I'm guessing you vibe-coded this and let the model hallucinate the decibel thresholds? I randomly googled a couple (Philly and Fort Worth) and the limits shown in the UI don't match up with any law I could find. Philly doesn't have a specific decibel threshold (the law is based on decibels above standard background noise) and Fort Worth's is 70 / 60, not 60 / 55, with higher limits in non-residential areas.
So it's safe to assume that most of the other limits hardcoded in the tool are wrong as well. Pretty irresponsible to release this without even taking the few minutes to research the laws yourself.
A few more concerning claims:
- "US One-Party Consent": This is not a thing, it varies by jurisdiction. Many states are two-party consent.
- "Standards database updated quarterly from public records": It would not appear that way!
- "Not Legal Advice": You are giving legal advice, and it's incorrect.
IANAL but I don't think you can invoke the two party consent laws regarding noises loud enough to be disruptive to other people through walls and such. The recording consent laws typically apply to things like phone calls I believe.
If I may give constructing feedback then, "noise evidence generator" sounds like you're creating fake noise proof to fool courts. The name needs to be revised.
"recorder" might feel less like it's inventing things.
More justification for my mortgage for my well-insulated, single family home with air and distance from my neighbors. Me and my subwoofers escape litigation facilitated by this SaaS.
I've searched but failed to find anything that works as a standard fiducial-type sound. e.g. With a recording of the sound produced by the fiducial from a set distance (and knowing temperature, pressure, humidity), you could calibrate/adjust recordings.
That is the crux you always need a calibrated sound to calibrate the mic. Best I think this app could hope to do is establish a relative reading. perhaps it could play a set of repeatable tones on the device(assuming being executed from a device that does not have external speakers that can be adjusted/ validate that sound was played from phone speaker and not headphone jack to a dac?) and record those then could replay sounds later to a calibrated device to establish reference. or start recording with some other regular repeatable sound like a vacuum cleaner with video showing how it was positioned for recording.
Congratulations on building a nice-looking app that addresses a legitimate need.
I'd like to see some proof that this is able to accurately measure noise level across a range of devices. The CDC have a sound meter app [1] which has been tested to 2db accuracy, and they only make that available on specific Apple devices because calculating noise level depends on the hardware.
I'm sorry to ask, but I'm seeing many cases of AI apps making accuracy claims based on the author’s ‘reasonableness spot checks’ but with no statistical testing that the outputs are accurate.
Cool idea but from personal experience it's probably not worth much. I'm lucky enough to live on the top floor apartment and I've had multiple issues with multiple neighbors over the years. Maybe the leasing office where I live is worse than others, but they basically just ignore complaints.
I came home from a holiday a few years ago and woke up to a dog howling at 7:00am. This was new. This new dog (a husky, lol) would completely lose it's mind any time the owner, my downstairs neighbor, would leave it alone for any amount of time. Like the moment he shut the door. I work from home so it was difficult to focus while this dog would scream it's lungs out for hours at a time during the day. I complained to the office and left a note on the neighbor's door and got no response from either. After a few weeks I decided to handle it, stayed up to 3:00am, walked into my bedroom (directly above his bedroom) and jumped up and down as hard as I could for as long as I could. I bruised my heel on the floor but in the morning there was no barking and it stopped completely, which means it could have ended any time, they just didn't care. I have several other stories that all ended more or less the same way. No one gives a fuck until they are personally affected. You must fight fire with fire.
The NIOSH app on my iPhone matches the calibrated meter I have so it’s feasible. However the app has the appearance of vibecoding. The proof of the pudding is in the eating so that doesn’t matter if it works. Have you tried using it?
I'm guessing you vibe-coded this and let the model hallucinate the decibel thresholds? I randomly googled a couple (Philly and Fort Worth) and the limits shown in the UI don't match up with any law I could find. Philly doesn't have a specific decibel threshold (the law is based on decibels above standard background noise) and Fort Worth's is 70 / 60, not 60 / 55, with higher limits in non-residential areas.
So it's safe to assume that most of the other limits hardcoded in the tool are wrong as well. Pretty irresponsible to release this without even taking the few minutes to research the laws yourself.
A few more concerning claims:
- "US One-Party Consent": This is not a thing, it varies by jurisdiction. Many states are two-party consent.
- "Standards database updated quarterly from public records": It would not appear that way!
- "Not Legal Advice": You are giving legal advice, and it's incorrect.
> let the model hallucinate the decibel thresholds?
Does it even matter ? Unless you're using calibrated hardware the measured dB will be "hallucinated" too
Yeah, there's that too. But it's especially bad to mislead people about what their local laws say!
Thank you for your feedback. I will verify it immediately and make adjustments accordingly.
IANAL but I don't think you can invoke the two party consent laws regarding noises loud enough to be disruptive to other people through walls and such. The recording consent laws typically apply to things like phone calls I believe.
所有分贝阈值是找的当地公开使用的官方标准来使用的
If I may give constructing feedback then, "noise evidence generator" sounds like you're creating fake noise proof to fool courts. The name needs to be revised.
"recorder" might feel less like it's inventing things.
Adjustments have been made, thank you again! That's a great suggestion!
Thank you for your suggestion, I will revise it!
"Noise Evidence Logger" perhaps? the 'generator' in the name also made me think this was for faking proof. neat app idea
was my first thought as well.
I 100% thought it was a tool for landlords to do illegal evictions lol
More justification for my mortgage for my well-insulated, single family home with air and distance from my neighbors. Me and my subwoofers escape litigation facilitated by this SaaS.
How do we know microphones on different devices are calibrated to record the same power readings? Do you have court-ready documents to show this?
I've added a calibration feature, hoping it will improve and help.
I've searched but failed to find anything that works as a standard fiducial-type sound. e.g. With a recording of the sound produced by the fiducial from a set distance (and knowing temperature, pressure, humidity), you could calibrate/adjust recordings.
That is the crux you always need a calibrated sound to calibrate the mic. Best I think this app could hope to do is establish a relative reading. perhaps it could play a set of repeatable tones on the device(assuming being executed from a device that does not have external speakers that can be adjusted/ validate that sound was played from phone speaker and not headphone jack to a dac?) and record those then could replay sounds later to a calibrated device to establish reference. or start recording with some other regular repeatable sound like a vacuum cleaner with video showing how it was positioned for recording.
I don't think you can really do that reliably either. Arbitrary consumer device microphones do not have flat frequency response
Congratulations on building a nice-looking app that addresses a legitimate need.
I'd like to see some proof that this is able to accurately measure noise level across a range of devices. The CDC have a sound meter app [1] which has been tested to 2db accuracy, and they only make that available on specific Apple devices because calculating noise level depends on the hardware.
I'm sorry to ask, but I'm seeing many cases of AI apps making accuracy claims based on the author’s ‘reasonableness spot checks’ but with no statistical testing that the outputs are accurate.
[1] https://www.cdc.gov/niosh/noise/about/app.html
Thank you for your feedback. I will adjust the accuracy settings immediately.
In my own reasonableness spot check, what measures as 37dB on my phone is showing as 49dB on my laptop
A calibration function has been added; hopefully this will help.
Cool idea but from personal experience it's probably not worth much. I'm lucky enough to live on the top floor apartment and I've had multiple issues with multiple neighbors over the years. Maybe the leasing office where I live is worse than others, but they basically just ignore complaints.
I came home from a holiday a few years ago and woke up to a dog howling at 7:00am. This was new. This new dog (a husky, lol) would completely lose it's mind any time the owner, my downstairs neighbor, would leave it alone for any amount of time. Like the moment he shut the door. I work from home so it was difficult to focus while this dog would scream it's lungs out for hours at a time during the day. I complained to the office and left a note on the neighbor's door and got no response from either. After a few weeks I decided to handle it, stayed up to 3:00am, walked into my bedroom (directly above his bedroom) and jumped up and down as hard as I could for as long as I could. I bruised my heel on the floor but in the morning there was no barking and it stopped completely, which means it could have ended any time, they just didn't care. I have several other stories that all ended more or less the same way. No one gives a fuck until they are personally affected. You must fight fire with fire.
How do you calibrate?
I mean, the sensitivity of the microphones varies a lot.
Legal proof needs a metrological chain of trust.
Understood, thank you for your feedback.
The NIOSH app on my iPhone matches the calibrated meter I have so it’s feasible. However the app has the appearance of vibecoding. The proof of the pudding is in the eating so that doesn’t matter if it works. Have you tried using it?
A calibration function has been added; hopefully it will be helpful.