Calculating pitch speeds from video

[BLB]

I know the frame rate of my camera (30 frames per second) and can use software to click through the video one frame at a time to count how many frames elapse through any sequence. .


I have Adobe Premiere which makes it easy to count frames,

For video analysis, depending on the video source (camcorder), you can load a video clip into Premiere Pro and change the speed setting of the clip to 50%. It will give you a 60 fps simulation by detinterlacing (separating) the 2 fields from each frame and make them 2 separate frames thus doubling the number of fps. Note, it's easier using a Canon Elph 300HS which can be set at either 120 or 240 fps all for a very affordable price. As for pitch speed, get a radar gun.

 

[guero_gordo]

Not clear why you would count frames instead of just using the elapsed time clock on the tape, maybe I'm dreaming on the precision.

Your math looks fine, and based on your calculated speeds, the block-o-tron came back with these travel times. In reality, of course, P releases well in front of the rubber.

Travel time from 40 feet at 43mph would be .63
Travel time from 40 feet at 48mph would be .56

 

[robinh]

Another member sent me a video to test. He knew the roughly what a radar gun showed at the game, but I didn't, nor did I know who the pitcher was. He did tell me that it was a state championship game. This was a little tougher because I had to estimate distances, but the pitchers circle helped, since her release was closer to the circle than the rubber. I estimated 37 feet to cross the plate. The video was from a high angle behind and to the left, though the backstop, so not ideal. I did the best I could and came up with speeds of 60 and 63 on three different pitches. I thought likely I was wrong, that seemed high for even a HS championship, although not out of the realm of possibility. I sent him my results and he replied that he had been told be the radar gun operator at the game that she was throwing 60+. The pitcher was Erin Gabriele, committed to Tennessee next year. So it appears that my system can at least get the ballpark speed from a decent video. If anyone else has a video that they know the actual speed on, and would like to send me to test, it would be fun to try it. Don't tell me the answer ahead of time, see how close I can come.

 

[Perfect Circle]

I don't have a radar gun and DD has never had a gun put on her pitches. I wanted a baseline speed as we start this off season so that we can measure her progress. So I had the idea that I could calculate it from video. I know the frame rate of my camera (30 frames per second) and can use software to click through the video one frame at a time to count how many frames elapse through any sequence. Then we measured exactly how far from where DD released the ball (six o clock position) to where I was catching it. (I made sure to be consistent with where I held my glove.) I set up the camera on a tripod far enough away to film the entire pitch and had her pitch 30 or so pitches.

Using this method I found her pitch speed to range 43 mph to 48 mph. This seems reasonable to me, she's 14 but pitches on the slow side for her age, the reason we are working on speed this off season.

Here's the details of what I did:

I came up with this formula and plugged it into a spreadsheet. Speed in Feet/Second = Distance in Feet/(Elapsed Frames/Video Frame Rate in frames per second.) Speed in MPH = Speed in FPS * 0.681818181818182, so that is added to the formula to convert the result to MPH.

This formula obviously calculates the average speed over the entire flight of the ball, so I did a little research and found that a softball loses 1 mph for every seven feet it travels. I added a line to the spreadsheet that uses that to calculate the speed at the release (Average speed in MPH + ((distance in feet /7)/2) = speed at beginning of flight.)

I loaded the video on my PC and counted frames pitch by pitch. I have Adobe Premiere which makes it easy to count frames, but you can do it in Quicktime or Windows Media Player, and probably most video players.

I found I could even count fractions of frames, it was obvious when the ball only traveled half the distance in the final frame that it had been traveling in previous frames. I rounded to the nearest half frame, I figure that's plenty precise enough for my purposes.

The spreadsheet allows me to enter any distance and frame rate, then it's very simple to plug in the frame count of a pitch and see the average and top speed for the pitch

Any flaws in my logic or formulas here?

I apparently can't attach the spreadsheet here, but I will be happy to email if anyone wants to try it or check my math. (request it at robin@robinhickman.com)

Got lotz of video to do... trying to find an way to use these number to translate to MPH...


<video width="512" height="288" controls="controls">
<source src="https://dl.dropboxusercontent.com/u/58923514/Banshees/swing_K.mp4" type="video/mp4" />
</video>

40' / .57sec = 70.175 = *681 = 47.789 + 3 + round off = 51mph

 

[Perfect Circle]

40' / .57sec = 70.175 <*681 > = 47.789 + 3 + round off = 51mph

 

[SoGASoftball]

I think Rick Pauly does this regularly from video. Not sure of his formula? Maybe he'll chime in.

 

[JAD]

While it is possible to calculate pitch speed from video, you are measuring distance and time, not true velocity - similar to a RevFire. Pocket Radars or a Bushnell are cheap alternative that do not require math or slow motion video frame counting. If nothing else ask if any of the other pitchers on your DD's team have one available.

 

[riseball]

I would think that the margin of error with respect to the distance traveled would make any derived estimates fairly inaccurate. I have seen this used for home to 1B, or pop times where distance is not a factor, but never with pitch speed.

 

[schwabach]

While it is possible to calculate pitch speed from video, you are measuring distance and time, not true velocity - similar to a RevFire. Pocket Radars or a Bushnell are cheap alternative that do not require math or slow motion video frame counting. If nothing else ask if any of the other pitchers on your DD's team have one available.

I don't understand this statement. If I know distance and time, then I can certainly calculate the average speed. And it should be possible to calculate speed lost due to air resistance. So, in theory at least, I should be able to calculate the speed at any point between the release point and home plate.

 

[MNDad]

Any formula would read slower than she actually throws so we’d need to add 5-7 mph. Or else we can make things easy and just round it to 60.





(sarcasm font)