TC value of each Ike's will be different, but the speed of the audio vs the image will remain constant ( no variation, acceleration of one vs the other ).

6- If an audio file is recording at 48khz and it is imported in a session at 44.1khz, it will be speeded up but the speed variation will be a constant.

Now fir questions:

7- in a situation where a camera is set to 48khz, 29.97ndf, and so is an audio recorder, both files are imported in the same editing software with the same settings, what could cause the audio to accelerate over the duration of a take? Same starting point in sync with the image and the sound on the camera ( wireless hop ) but over the course of a minute then be out of sync, the gap getting wider and wider over time.

8- is there a software that will take video files with TC and audio files with TC and resync them automatically simply by looking at the TC values? Something that will enable then to import the newly synced and linked video and audio files.

9- If a camera and an auditor recorder were incorrectly set at the beginning of the day ( wrong TC frame rate, ex: 23.98 cam and 24 audio ), is there a software or a way of changing the frame rate of the audio files but also have the values corrected in order to sync properly with the cam files?

I know it is a lot o things but I believe these are Infos that would greatly help me and a lot of people.

Thank you for your time and experience in advance.

[ccsnd]

Allright, I consider myself knowledgeable when it comes to TC but since I am working a lot in situations where things I can't explain happens, I would like to comfirm some of my "certainty". So here is a list of what I believe to be true. Please add comment on the list as per there validity.

1- TC information is taking at the beginning of a recording and the value at that time is then stamped into the Metadata field ( TC start or something of the sort. ).

2- Speed of sound is decided by sample rate NOT by TC. TC is simply a clock that runs independently to the actual audio or image and is read and recorded at the moment the recording starts.

3- If two device have quality TC generators that have been jam synced together, jam syncing them, their TC clock should remain stable and in sync for a solid 5-6 hours.

4- regardless if TC frame rate, exactly one minute recorded is exactly one minute in length, even if one was to change the frame rate using something like Wave Agent.

5- If a camera is set to 23.98ndf a d an audio recorder to 24, if both have the same sampling rate audio wise ( 48khz ), only the TC value of each Ike's will be different, but the speed of the audio vs the image will remain constant ( no variation, acceleration of one vs the other ).

6- If an audio file is recording at 48khz and it is imported in a session at 44.1khz, it will be speeded up but the speed variation will be a constant.

Now fir questions:

7- in a situation where a camera is set to 48khz, 29.97ndf, and so is an audio recorder, both files are imported in the same editing software with the same settings, what could cause the audio to accelerate over the duration of a take? Same starting point in sync with the image and the sound on the camera ( wireless hop ) but over the course of a minute then be out of sync, the gap getting wider and wider over time.

8- is there a software that will take video files with TC and audio files with TC and resync them automatically simply by looking at the TC values? Something that will enable then to import the newly synced and linked video and audio files.

9- If a camera and an auditor recorder were incorrectly set at the beginning of the day ( wrong TC frame rate, ex: 23.98 cam and 24 audio ), is there a software or a way of changing the frame rate of the audio files but also have the values corrected in order to sync properly with the cam files?

I know it is a lot o things but I believe these are Infos that would greatly help me and a lot of people.

Thank you for your time and experience in advance.

1-6 seems you have a good grasp on the difference between WC and TC

7. My guess is that they are using final cut. Apple in their infinite wisdom decided that when you import files, if there is a timecode mismatch between audio and video, it will automatically do a pull down regardless of what sample rate you are using. This will speed up or slow down the audio based on TC rather than sample rate.

Brilliantly, there is no way to turn this "feature" off.

8. most pro video editing programs do this

9. http://www.sounddevi...ducts/waveagent- This will allow you to change your TC value so apple doesn't ruin your audio. It's free.

Final cut is the only program I know that creates this issue, and it is completely moronic.

[pvanstry]

#7 is amazing, thanks a lot CCalandro!!!! Got our DIT and i going nuts on a friend 552 files and C300. They were in sync at the begining but would change speed over time and rapidly out of sync during the same take.

#8, I mean i would like a dedicated software that does that, i can't remember but a DIT on a set had something and can't find it anymore.

#9 I know that i can change the Frame rate stamps, but i am looking at a software that will recalculate the value of the TC start stamp with the new Frame rate value, basically making the files like if i had rolled at the correct frame rate so it does sync correctly using the TC only.

Thanks a lot.

[pvanstry]

#8, i foundation the software, it's Sync N link. But beware it has problem with Dropframe code.

[Philip Perkins]

So basically the camera dept hosed you by setting their camera to a diff TC rate than the one you used (or they THOUGHT they were using). +1 for Sync N Link. Why was anyone using dropframe TC in production? It is possible to import audio into FCP7 with a diff TC rate than the "project" setting (hopefully what the camera was set at) and have it not pull-up or down. Easiest deal however is to batch change the TC flag on your audio files in Wave Agent or BWF Widget to match what they did. As a desperate measure you can convert your audio files from BWAV to regular wav files w/o TC metadata and then FCP will play them at the correct speed. At that point PluralEyes is your friend.

philp

[studiomprd]

1. yes, in file based sound

2. " Speed of sound is decided by sample rate NOT by TC. "

Speed of sound is decided by word clock, NOT by TC or sample rate.

3. if the clocks are up to SMPTE spec's, and working properly

4. " exactly one minute recorded is exactly one minute in length " yes, but! ... not " if one was to change the frame rate " (this is where it isn't so simple, and where the professionals earn their professional rate

5. " the speed of the audio vs the image will remain constant " yes, unless you make changes...

6. not necessarily, it depends

7. " what could cause the audio to accelerate over the duration of a take? " accelerate is constant increasing speed..?? perhaps you mean run at a faster rate (a constant, not accelerating continuously).. well there are a lot of things that might cause the rates not to match, one being integer, and the other being non-integer. that goes to changing the frame rate so that one is actually going at a slightly (.1%) different rate

" in sync at the begining but would change speed over time and rapidly out of sync during the same take. "

you are almost certainly describing this incorrectly. they start in sync but drift apart over the take because one is running at a constant, but slightly different (by .1%) rate

8. " is there a software that will ... " probably

9. " If a camera and an auditor recorder were incorrectly set " " a software that will recalculate the value of the TC start stamp with the new Frame rate value, "

there maybe a program like that, but probably not. sorry, mistakes happen, and it takes time/effort/money to correct them

Ah, Ha...

" these are Infos that would greatly help me and a lot of people. "

this is nothing new, and is stuff that high paid professionals need to know.

the .1% difference in speed happens when the frame rate of an element is changed. at the non-integer rates, the 30th (or 24th) frame are completed in the next real time clocks second, and yes, that means the contents of that element are slowed down.

" problem with Dropframe code. "

sometimes there are non industry standard drop frame codes. the only DF code that is industry standard is 29.97DF, which is a counting thing, --not a frame dropping thing-- dropping certain frame numbers, much as the calendar drops 29 February 3 out of 4 years!.

This is all stuff you should know, and requires some reading... Unlike Final Cut, (and like PluralEyes) Google is your friend!

" So basically the camera dept hosed you by setting their camera to a diff TC rate than the one you used "

that is one way of seeing it...

and if it were simple, then we wouldn't be worth the fancy rates we expect.

[Andy P]

#8..

here's how to do it in FCP.

[pvanstry]

Senator, very informative post.

But I want I clarify something, when you say that the speed of the files is dictated by word clock and not sample rate, I am confused. Word clock is simply a way of locking two digital devices to have exactly the same sample rate by making one of the device slave ( not using its internal word clock ) and the other the master ( by connecting them together so that the master shares its word clock with the slave. Unless I am wrong here ( I don't believe I am ) we simply said the same thing in different way. I implied within one device and you implied with multiple devices.

Makes sens?

Thanks

[ccsnd]

Senator, very informative post.

But I want I clarify something, when you say that the speed of the files is dictated by word clock and not sample rate, I am confused. Word clock is simply a way of locking two digital devices to have exactly the same sample rate by making one of the device slave ( not using its internal word clock ) and the other the master ( by connecting them together so that the master shares its word clock with the slave. Unless I am wrong here ( I don't believe I am ) we simply said the same thing in different way. I implied within one device and you implied with multiple devices.

Makes sens?

Thanks

A word is (in this case) a 16 or 24 bit series of numbers.

I word clock is the source that triggers the data capture.

The word clock sends a on off voltage trigger that tells the computer to sample the audio and capture the data. If the word clock is set at 48k, that happens 48k times / second

If you try to connect digital IO with 2 seperate clocks, the minute difference between the clocks will create clicking and popping.

When you use WC output into another devices WC in, you are just sending that devices ON/OFF reference to the new device so they share a similar clock. In this case you are bypassing the internal clock on the clocked device.

I hope this helps and was not too confusing.

[Marc Wielage]

#7 is amazing, thanks a lot CCalandro!!!! Got our DIT and i going nuts on a friend 552 files and C300. They were in sync at the begining but would change speed over time and rapidly out of sync during the same take.

That tells me that the camera probably has a drifting internal clock crystal used to reference its own timecode generator, or it was set to the wrong rate (or both). How much of a drift would answer the question. If it's 3 frames a minute, that's almost always a sign that one is set to an integer value (like 30.00 or 24.00), and the other is set to a non-integer value (like 29.97 or 23.98).

You can avoid this situation by using an external timecode box and forcing the camera to jam to that every so often, or actually running the camera with external timecode. Chances are, the box will be far more accurate than the crappy generators found inside the camera -- though there are a few good ones, like the ones in the Arri Alexa. I've also had good luck with the Sony F35, but terribly luck with the cheaper Sony F3.

Don't forget the need to always, always do a workflow test prior to a shoot: roll the camera, roll sound, get a timecode slate, clap at the head, run for about 3-4 minutes, then clap at the tail. Take the files and load them into an editing program, like them up, and see if they match and stay precisely in sync to the frame. If they don't, re-examine the reference issue, try different settings, reboot the equipment and try again. You'll eventually find the right settings that will work. And make sure the post department and/or editor is involved in these tests and decisions.

If I were confronted with a constant drift after the fact, I'd load the audio files into Pro Tools and try to come up with a setting for TCE (time compression/expansion) that would solve the problem. I would start with a .1% speed change (plus or minus) and see what works just through trial and error. Speed change/drift problems can be horrifically frustrating, but there are solutions; I can recall a project where the crew mistakenly set some cameras to 25fps and others to 24fps, and another where they set a film camera to 23.98, not knowing that the film speed is always pulled down .1% in telecine. We were eventually able to get it all synced up, but only with a lot of cursing and turning over chairs and beating our heads on the wall for several hours.

[Jay Rose]

Speed of sound is decided by word clock, NOT by TC or sample rate.

Speed of sound is determined by density of medium. It's approximately one foot per millisecond in air at normal room temperature / normal barometric pressure.

Rate of audio playback, compared to rate of record, is something else entirely.

Sorry, couldn't resist.

[VM]

You can avoid this situation by using an external timecode box

552 files and C300

Marc is right : A C300 absolutely needs a timecode box.

[Marc Wielage]

Yeah, I think the C300 -- while it can make good pictures -- is still a borderline toy camera. My opinion is that it's simply a redesigned EOS-D5 in a large case, optimized for filmmaking, but has a cheap audio section and a marginal timecode circuit. But Sony is no better with the F3, and my fear is that the upcoming F5 and F55 won't be much better for timecode stability.

[studiomprd]

" I implied within one device "

each device has its own word clock... CC explained it.

Jay is correct about the speed of sound, aka Mach1.

the speed of sound recording is what I was addressing.

no matter what the recording frame rates are, it is always real-time unless and until something gets pulled (up or down).,

[pvanstry]

I understand WordClock no problem, as sampling rate, an digital audio intricacies. What as always been a little grayer was framerate. But going thru notes, passed experiences, reading, going thru the answers in this post etc... As confirmed a lot of gray area I was operating with. I don't like not being sure. What really confused me was that an audio file was played back faster then it had been recorded an that everyone was saying that no sampling rate change had been done or Time compression expansion either. This was making me doubt some of my TC beliefs. TC has nothing to do with speed change of a file. It is simply an Independant clock that runs at a specified framerate. When one press records, the recorder looks at the exact value of the TC clock and stamps the file with a TC Start value and records it into the metadata field for it.

Thanks guys for all the info.

[ccsnd]

I understand WordClock no problem, as sampling rate, an digital audio intricacies. What as always been a little grayer was framerate. But going thru notes, passed experiences, reading, going thru the answers in this post etc... As confirmed a lot of gray area I was operating with. I don't like not being sure. What really confused me was that an audio file was played back faster then it had been recorded an that everyone was saying that no sampling rate change had been done or Time compression expansion either. This was making me doubt some of my TC beliefs. TC has nothing to do with speed change of a file. It is simply an Independant clock that runs at a specified framerate. When one press records, the recorder looks at the exact value of the TC clock and stamps the file with a TC Start value and records it into the metadata field for it.

Thanks guys for all the info.

Exactly.

You are ahead of a lot of other people in your TC/WC knowledge.

If audio plays back at a different speed than was recorded, you are probably using final cut - and that there's your problem.

[pvanstry]

Yep man, that sample rate change ( pull up or down ) in Final Cut due to a mismatch in framerate bethween file ( being imported ) and session was the problem. What a stupid thing to do. There must be a rational reason for it but WHAT? Anyway thanks a lot for all the info ( including the senator ).

I am so glad we have this forum!!!!

[ccsnd]

There must be a rational reason for it but WHAT?

You would think, but there isn't.

not that I can find.

[Philip Perkins]

You would think, but there isn't.

not that I can find.

They added this on purpose when Apple bought CinemaTools and incorporated it into FCP. They thought it would be a helpful thing for people editing movies to have an automated "pull" feature--instead it turned into a major headache that still screws us up today. I'm hoping FCP 10 doesn't do this, but so far none of my clients are using it and I haven't bought it so I don't know for sure. That feature needs to die.

philp

[studiomprd]

" TC has nothing to do with speed change of a file. "

er...

I wouldn't say that.

[VASI]

Scenario Time Code (for EU and US Broadcast)

Camera: 1080i50

Sound: 25fps EU

Camera: 1080i60

Sound: 29.97 US

Right?

[Marc Wielage]

Not necessarily. I assume what you really mean is:

Camera: 1080i 59.97 [commonly referred to as "60Hz," but it's really not], 29.97ndf timecode

Sound: 29.97ndf

That will work. But generally, only reality shows and documentaries shoot at this rate. And there are some news-related shows that shoot 29.97 drop-frame, which is yet another complication.

Some cameras and recorders have a true 60 Hz setting, but I wouldn't advise using it unless you have absolute confirmation in writing from post that this is what they want. And I would warn them "you'll be sorrrrrrrrrrry!"

[VASI]

Thank you Marc!

[studiomprd]

" Right? "

sometimes

[VASI]

" Right? "

sometimes

or Left?