who made the mistake? Slow motion

film_autre

film_autre

Hi guys,

im baffled....

artist wanted slow motion, so my deliverable is 25p, we shot at 50fps (so we could slow down to 50fps) the original song tempo was 60 (bpm), according to my math, double frame rate = double tempo/bpm so he supposedly made a 120 bpm version...

we shoot it all etc.. however when trying to sync the video footage interpreted to 25fps the actual video is a bit slower, I can line it up to match a word then after a few words it drifts out of sync and doesnt keep up..

my guess is that he has not bounced the song correctly or something like that?

70% of the shots were slow motion but I do have a usable normal motion selection, I can cut most of the slow motion to fit the song but will miss out on alot of stuff I wanted, but does anyone have any idea what went wrong?

to clarify

songs original tempo was 60, he made a 120 version
videos original delivery frame rate is 25fps, shot at 50fps for the slow motion parts.
song doesnt sync...
 
Sort by date Sort by votes
There are various possibilities. There is one obvious issue however: You filmed at 50 frames per second, the artist created music at 2 beats per second. Your math is correct but you assumed that a second is a constant, it isn't, you haven't defined what a second is! Your camera has defined how long a second is and your artist's DAW has also (and independently) decided how long a second is. So now you've got two different definitions of a second and you're trying to match them up. Generally, I wouldn't expect the two different definitions of a second to become apparent for several minutes, rather than after a few words but by subsequently halving the speed of everything you've probably at least doubled the discrepancy.

G
 
Upvote 0 Downvote
AudioPostExpert said:
There are various possibilities. There is one obvious issue however: You filmed at 50 frames per second, the artist created music at 2 beats per second. Your math is correct but you assumed that a second is a constant, it isn't, you haven't defined what a second is! Your camera has defined how long a second is and your artist's DAW has also (and independently) decided how long a second is. So now you've got two different definitions of a second and you're trying to match them up. Generally, I wouldn't expect the two different definitions of a second to become apparent for several minutes, rather than after a few words but by subsequently halving the speed of everything you've probably at least doubled the discrepancy.

G
Click to expand...

so how does one achieve the effect correctly then? everything iv seen seems to point to what I have done...
 
Upvote 0 Downvote
I don't know how to do the quote thing, but in response to APE's response about what defines the second, I sort of get it (I still remember having to tether the camera to the Nagra to compensate for this) but in the digital age, I thought this wasn't a issue. I record with my BMPCC and Zoom recorder all the time and they stay in sync. What am I missing?
 
Upvote 0 Downvote
AudioPostExpert said:
There are various possibilities. There is one obvious issue however: You filmed at 50 frames per second, the artist created music at 2 beats per second. Your math is correct but you assumed that a second is a constant, it isn't, you haven't defined what a second is! Your camera has defined how long a second is and your artist's DAW has also (and independently) decided how long a second is. So now you've got two different definitions of a second and you're trying to match them up. Generally, I wouldn't expect the two different definitions of a second to become apparent for several minutes, rather than after a few words but by subsequently halving the speed of everything you've probably at least doubled the discrepancy.

G
Click to expand...

Is this where a timecode generator comes into play? Curious.
 
Upvote 0 Downvote
I've done this before without any problems.
Did you record with sound?
In that case you can compare the interpreted sound to the original song.

Next time use your NLE to speed up the track yourself, so you know what was done to it.

Is there a change that the artist made the speed 120% instead of 120bpm? :P
(In that case 41,7 fps would sync...)

Ask for the 120bpm version, double the speed and compare to the original and you'll see whether that is the problem.

APE:
if second a is not equal to second b and both are sped up by doubling the speed and then interpreted at half the speed the difference between a and b would still be the same. It would only show twice as fast, because you have twice as many frames with that difference.
Like you said: it wouldn't show that fast. Unless the difference so pretty big it would cause problems all the time...

Somehow I have to think of using a smpte synctrack on tape to make sure any irregulaties in tape speed wouldn't mess up the sync of various MIDI programmed tracks...
 
Upvote 0 Downvote
WalterB said:
I've done this before without any problems.
Did you record with sound?
In that case you can compare the interpreted sound to the original song.

Next time use your NLE to speed up the track yourself, so you know what was done to it.

Is there a change that the artist made the speed 120% instead of 120bpm? :P
(In that case 41,7 fps would sync...)

Ask for the 120bpm version, double the speed and compare to the original and you'll see whether that is the problem.

APE:
if second a is not equal to second b and both are sped up by doubling the speed and then interpreted at half the speed the difference between a and b would still be the same. It would only show twice as fast, because you have twice as many frames with that difference.
Like you said: it wouldn't show that fast. Unless the difference so pretty big it would cause problems all the time...

Somehow I have to think of using a smpte synctrack on tape to make sure any irregulaties in tape speed wouldn't mess up the sync of various MIDI programmed tracks...
Click to expand...

I have also done this before which is why I'm baffled, I already got the 120 version and it doesn't match when slowed down by 50% I think he has messed up..

Music artists.. Gotta love them
 
Upvote 0 Downvote
well just had a call with the producer and he confirmed he accidently bounced it at 119 BPM.. anyone renting out a shotgun?

on a serious note, anyone got any ideas how I can fix this? im crap at math and just about figured double...
 
Upvote 0 Downvote
film_autre said:
anyone renting out a shotgun?
Click to expand...


Sure, what of you need? :D:lol:


latest



image.php
 
Upvote 0 Downvote
film_autre said:
so how does one achieve the effect correctly then? everything iv seen seems to point to what I have done...
Click to expand...

I'm not sure what you've seen or where you've seen it, but professionally we "sync-lock" cameras and sound systems together during recording/filming and we sync-lock picture and sound together during post. Sync/lock effectively eliminates sync issues, as one piece of kit (masterclock or sync-generator) defines what a second is and then all the other equipment "locks" to that definition.

ILikef11 said:
I sort of get it (I still remember having to tether the camera to the Nagra to compensate for this) but in the digital age, I thought this wasn't a issue. I record with my BMPCC and Zoom recorder all the time and they stay in sync. What am I missing?
Click to expand...

What you're missing is that they don't stay in sync in the digital age! Clocking is certainly more accurate in this digital age than it was in the mechanical age but in consumer audio recorders and DSLRs they're using cheap clocking circuitry and even with sophisticated clocking circuitry there would still be drift, just less drift. Nevertheless, the amount of drift you're experiencing is obviously not enough for you to notice but, as the length of your takes increases, the drift would become more apparent, as it's cumulative. As you can't make clocks accurate enough to always perfectly agree with each other, the only sensible way to ensure avoiding any sync issues is not to be reliant on different clocks agreeing with each other in the first place! Hence the master-clock/slave-clock topology utilised in pro film/TV workflows.

WhiteOpus said:
Is this where a timecode generator comes into play? Curious.
Click to expand...

Yes, or at least partially. A TC-gen is part of the equipment that is required. Also required is an audio recorder and/or camera which can accept an incoming TC/ref signal and modulate it's own internal clock signal to the incoming signal. This is called chase-lock and obviously this functionality comes at a price, an additional price which just isn't justified in cameras primarily designed to take stills . The only exception I know of is the GH4 but you have to buy a relatively expensive option to get this functionality. AFAIK, all pro video/film cameras have this functionality built-in and that's one of the reasons why they are expensive relative to consumer video cameras and DSLRs, same with the better pro field audio recorders.

WalterB said:
if second a is not equal to second b and both are sped up by doubling the speed and then interpreted at half the speed the difference between a and b would still be the same.
Click to expand...

The error would be the same over the same period of time but double overall as the period of time is doubled by halving the speed. And, this is assuming a perfect "interpretation", which is virtually never the case because we are once again relying on an imperfect clock.

film_autre said:
well just had a call with the producer and he confirmed he accidently bounced it at 119 BPM..
Click to expand...

I fairly commonly analyse the music I'm provided to mix. Virtually never is it exactly at the BPM the composer thought he/she was working at, due to clocking inaccuracies. A 120bpm song would typically measure anywhere between 119.5 and 120.5 BPM. No one is going to notice if a song/track is at say 119.9 bpm instead of 120 bpm, unless they try to sync it with something else which is 120 bpm. Music videos are usually shot to the song, so it doesn't matter what bpm the music is at and incidental music usually doesn't need perfect sync and even when it does, music cues often don't last long enough for the drift to be a serious issue. One does sometimes run into this issue with amateur film/TV composers but real pros use sync-locked systems.

film_autre said:
on a serious note, anyone got any ideas how I can fix this? im crap at math and just about figured double...
Click to expand...

Assuming the music is at exactly 119 bmp: Each second (1,000ms) of footage currently equals about 1008.4ms of music. Time compressing the music by 0.833% should sync them up. IE., The music needs to be 99.167% of it's current duration.

G
 
Last edited:
Upvote 0 Downvote
Ask the artist for a 59,5 bpm version? ;)
(Although that would screw up the normal shots.)

Be realistic: it will cause troubles in editing.
The easiest thing would be to reshoot the slowmo part at the artist's expense...
 
Upvote 0 Downvote
You must log in or register to reply here.
Back
Top