Home brewed mic for Iphone Timegrapher app?

[bsoderling]

Hi all,

 

I’m a fan of the timegrapher app for Iphone and have been using it for a long time, with reasonable result for my needs.

 

My assumption is that the weak spot is the mic where you are using the built-in mic in the head set that comes with the phone.

 

My question is if there’s any experience out there from building/adapting an alternative mic to plug into the head set input on the phone.

 

I have seen many posts where people have designed pc s/w and built mic’s that plug into your pc and I would guess one could do the same with the Iphone.

 

Any feedback is appreciated.

 

 

 

 

Sent from my iPad using Tapatalk

[MyRetroWatches]

It’s funny your should post this now. I have just uploaded my latest video on my channel and it’s about the iPhone app vs iOS app.

App came off interestingly should we say.

Due to some of the feedback and my own curiosity I am going to do one short video test using my lapel mic that I would use for my voice. Also to remove the caseback.

I will test first on the timegrapher and then on the app and then compare.
Here is my video

Sent from my iPhone using Tapatalk

[bsoderling]

Well...strange things happen all the time.

Truly inspiring and I must agree with your fundamental conclusion, if you want something that works simply and flawlessly out of the box, the Weishi is definitely not over priced.

But still...there’s something about exploring the underlying reasons and possibly try to improve what you have. Kind of the same sentiment that makes watch work interesting...

Some quick feedback from my experiences that might be interesting for others.

- I find it consistently easier to get good readings with an open case and the mic directly over the escapement. Again I guess that has to do with limitations of the mic (and possibly the signal conditioning that follows).

- I usually step back on the sensitivity to the point where I get no readings (or zero sensitivity) and then step up to where I get the most stable plot and reading. Sometimes this is significantly lower than the default 12 setting.

- I tend to always use the noise cancellation option in the settings menu. Especially when it’s hard to get an amplitude reading this allows to increase the sensitivity without getting rubbish data.

- It’s difficult to get completely clear lines, there will always be some outliers. To me this indicate that the s/w that detects the relevant points in the signal isn’t smart enough to sort these out as erroneous. Alternatively it’s the signal conditioning/filtering that could be better.

- Fundamentally I wouldn’t trust the average readings produced when the plot looks like as with your 2nd watch. Garbage in => garbage out and averaging shouldn’t make it more reliable (unless there is actually more going on that we don’t see that removes outlier data)

To conclude, it would be great with a follow up where you try to tweak the performance a bit. But I realize I’m probably a bit too nerdy here... :-)

And I still wouldn’t be surprised if a better mic and some analogue filtering could do magic with the performance of the app.

I know there are posts on that subject on the forum and I’ll see what I can dig out there.

And you got yourself a new follower on youtube... :-)










Sent from my iPad using Tapatalk

[praezis]

I suppose the main issue is the app, not so much the mic.

I posted about unsatisfacting results with different Iphone apps here in the forum (maybe post was deleted by a kind moderator). I used special piezo mics as well  as direct signal input with different volumes-

Frank

[MyRetroWatches]

Well...strange things happen all the time.

Truly inspiring and I must agree with your fundamental conclusion, if you want something that works simply and flawlessly out of the box, the Weishi is definitely not over priced.

But still...there’s something about exploring the underlying reasons and possibly try to improve what you have. Kind of the same sentiment that makes watch work interesting...

Some quick feedback from my experiences that might be interesting for others.

- I find it consistently easier to get good readings with an open case and the mic directly over the escapement. Again I guess that has to do with limitations of the mic (and possibly the signal conditioning that follows).

- I usually step back on the sensitivity to the point where I get no readings (or zero sensitivity) and then step up to where I get the most stable plot and reading. Sometimes this is significantly lower than the default 12 setting.

- I tend to always use the noise cancellation option in the settings menu. Especially when it’s hard to get an amplitude reading this allows to increase the sensitivity without getting rubbish data.

- It’s difficult to get completely clear lines, there will always be some outliers. To me this indicate that the s/w that detects the relevant points in the signal isn’t smart enough to sort these out as erroneous. Alternatively it’s the signal conditioning/filtering that could be better.

- Fundamentally I wouldn’t trust the average readings produced when the plot looks like as with your 2nd watch. Garbage in => garbage out and averaging shouldn’t make it more reliable (unless there is actually more going on that we don’t see that removes outlier data)

To conclude, it would be great with a follow up where you try to tweak the performance a bit. But I realize I’m probably a bit too nerdy here... :-)

And I still wouldn’t be surprised if a better mic and some analogue filtering could do magic with the performance of the app.

I know there are posts on that subject on the forum and I’ll see what I can dig out there.

And you got yourself a new follower on youtube... :-)










Sent from my iPad using Tapatalk

Thanks , detailed reply!
All makes sense.
Don’t worry there will be a short follow on to the first video . Probably within a few days. I will try to see how I can better the reading after getting a datum reading off the TG .



Sent from my iPhone using Tapatalk

[bsoderling]

Hi,

Saw your new video and it’s pretty much in line with my experience. On some watches it works better on some not so good.

Interesting alternative mic trial. The mic does have an influence but it will be a tough task to find one that really makes a change in the positive direction without lots of trial and (mostly) error.

Looking forward to coming videos!




Sent from my iPad using Tapatalk

[AndrewR]

I had a good exchange with the developer of the app I've been using. He said that if you're not getting the volume input at 4-5 out of 5 then the app/iPhone doesn't have enough resolution for calculating the amplitude (for example). I guess low volume might impair the other functions, but not the same extent. 

Therefore I was interested to see what external mic options might help this as well. Might be you can't beat the ultra hi-tech internal mic anyway?