Frame rate calculator
NTSC video runs at 29.97 fps instead of 30, a legacy of fitting colour into the black-and-white broadcast signal in 1953. Over an hour that 0.1% difference accumulates to 3.6 seconds of drift between timecode and real time. Drop-frame timecode fixes it by skipping frame numbers, not actual frames; at the start of most minutes, which is why professional timecode uses semicolons and why the numbers occasionally jump.
Duration is frame count divided by frame rate. Fourteen hundred and forty frames at 24 fps is exactly 60 seconds, or timecode 00:01:00:00. It runs the other way too: paste 00:01:30:12 at 25 fps and it comes back as 2,262 frames.
How to convert frames to time
Conforming 24 fps film to 25 fps for European broadcast simply plays it 4% fast: the classic PAL speedup, which shortens a two-hour film by nearly five minutes and raises the pitch of the audio by most of a semitone unless corrected. Going the other way, 3:2 pulldown spreads 24 frames across 30 by alternating how many fields each frame occupies. Neither is lossless, and knowing which was applied explains most of the odd durations in an archive.
Reading a timecode
SMPTE timecode is HH:MM:SS:FF, and the last pair is frames within the current second counting from zero, so at 25 fps it runs 00 to 24 and never reaches 25. That is the part that catches people out when they treat it as a fourth clock field. A minute holds 1,500 frames at 25 fps, 1,440 at 24 and 1,800 at 30, which is why the same timecode describes a different amount of footage at every rate and why a conform has to be stated and never assumed.
Drop-frame drops numbers, not frames
This tool counts non-drop. Drop-frame timecode exists because 29.97 fps drifts about 3.6 seconds an hour against the wall clock, and it corrects that by skipping frame numbers at the start of most minutes without skipping any actual frames, which is why professional timecode writes a semicolon instead of the last colon, and why the numbers occasionally jump. Nothing is lost from the picture; only the labels move, so a drop-frame duration matches the clock while a non-drop one matches the frame count, and the two disagree by those 3.6 seconds an hour.
The rates that are not whole numbers
23.976, 29.97 and 59.94 are not arbitrary: each is its whole-number rate multiplied by 1000/1001, a factor introduced when colour was added to NTSC and the subcarrier had to be fitted between the existing audio and video frequencies. That thousandth is the entire reason drop-frame exists, and it propagates into every conform that touches American footage. Enter 23.976 or 59.94 here and the arithmetic treats them as their nearest whole rate, which is the right answer for non-drop frame counting and the wrong one if you need drop-frame numbering or a duration matched to a wall clock.
What people use it for
- Converting a frame count to a duration
- Working out timecode for an edit
- Turning a timecode back into a frame count
- Understanding a conform between frame rates
- Checking clip length in frames
Questions
Divide the frame count by the frame rate. 1440 frames at 24 fps is 60 seconds.
Paste it into the timecode tab with the frame rate. 00:01:30:12 at 25 fps is 2,262 frames.
1,500 at 25 fps, 1,440 at 24 fps and 1,800 at 30 fps.
Frames within the current second, counting from zero. At 25 fps it runs 00 to 24.
24 fps slowed by 0.1% for NTSC compatibility. It is the standard rate for most film-originated video.
A numbering scheme that skips frame numbers to keep 29.97 fps timecode aligned with real time. No frames are actually dropped, only labels.
No, it counts non-drop. Enter 23.976 or 59.94 and the arithmetic uses the nearest whole rate, which is correct for non-drop and not for drop-frame numbering.
Almost always a 29.97 against 30 mismatch. Over an hour that is about 3.6 seconds of difference.
PAL speedup: 24 fps played at 25 is 4% faster, shortening a two-hour film by about five minutes.
24 or 25 for a filmic look, 50 or 60 for sport and slow motion. Match the delivery standard where there is one.
The frames tab shows the same count at 25 and 30 fps beside the answer. 1,440 frames is 60 seconds at 24 fps, 57.6 at 25 and 48 at 30, because the frames are fixed and only the rate they play at changes.
The method for fitting 24 fps material into 30: frames alternate between occupying two fields and three, spreading four film frames across five video ones. It is what makes 24 fps look acceptable on a 60 Hz signal, and it is not lossless.