Page 1 of 1

A Chronograph Watch Chronograph

Posted: Sun Aug 16, 2009 8:10 am
by boyntonstu
A Chronograph Watch Chronograph

I was thinking about carrying my laptop outside to utilize its sound card to record shots for fps measurements

Audacity is a terrific free program that allows you to see the sound graphs in fine detail and to calculate the velocity of your projectile.

As I was about to use the laptop I glanced at my watch, and I suddenly remembered that it has a built in MP3 recorder.

Darn if this isn't the simplest method to measure velocity with 2 thunks.

To increase the thud sound level at the plywood target I leaned an old aluminum pan against it as a rattle.

I paid about $40 for my watch about a year ago.

Mine has 1 Gb of memory and it can record 27 hours of audio.

It also has rechargeable batteries.

Highly recommended for a very accurate watch and for recording conversations without being obvious.

Check it out:

http://cgi.ebay.com/NEW-Sungale-WMP3218 ... 286.c0.m14

http://tinyurl.com/mdqpkp

UNGALE
Sungale WMP3218 128MB MP3 Watch w/ Earphones
Model: WMP3218-128

Condition: New
Description:
* 5-in-1 MP3 watch player: MP3, USB, recorder, watch & repeat
* Supports MP3 & WMA formats
* Citizen quartz mechanism
* 5 EQ modes
* Built-in, sensitive microphone for voice recording
* 3-10 ATM water-resistant
* Shockproof without sound jumping
* Red, yellow & green light indicators
* Up to 10 hours playback with built-in, Li-Ion polymer battery
* Stereo retractable earphone
* Stylish watch case & ashen dial design
* Comfortable leather wristlet

* Product Type: Flash MP3 Player
* Product Features: Voice Recorder
* Flash Memory Capacity: 128MB
* Audio Formats: MP3, WMA

FWIW Any portable MP3 recorder will work fine.

Place it midway between the muzzle and the target.

Measure the time interval in seconds with Audacity and divide it into the distance in feet.

For example: time is .1 second and distance is 10 feet.

10/.1 = 100 fps.

BoyntonStu

Posted: Sun Aug 16, 2009 12:49 pm
by jimmy101
That's an excellent way to record data in the field and has been discussed before. First time I've heard of a watch with a record function though. Does it save the data as an MP3 (compressed and lossy) or as a WAV (non-compressed and non-lossy)?

With a PDA with a record option you could probably write an app to do the math directly on the PDA, similar to SoftChrono.
Quick someone write an iPhone app ...
It could be done in a web browser with Javascript pretty easily.

Hmm, could also put a web server somewhere to do the number crunching then email (or post) the .wav (or mp3) file to the server and have the server send back the calculated results, graph etc.

Posted: Sun Aug 16, 2009 12:59 pm
by Technician1002
jimmy101 wrote:That's an excellent way to record data in the field and has been discussed before. First time I've heard of a watch with a record function though. Does it save the data as an MP3 (compressed and lossy) or as a WAV (non-compressed and non-lossy)?

With a PDA with a record option you could probably write an app to do the math directly on the PDA, similar to SoftChrono.
Quick someone write an iPhone app ...
It could be done in a web browser with Javascript pretty easily.

Hmm, could also put a web server somewhere to do the number crunching then email (or post) the .wav (or mp3) file to the server and have the server send back the calculated results, graph etc.
You are embarrassing yourself. Read the original post again under specifications.
Audio Formats: MP3, WMA
Nice watch. My concerns are more along the lines of the amount of audio compression. Compression may make some of the audio events not appear as they really are and may make finding the event from echos difficult.

Some compression may attenuate on the main shot and not release until after the impact to an echo of the impact may appear to be the impact. A recorder without any compression is best for this type recording.

I doubt the watch has manual gain controls. If it has no compression, it may clip the recording with the resulting loss of detail of the event.

Posted: Sun Aug 16, 2009 1:01 pm
by ramses
Image

JK, but because you listed all the specs it seemed like it could have been.

I have used this method with the ancient, archaic, Tape recorder. It works rather well.

When I read the title I thought you were going to modify the watch so that the "start" and "stop" buttons of the timer were actuated by photogates, and just read the time off the watch display. I had considered this, but if the photogates were a reasonable distance apart, (like 3ft), the max velocity would be 300fps. (3/.01).

Still, nice watch :D

Posted: Sun Aug 16, 2009 2:25 pm
by boyntonstu
Technician1002 wrote:
jimmy101 wrote:That's an excellent way to record data in the field and has been discussed before. First time I've heard of a watch with a record function though. Does it save the data as an MP3 (compressed and lossy) or as a WAV (non-compressed and non-lossy)?

With a PDA with a record option you could probably write an app to do the math directly on the PDA, similar to SoftChrono.
Quick someone write an iPhone app ...
It could be done in a web browser with Javascript pretty easily.

Hmm, could also put a web server somewhere to do the number crunching then email (or post) the .wav (or mp3) file to the server and have the server send back the calculated results, graph etc.
You are embarrassing yourself. Read the original post again under specifications.
Audio Formats: MP3, WMA
Nice watch. My concerns are more along the lines of the amount of audio compression. Compression may make some of the audio events not appear as they really are and may make finding the event from echos difficult.

Some compression may attenuate on the main shot and not release until after the impact to an echo of the impact may appear to be the impact. A recorder without any compression is best for this type recording.

I doubt the watch has manual gain controls. If it has no compression, it may clip the recording with the resulting loss of detail of the event.
Several bit rates are specified in the MPEG-1 Layer 3 standard: 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 192, 224, 256 and 320 kbit/s, and the available sampling frequencies are 32, 44.1 and 48 kHz. A sample rate of 44.1 kHz is almost always used, because this is also used for CD audio, the main source used for creating MP3 files. A greater variety of bit rates are used on the Internet. 128 kbit/s is the most common, because it typically offers adequate audio quality in a relatively small space. 192 kbit/s is often used by those who notice artifacts at lower bit rates. As the Internet bandwidth availability and hard drive sizes have increased, 128 kbit/s bit rate files are slowly being replaced with higher bit rates like 192 and 256 kbit/s, with some being encoded up to MP3's maximum of 320 kbit/s. It is unlikely that higher bit rates will be popular with any lossy audio codec because file sizes at higher bit rates approach those of lossless codecs such as FLAC.

By contrast, uncompressed audio as stored on a retail CD has a bit rate of 1,411.2 kbit/s (16 bit/sample × 44100 samples/second × 2 channels / 1000 bits/kilobit).Several bit rates are specified in the MPEG-1 Layer 3 standard: 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 192, 224, 256 and 320 kbit/s, and the available sampling frequencies are 32, 44.1 and 48 kHz. A sample rate of 44.1 kHz is almost always used, because this is also used for CD audio, the main source used for creating MP3 files. A greater variety of bit rates are used on the Internet. 128 kbit/s is the most common, because it typically offers adequate audio quality in a relatively small space. 192 kbit/s is often used by those who notice artifacts at lower bit rates. As the Internet bandwidth availability and hard drive sizes have increased, 128 kbit/s bit rate files are slowly being replaced with higher bit rates like 192 and 256 kbit/s, with some being encoded up to MP3's maximum of 320 kbit/s. It is unlikely that higher bit rates will be popular with any lossy audio codec because file sizes at higher bit rates approach those of lossless codecs such as FLAC.

By contrast, uncompressed audio as stored on a retail CD has a bit rate of 1,411.2 kbit/s (16 bit/sample × 44100 samples/second × 2 channels / 1000 bits/kilobit).

At 1,000 ft/sec a projectile would take 50 milliseconds to hit a target at 20 feet (1,000/20ft).

The worst MP3 spec would easily capture the 2 thuds and Audacity would easily delineate and you would be able measure the blip time differential.

Anyone measuring projectile velocity using an MP3 recorder would have little reason to be embarrased.

BoyntonStu

Posted: Sun Aug 16, 2009 2:39 pm
by jimmy101
Technician1002 wrote:You are embarrassing yourself. Read the original post again under specifications.
Not really. I just refuse to wade through twenty lines of irrelevant spec's to find the relevant part. :D Of course, you are still missing an important spec: the actual sample rate. Both MP3 and WAV (WMA) support different sample rates. The recording might be at 8KHz, which is more than adequate for a simple voice recording, but not as fast as you would like in a chrony.

I wouldn't worry too much about compression (in terms of reduction in amplitude, not in data compression which is a reduction in the frequency spectrum). Worse case scenario the recorder clips the top of the peak off, so you just use the center of the flat peak top as the top of the peak. More than accurate enough for a shooting chronometer given the uncertainty in all the other parameters. As long as it doesn't clip too much off. If you're really anal the peak top can be adequately reconstructed from the adjacent non-clipped data.
ramses wrote: I had considered this, but if the photogates were a reasonable distance apart, (like 3ft), the max velocity would be 300fps. (3/.01).
Not quite. The maximum recordable velocity depends on how accurate you want it to be. For "3/0.01" you would be able to tell the difference between 150 and 300 FPS but nothing in between or beyond 300 FPS. If you want say 10% accuracy at 300 FPS, and your timer is 0.01 second resolution, then the gates would need to be 30 feet apart. That is, you need at least ten data points between the two events.

For 1% resolution you need 100 data points between the two events. For 300 FPS with 3' gates you need a resolution of 0.0001 seconds (100uSec). Unfortunately that's about ten times better resolution than even high end stop watches made for timing human events.

But it's a peace of cake for a 48KHz audio system where the time resolution is 0.000021 seconds (21uSec).

Posted: Sun Aug 16, 2009 3:47 pm
by boyntonstu
jimmy101 wrote:
Technician1002 wrote:You are embarrassing yourself. Read the original post again under specifications.
Not really. I just refuse to wade through twenty lines of irrelevant spec's to find the relevant part. :D Of course, you are still missing an important spec: the actual sample rate. Both MP3 and WAV (WMA) support different sample rates. The recording might be at 8KHz, which is more than adequate for a simple voice recording, but not as fast as you would like in a chrony.

I wouldn't worry too much about compression (in terms of reduction in amplitude, not in data compression which is a reduction in the frequency spectrum). Worse case scenario the recorder clips the top of the peak off, so you just use the center of the flat peak top as the top of the peak. More than accurate enough for a shooting chronometer given the uncertainty in all the other parameters. As long as it doesn't clip too much off. If you're really anal the peak top can be adequately reconstructed from the adjacent non-clipped data.
ramses wrote: I had considered this, but if the photogates were a reasonable distance apart, (like 3ft), the max velocity would be 300fps. (3/.01).
Not quite. The maximum recordable velocity depends on how accurate you want it to be. For "3/0.01" you would be able to tell the difference between 150 and 300 FPS but nothing in between or beyond 300 FPS. If you want say 10% accuracy at 300 FPS, and your timer is 0.01 second resolution, then the gates would need to be 30 feet apart. That is, you need at least ten data points between the two events.

For 1% resolution you need 100 data points between the two events. For 300 FPS with 3' gates you need a resolution of 0.0001 seconds (100uSec). Unfortunately that's about ten times better resolution than even high end stop watches made for timing human events.

But it's a peace of cake for a 48KHz audio system where the time resolution is 0.000021 seconds (21uSec).
Audacitry places pips every .0010 seconds.

If you are off by .0010 seconds your error at 1,000 fps is minuscule.

As long as the bag/thump raises the amplitude high enough for identification, you are good to go.

A friend of mine commented:

That's a sweet looking watch for having a mp3 player built in! I was picturing something very different. That should work great for recording in the field, much better than dragging the laptop around. Come to think of it, it could come in very handy at work... You aren't kidding about Audacity, that has got to be one of the most valuable free programs out there. I've compared the results I've gotten with Audacity to the results of my chronograph, and there isn't very much difference in readings. Audacity will give slightly slower readings, due to the longer velocity average measured (10+ yards as opposed to 1 foot).

I believe that I will trust MP3 until I am a able to compare it to a uncompressed audio or an accurate chronometer.

3/0.1 was used as an example of how to do the calculation not as a suggestion for experiment.

I would never fire into a target that close out of fear thatf the target would fire back at me.

I believe that 20 feet is a good distance if you take precautions.

BoyntonStu

Posted: Sun Aug 16, 2009 8:45 pm
by rp181
Its quite easy to hook up the stopwatch function to a couple of photogates, for direct time measurment. Or simpler yet, A break wire chronograph with schimdt triggers and a 555 monostable.

Posted: Mon Aug 17, 2009 11:31 am
by jimmy101
boyntonstu wrote:Audacitry places pips every .0010 seconds.
Actually, I believe it puts a pip at every data point. At 44KHz that's a dot every 0.000023 seconds. If you are seeing pips at 1mS then that means the recording was done at 1KHz.
boyntonstu wrote:If you are off by .0010 seconds your error at 1,000 fps is minuscule.
Assuming the distance to the target is large. Even then 1000 FPS at 20' the total time is 20mS. With pips every 1mS the error is 5%. If the MP3/WAV recorder is actually sampling at say 44KHz then Audacity will give you all the data and the sample error is much less. But, the larger the distance the less accurately your measurement reflects the muzzle velocity. With a small dense projectile the error isn't to bad. For a lighter ammo (like a spud) the error over 20 feet is going to be more significant.
boyntonstu wrote:As long as the bag/thump raises the amplitude high enough for identification, you are good to go.
Yep. The amplitude, and to a large extent, the fidelity is not critical.
boyntonstu wrote:I believe that I will trust MP3 until I am a able to compare it to a uncompressed audio or an accurate chronometer.
Though I would take a quick look at exactly what the sample rate the watch is using. If it is meant just to do voice notes it may well be sampling at 1KHz. As you said, that's OK for large target distances. And, if the gun is firing at significantly less than 1000 FPS then the time resolution becomes even less critical.
boyntonstu wrote:3/0.1 was used as an example of how to do the calculation not as a suggestion for experiment.

I would never fire into a target that close out of fear thatf the target would fire back at me.

I believe that 20 feet is a good distance if you take precautions.
The target will only "fire back" if it is nearly perpendicular to the projectiles path. Tilt the target to 45 degrees and you'll have a very good idea where the shell is going. That lets you bring the target nearer the muzzle. If you have sufficiently high sample rate (like 44KHz or 48KHz) then you have enough resolution even at the short range. The short range gives a better estimate of the muzzle velocity.

Posted: Mon Aug 17, 2009 11:55 am
by boyntonstu
jimmy101 wrote:
boyntonstu wrote:Audacitry places pips every .0010 seconds.
Actually, I believe it puts a pip at every data point. At 44KHz that's a dot every 0.000023 seconds. If you are seeing pips at 1mS then that means the recording was done at 1KHz.
boyntonstu wrote:If you are off by .0010 seconds your error at 1,000 fps is minuscule.
Assuming the distance to the target is large. Even then 1000 FPS at 20' the total time is 20mS. With pips every 1mS the error is 5%. If the MP3/WAV recorder is actually sampling at say 44KHz then Audacity will give you all the data and the sample error is much less. But, the larger the distance the less accurately your measurement reflects the muzzle velocity. With a small dense projectile the error isn't to bad. For a lighter ammo (like a spud) the error over 20 feet is going to be more significant.
boyntonstu wrote:As long as the bag/thump raises the amplitude high enough for identification, you are good to go.
Yep. The amplitude, and to a large extent, the fidelity is not critical.
boyntonstu wrote:I believe that I will trust MP3 until I am a able to compare it to a uncompressed audio or an accurate chronometer.
Though I would take a quick look at exactly what the sample rate the watch is using. If it is meant just to do voice notes it may well be sampling at 1KHz. As you said, that's OK for large target distances. And, if the gun is firing at significantly less than 1000 FPS then the time resolution becomes even less critical.
boyntonstu wrote:3/0.1 was used as an example of how to do the calculation not as a suggestion for experiment.

I would never fire into a target that close out of fear thatf the target would fire back at me.

I believe that 20 feet is a good distance if you take precautions.
The target will only "fire back" if it is nearly perpendicular to the projectiles path. Tilt the target to 45 degrees and you'll have a very good idea where the shell is going. That lets you bring the target nearer the muzzle. If you have sufficiently high sample rate (like 44KHz or 48KHz) then you have enough resolution even at the short range. The short range gives a better estimate of the muzzle velocity.
A good explanation. Thanks.

Since I wear my chronograph wherever I go, it is always available for recording voice and for measuring projectile velocities.

It even tells time!

BoyntonStu