Let's get honest with SPL

[Blumlein 88]

My app shows cca 10db less than GLM mic. Maybe phone does not measure bass?

Few phones respond well below 200 hz.

 

[kongwee]

Most likely your readings are no good. With a test tone you get standing waves. Change frequency and you are in a different pattern of standing waves.

Isn't that the real world. Even you record at fixed frequency for 20 second, your waveforms is still not the same at every cycle.

 

[hege]

that's because most microphones on a phone are C-weighted by default. I experienced similarily weird calibration stuff when i wasn't using relatively high frequencies.

Few phones respond well below 200 hz.

Not sure that even the microphone frequency response matters that much, as long as there aren't any major peaks. Atleast A-weighting should be measured well.

The question is, how is any phone app supposed to know how sensitive a random phone models microphone is? It's no different from any random microphone you might buy. Unless you have calibration data for it, or calibrate it yourself to a known volume level, the results are meaningless. UMIK-1 comes with the needed calibration data, and it's USB so you can't mess things up with some pre-amplifier volume control etc.

Unless the software specifically guarantees some results for your phone model, like Decibel X ("Tested and calibrated for all Apple devices"), it's a crapshoot. Decibel X allows to calibrate yourself (easy to do with UMIK-1 beside). It can even correct for frequency response besides sensitivity, but I was too lazy for that.

 

[fpitas]

Not sure that even the microphone frequency response matters that much, as long as there aren't any major peaks. Atleast A-weighting should be measured well.

The question is, how is any phone app supposed to know how sensitive a random phone models microphone is? It's no different from any random microphone you might buy. Unless you have calibration data for it, or calibrate it yourself to a known volume level, the results are meaningless. UMIK-1 comes with the needed calibration data, and it's USB so you can't mess things up with some pre-amplifier volume control etc.

Unless the software specifically guarantees some results for your phone model, like Decibel X ("Tested and calibrated for all Apple devices"), it's a crapshoot. Decibel X allows to calibrate yourself (easy to do with UMIK-1 beside). It can even correct for frequency response besides sensitivity, but I was too lazy for that.

Besides a cal, some method to properly hold the phone is a good idea, if you expect reproducible results from test to test.

 

[sajgre]

Few phones respond well below 200 hz.

I believe so too, so this thread is only useful if people with calibrated mics chime in.

 

[Blumlein 88]

Not sure that even the microphone frequency response matters that much, as long as there aren't any major peaks. Atleast A-weighting should be measured well.

The question is, how is any phone app supposed to know how sensitive a random phone models microphone is? It's no different from any random microphone you might buy. Unless you have calibration data for it, or calibrate it yourself to a known volume level, the results are meaningless. UMIK-1 comes with the needed calibration data, and it's USB so you can't mess things up with some pre-amplifier volume control etc.

Unless the software specifically guarantees some results for your phone model, like Decibel X ("Tested and calibrated for all Apple devices"), it's a crapshoot. Decibel X allows to calibrate yourself (easy to do with UMIK-1 beside). It can even correct for frequency response besides sensitivity, but I was too lazy for that.

Some phone apps do detect the model phone and are more or less calibrated. Apple isn't the only one for which that is true.

EDIT to add: Decibel X is on Android too and is pre-calibrated to phones. Many people report it drains their battery and have to force stop it after using it.

 

[Blumlein 88]

Isn't that the real world. Even you record at fixed frequency for 20 second, your waveforms is still not the same at every cycle.

It isn't a matter of same cycle. Your speaker could be putting out exactly the same sound level at two different frequencies. Yet you are in a peak for one frequency with your microphone and a valley for another. So you would not expect a sound level meter to report the same level because they physically are different at your location in the room.

 

[Prana Ferox]

I live in an apartment and am roughly never as loud as I want to be.

One thing I haven't seen mentioned, people can have a volume presence for music, likely louder for 'critical listening', but they're also likely to have a preferred volume for TV and especially movies, and that's likely to be a lot higher.

 

[hege]

Some phone apps do detect the model phone and are more or less calibrated. Apple isn't the only one for which that is true.

I doubt many people even realize this is a thing, they just download some random app (there are dozens) and happily post the numbers. Maybe someone will stumble on this and other threads and is educated, but this will end up in the archives soon too.

 

[bo_knows]

96dB anechoic @1M and 96dB in-room at LP (Let say 3M for instance) are indeed two different things...

But AFAIC, my regular SPL for most critical listening is about 80dB+ at LP. Could be 10dB lower for chilling out, and 10dB over for loud, physical listening.

Same here, C or Z weighting via iPhone using the Decibel X app.

 

[MAB]

More interesting if you are at 96dB SPL listening position, and you turn -0.5dB at your source, what is the result of dB SPL at your listening position.

I would turn it down by more than 0.5dB, just to save hearing.
Joking aside, can you show this non-linearity with data? You make such odd observations, it would be good to see you back them up with some data.

 

[Trdat]

There's similar topics already, you'll get just the same vague answers with mobile apps, no defined weighting etc.

"96dB" alone means nothing. Is the value A-weighted? C/Z-weighted? Average? Peak? All this also depends on how bassy music you listen, if you apply some house curve, amount of acoustic treatment etc.

90-95dB(A) is the max I physically like to listen when rocking out. It's loud, but I have an extremely treated/damped room which doesn't attack your ears like an echoing untreated room would. Now depending on the bass, C/Z-weighting for the same can display up to 110dB(Z). Peak value can be as high as 120dB.

Once in a while I will play at 90 to 95 db C weighted(proper SPL meter) for 10-15 min. But during my once a week sessions I still go loud but maintain that between 80 to 90db for around 30min and keep it at less than 80 db for the next hour so.

I also have a very damped room, between .2 and .5 for RT60. Ears rarely hurt, or ring after, in saying that I always hold my hands over my ears in the metro never measured the SPL in the metro system but its painfully loud.

 

[AdamG]

If you measure, what is the SPL at your listening position?

This has often been mentioned here and there, but I'd like to compile it and under the same criteria for everyone; only if you have an SPL meter and you measured at your listening position.

I have to compare your answers, it's strange to me that someone would reach 96dB... I read what Amir wrote about Metas not being happiest little speaker at 96dB, but even approaching 90dB and the sound is overwhelming. Does anyone really ever listen at 96 or over? Even if it's just a few songs...

If I might make a suggestion? Would you consider changing this thread into a Poll? Set a few ground rules (Like normal casual listening level and not Movie or Party music levels) and create vote ranges in dbs groupings. Example: 50dB - 55db; 56dB-60dB; 61dB-65dB; and go up to maybe 100dB? This way will make the results available to everyone and easy to understand as the Poll results will be self evident. Just a suggestion. Might have to start a new thread? If you do we can close this thread and leave a redirect link to the new one.

 

[Goldenyears]

Average of 56.76. Hi 78db. ipad app.

 

[Blumlein 88]

Screenshot of REW. Red is peak hold. 64 K FFT. This covered about the middle third of the song. Comfortable listening level. I might at times turn it up 5 or 6 db more. The song is Flight of the Cosmic Hippo by Bela Fleck and the Flecktones. Has more than the normal amount of low frequency bits in it.

You can see the Z wtd level is nearly 87 db, while A wtd was merely 72 db. Plus there were peaks of 90 db in the song.

Now here is Bad Religion by Godsmack. This is about as loud as I would listen to it. Same volume setting. You see lower peak levels yet higher average levels and A-wtd seems to correspond more or less with how loud it was subjectively. Much louder than the previous song. I think B-wtd correlates best with subjective impressions vs SPL meters, but I didn't have that as a choice.

Here is Bird on a Wire from Jennifer Warnes Famous Blue Raincoat. I'd listen to this maybe 3-5 db louder at times, but this was a comfortable enjoyable level. So I'd say from this sampling I mostly listen at average A-wtd levels of 75-80 db and consider it is getting too loud if you pass 85 db levels measured this way.

 

[sarumbear]

For music I have the loudness of my system normalized to -16 LUFS and I can comfortably listen to my system at -15dBFS max volume (peaking at 90dBSPL for a song with -16 LUFS loudness) for 30 to 40 minutes at a time.

I think it's also dependent on room acoustics and material, if the sound is just not good at 75dBSPL turning the volume up will just increase your uncomfort.

What has LUFS to do the with SPL? LUFS or dBFS is not even remotely related to SPL. They are different units.

 

[sarumbear]

that's because most microphones on a phone are C-weighted by default. I experienced similarily weird calibration stuff when i wasn't using relatively high frequencies.

C weighting only affects the extreme octaves. For all practical purposes (like music monitoring) it is flat.

Frequency Weightings - A-Weighted, C-Weighted or Z-Weighted

A description of the frequency weightings used for sound level meters - A-Weighted, C-Weighted or Z-Weighted.

www.noisemeters.co.uk

 

[hege]

Relaxing with some late nite house mixes, 72dB(A), 88dB(Z), 100dB peaks. This is pretty much my comfy relaxing level, could do it all nite safely without fatigue. It's just nice to know there's still 20dB+ clean headroom when I feel rowdy. 10dB increase from here is already a quite big jump into feeling more physical.

 

[digitalfrost]

I have a calibrated SPL meter. My system is setup for 78dBC as standard - that is the point where loudness correction is flat and my room curve is as it's supposed to be. I usually do not reach this when casually listening, but I do when actively listening and I might also go to 83dBC or 86dBC when I'm feeling it. Louder than that is uncomfortable for me.

The max volume I have measured with this system is 107dBC during the explosion at the beginning of Edge of Tomorrow (when their craft gets shot down).

My room volume is ~56m³ which is ~2000 cu. ft. There's a table here for recommended reference SPL levels https://www.atsc.org/wp-content/upl...tablishing-and-maintaining-audio-loudness.pdf (see Table 10.2 Reference Sound Pressure Level).

There's also other standards that give slightly less SPL but more or less this is in the correct ballpark https://www.masteringthemix.com/blogs/learn/the-perfect-monitoring-levels-for-your-home-studio

e: To be more precise, I use ReplayGain for my music which will typically result in music that is 14dBFS below full scale. I achieve 78dBC with ReplayGain applied at a volume setting of 33%, that is -9.6dB below full volume. So the theoretical max based on this is 14+10+78 = 102dB. If you add in stereo speakers, stereo subwoofers and room gain, that is how I achieved the 107dB.

 

[DonH56]

About 60 dB average on a C-weighted meter, 70 dB when playing loud or watching a loud movie. Music 70 to 80 dB if I really like the song and push the volume way high, again average C-weighted readings, so deep bass is probably at least 10 to 20 dB higher.