frame rate and shutter speed

[offbeatbryce]

I honestly can't remember where I read this but it was somewhere online. Something about how if you are filming at 30fps your shutter should be 1/60th. Is that always the case? What if your camera doesn't shoot exactly 30fps but rather 29.97? I have usually filmed with shutter 1/30th using 29.97.

 

[Alex H.]

Traditionally, in film, the shutter was set at 180° (twice the frame rate) unless a special shot or overall production look required something different. With film running at 24fps, a 180° shutter would be 1/48.

In doing everything possible to get the "film" look in the digital world, folks tend to try and shoot at 24p with a 180° shutter, but many cameras can only get close to that with 1/50, which works fine. I tend to use 1/60 under most artificial lighting sources to eliminate flicker or pulsating.

Plus, the slower the shutter speed, the more the motion blur. So, for you shooting at 30p, 1/30 can add more motion blur and thus result in a softer-looking image. 1/60 may help you keep a sharper image.

But then, there are times that call for higher shutter speeds, such as the "Saving Private Ryan" look. Just remember that a higher shutter speed means less light entering the camera.

The "180° rule" isn't really a rule at all. It's a basic starting point. Just like frame rate, aperture, and lens/focal length, pick your shutter speed for the look you want. Light accordingly.

You might want to look into the 24p and Frame Rates tutorial video from David and Barry. It'll go into greater detail not just about frame rates but how shutter speed affects the image.

 

[offbeatbryce]

Traditionally, in film, the shutter was set at 180° (twice the frame rate) unless a special shot or overall production look required something different. With film running at 24fps, a 180° shutter would be 1/48.

In doing everything possible to get the "film" look in the digital world, folks tend to try and shoot at 24p with a 180° shutter, but many cameras can only get close to that with 1/50, which works fine. I tend to use 1/60 under most artificial lighting sources to eliminate flicker or pulsating.

Plus, the slower the shutter speed, the more the motion blur. So, for you shooting at 30p, 1/30 can add more motion blur and thus result in a softer-looking image. 1/60 may help you keep a sharper image.

But then, there are times that call for higher shutter speeds, such as the "Saving Private Ryan" look. Just remember that a higher shutter speed means less light entering the camera.

The "180° rule" isn't really a rule at all. It's a basic starting point. Just like frame rate, aperture, and lens/focal length, pick your shutter speed for the look you want. Light accordingly.

You might want to look into the 24p and Frame Rates tutorial video from David and Barry. It'll go into greater detail not just about frame rates but how shutter speed affects the image.

Interesting, thanks. I've never noticed much blur on my camera at 1/30. Does the blur show up in only certain certain situations? I usually interview people and don't see much motion blur.

 

[Alex H.]

I've never noticed much blur on my camera at 1/30. Does the blur show up in only certain certain situations? I usually interview people and don't see much motion blur.

The faster the motion, the more the visible blur with slower shutter speeds.

For a seated interview with minimal motion, you might not notice it very often. Shoot faster motion (people running, driving shots, shaky handheld, etc.) and you'll see it. I'd be interested, though, to see if you can spot the difference between 1/30 and 1/60 with a seated interview. You ought to set up and light a mock interview. Shoot a couple minutes at 1/30 and a couple at 1/60. Put them side-by-side in post and see if you can spot any difference.

 

[offbeatbryce]

The faster the motion, the more the visible blur with slower shutter speeds.

For a seated interview with minimal motion, you might not notice it very often. Shoot faster motion (people running, driving shots, shaky handheld, etc.) and you'll see it. I'd be interested, though, to see if you can spot the difference between 1/30 and 1/60 with a seated interview. You ought to set up and light a mock interview. Shoot a couple minutes at 1/30 and a couple at 1/60. Put them side-by-side in post and see if you can spot any difference.

tested this like you suggested. There is a slight difference. I did notice that syncing audio from a tascam to the video at 1/30 resulted in better looking results than when I went to sync the tascam audio to the shot at 1/60.

I even tested this using two camera's. two camera's were side by side. Did a test where both cameras shot at 1/60. Synced them up using the automatic syncing in premiere to the wavforms. Looked like it drifted out of sync between both camera's. When I did a second test with both camera's at 1/30 it didn't look it drifted at all.

also, I've had issues in the past where someone told me my video looked out of sync with the external audio. That was when I used to shoot 29.97fps at 1/60. Ever since I switched to 1/30 at 29.97fps I've had no complaints of external audio looking out of sync.

both my tascam recorder test and my two camera sync test recorded both audio at 48khz at 24bit.

 

[Alex H.]

I did notice that syncing audio from a tascam to the video at 1/30 resulted in better looking results than when I went to sync the tascam audio to the shot at 1/60.

both my tascam recorder test and my two camera sync test recorded both audio at 48khz at 24bit.

This really makes no sense whatsoever, as the shutter speed should have absolutely no effect on audio recording and reproduction. Shutter speed only affects the amount of time each frame of video is exposed, but has no impact on the length of the frame (which is determined strictly by frame rate). So, one frame of 30p video exposed at 1/30 is exactly the same length on playback as one frame of 30p video exposed at 1/60.

Just curious... what two cameras are you using? And which Tascam recorder? This may prove a theory that I have...

Audio does not rely on time code or frame rate, and shutter speed, focal length, and aperture are nowhere near the sound side of things in-camera. The only things that matter for sound are bit depth and sample rate.

Now, where time code does come into play is in sync sound, where a time stamp is embedded in the metadata to identify the start time of the sound file, but again the TC and frame rate are largely irellevant after that. Audio in a TC-synced recorder alongside a camera relies instead on the word clock, a timing crystal that regulates the true sample rate between the camera and the recorder.

But for what you're doing (though until I know which cameras and which Tascam, I am only assuming here), my guess is non-sync? Meaning, your Tascam doesn't utilize TC and is not synced to your cameras. Here is where things can get tricky with sound. Timing crystals are not uniform from one device to the next. With non-sync, the recorder and the camera may be recording sound at slightly different sample rates (just above or just below a true 48kHz). BUT, when placed on a timeline in Premiere, the audio is played back at a true 48,000Hz. So, anything recorded at a rate slightly above 48kHz is actually playing back slightly slower, and anything recorded at a rate just slightly under 48kHz is playing back slightly faster. This is why we experince drift in non-sync sound.

My theory in your 2-camera test is that both cameras were running a slightly different offsets, and the Tascam was running at an offset that was more compatible with one of the cameras than the other. I wonder if your results would differ using only the one camera that stayed closest in sync wtih the Tascam in your test.

also, I've had issues in the past where someone told me my video looked out of sync with the external audio. That was when I used to shoot 29.97fps at 1/60. Ever since I switched to 1/30 at 29.97fps I've had no complaints of external audio looking out of sync.

Again, I think correlation is not causation here. I wonder if something else changed in your workflow. Again, shutter speed has absolutely no effect on sound recording. None.

 

[offbeatbryce]

This really makes no sense whatsoever, as the shutter speed should have absolutely no effect on audio recording and reproduction. Shutter speed only affects the amount of time each frame of video is exposed, but has no impact on the length of the frame (which is determined strictly by frame rate). So, one frame of 30p video exposed at 1/30 is exactly the same length on playback as one frame of 30p video exposed at 1/60.

Just curious... what two cameras are you using? And which Tascam recorder? This may prove a theory that I have...

Audio does not rely on time code or frame rate, and shutter speed, focal length, and aperture are nowhere near the sound side of things in-camera. The only things that matter for sound are bit depth and sample rate.

Now, where time code does come into play is in sync sound, where a time stamp is embedded in the metadata to identify the start time of the sound file, but again the TC and frame rate are largely irellevant after that. Audio in a TC-synced recorder alongside a camera relies instead on the word clock, a timing crystal that regulates the true sample rate between the camera and the recorder.

But for what you're doing (though until I know which cameras and which Tascam, I am only assuming here), my guess is non-sync? Meaning, your Tascam doesn't utilize TC and is not synced to your cameras. Here is where things can get tricky with sound. Timing crystals are not uniform from one device to the next. With non-sync, the recorder and the camera may be recording sound at slightly different sample rates (just above or just below a true 48kHz). BUT, when placed on a timeline in Premiere, the audio is played back at a true 48,000Hz. So, anything recorded at a rate slightly above 48kHz is actually playing back slightly slower, and anything recorded at a rate just slightly under 48kHz is playing back slightly faster. This is why we experince drift in non-sync sound.

My theory in your 2-camera test is that both cameras were running a slightly different offsets, and the Tascam was running at an offset that was more compatible with one of the cameras than the other. I wonder if your results would differ using only the one camera that stayed closest in sync wtih the Tascam in your test.



Again, I think correlation is not causation here. I wonder if something else changed in your workflow. Again, shutter speed has absolutely no effect on sound recording. None.

one camera is a canon xa10 the other is xa25. Both were recorded at 24 bit 48khz. Tascam 44WRL recorder. Audio was recorded at 48khz. I can confirm that for you with screenshots.

I just take the tascam audio and drop it in and put a marker at the start of the clap of my hands in front of one of the cameras. Then I take the recorded audio from the tascam and sync the clap from the camera showing my hands to the tascam audio. Sync is fine for the first minute but after that it drifts unless i'm at 1/30 shutter

 

[Alex H.]

one camera is a canon xa10 the other is xa25. Both were recorded at 24 bit 48khz. Tascam 44WRL recorder. Audio was recorded at 48khz. I can confirm that for you with screenshots.

No need for screen shots. I can guarantee that none of the three are recording at a true 48kHz (or, 48,000Hz). Yes, the cameras and the recorder have selections between 44.1kHz, 48kHz, and 96kHz. However, what I explained earlier was that the timing crystals in these devices aren't spot-on-accurate. I promise you, one of them may be recording at something like 47,097Hz and another at something like 48,003Hz. The offsets may be even higher than that. And in my experience, Canon HD camcorders and Tascam audio recorders drift like a loveless marriage.

I just take the tascam audio and drop it in and put a marker at the start of the clap of my hands in front of one of the cameras. Then I take the recorded audio from the tascam and sync the clap from the camera showing my hands to the tascam audio. Sync is fine for the first minute but after that it drifts unless i'm at 1/30 shutter

Again, I really think this is a sample rate offset issue between the two cameras and the recorder. Shutter speed has absolutely no effect on audio sample rate.

 

[Run&Gun]

Something I stumbled across several years ago… Make sure the camera and recorder are actually using the same frame rate for timecode. Some cameras are switchable(like the C300) and some cameras actually output a different TC than you would think based on the shooting framerate(P2 VariCam). Found it out the hard way the first time we did a multi-cam shoot with two VariCam's, a C300 and SoundDevices 664. We were shooting 24P(23.98) on all cameras, but the Vari's still output 30(29.97) timecode and we didn't know that the C300 could output either 30(29.97) or 24(23.98) and of course the 664 can output and receive lots of different TC rates. Luckily we were just testing the 664's recording function at the time and not being able to have TC synced audio files didn't matter.

Bottom line, make sure you know the real TC rates of the equipment you're using and as Alex said, shutter speed should have ZERO effect on TC, framerate, sample rate, etc.

 

[Alex H.]

Make sure the camera and recorder are actually using the same frame rate for timecode.

You do realize, with the cameras and recorder in question, we aren't dealing with devices that even have the ability to sync to TC, right? The Tascam DR-44WL neither has TC or word clock i/o, nor the ability to utilize different frame rates. This is only going to confuse the situation.

 

[Run&Gun]

You do realize, with the cameras and recorder in question, we aren't dealing with devices that even have the ability to sync to TC, right? The Tascam DR-44WL neither has TC or word clock i/o, nor the ability to utilize different frame rates. This is only going to confuse the situation.

I was just throwing it out there, because it's something that can and has tripped up many an experienced person. If the OP's gear doesn't fall under that, then "let's move along, nothing to see here…".

 

[Razz16mm]

29.97 IS 30fps. It is the NTSC timing offset for 30fps left over from analog days. (30/1.001). A difference of one frame in 1000.

 

[Barry_Green]

29.97 IS 30fps. It is the NTSC timing offset for 30fps left over from analog days. (30/1.001). A difference of one frame in 1000.

Which will cause a visible sync error in about a minute. Your footage will be two frames out of sync with your audio after a minute, if you try to use 30fps timecode. If you're doing short takes, no big deal. If you're doing a three-minute wide shot, it's a very big deal. 29.97 is NOT 30fps, they are different for a reason, there is a very good reason that Drop Frame timecode was invented.

 

[David Jimerson]

And in my experience, Canon HD camcorders and Tascam audio recorders drift like a loveless marriage.

^^^
Dialogue-worthy.

 

[offbeatbryce]

Anyway, I recently made a documentary film shot at 29.97fps and a 1/30 shutter. Everyone gave a standing ovation including award winning filmmakers. Explain how that's possible if my shutter was wrong.

 

[Mark Watson]

In case you missed it...

... The "180° rule" isn't really a rule at all. It's a basic starting point. Just like frame rate, aperture, and lens/focal length, pick your shutter speed for the look you want. Light accordingly.

 

[deltoidjohn]

This really makes no sense whatsoever, as the shutter speed should have absolutely no effect on audio recording and reproduction. Shutter speed only affects the amount of time each frame of video is exposed, but has no impact on the length of the frame (which is determined strictly by frame rate). So, one frame of 30p video exposed at 1/30 is exactly the same length on playback as one frame of 30p video exposed at 1/60.

It can have a minute effect in those instances where the audio always seems to be 1 frame off in either direction. The motion blur means you can be be slightly off while still appearing to be perfectly in sync, because the longer time sample of the 1/30th frame means more micro-expressions are carried

As a practical example, say the visual strike of a clapperboard occurs in between the 1/60th exposures - this will result in your audio always appearing slightly off, because no amount of sliding back and forth by one frame will cause it to sync perfectly. With a 1/30th (360 degree) exposure, that clapper strike will always be recorded - albeit more blurred - and the audio can always be synced to the frame where the strike occurs.

Of course the way around this is to use audio time units to slide your sound in fractions of a frame-length, and only the most astute audience members would ever notice anyway, but you can't completely throw away the idea that motion blur has no effect on sound syncronisation.

For an exaggerated, real world demonstration, refer to any of the many 90's grunge music videos which use the slow-shutter/slow frame rate effect.