|
Post by patate91 on Aug 21, 2020 16:29:19 GMT
I was reading about amplifier's power needed do drive speakers and there's a couple of things I've read that I find suspicious.
First a lot of people says they listen at 85 db. Personnaly I listen at 75 db, C weighted, when I want it loud. Usually it will as low as 65 db.
Second about power for peaks. There's people who says that 20 db + peaks needs to be planned when looking at amplifier's power.
My questions are : is there a lot of songs that have 20db peaks above the song's average SPL. What is done in the audio industry about that?
|
|
|
Post by Michael Lawrence on Aug 21, 2020 16:51:12 GMT
Hi Patate91 - You're asking about a concept called "crest factor" which is the ratio between the RMS level of a signal and its peaks. Typical music has a crest factor between 10 and 20 dB, depending on the genre and the style. Because different SPL metrics have different integration times, they are affected differently by the crest factor of a signal. Some metrics react more to the peaks, which others are closer to the rms level. Here is Peak C, SPL C Slow, SPL C Fast, and LCeq10 all for the same song. They are all C weighted metrics but they give very different answers, due to crest factor. That's why it's important always to indicate both the frequency weighting (which you did) and the time-domain information of an SPL measurement for it to be technically meaningful. With regards to your second question, a peak is, by definition, a very short-term event, and since power is energy over time, peaks dissipate very little power. Current sourcing abilities of an amplifier are therefore typically informed by the RMS voltage it will need to sustain through the load, and the load's impedance, and then making sure the voltage rails of the amplifier can accommodate the peaks with the desired crest factor. In a live sound reinforcement situation, where headroom is extremely expensive, the mixes tend to be a bit more limited and compressed, and the occasional peak limiting (6 dB or so) on transients is not considered to be extremely detrimental. For lots more on this, check out some of the work done by SynAudCon and Pat Brown with the CAF Viewer program.
|
|
|
Post by patate91 on Aug 21, 2020 17:20:04 GMT
Hi Patate91 - You're asking about a concept called "crest factor" which is the ratio between the RMS level of a signal and its peaks. Typical music has a crest factor between 10 and 20 dB, depending on the genre and the style. Because different SPL metrics have different integration times, they are affected differently by the crest factor of a signal. Some metrics react more to the peaks, which others are closer to the rms level. Here is Peak C, SPL C Slow, SPL C Fast, and LCeq10 all for the same song. They are all C weighted metrics but they give very different answers, due to crest factor. View AttachmentThat's why it's important always to indicate both the frequency weighting (which you did) and the time-domain information of an SPL measurement for it to be technically meaningful. With regards to your second question, a peak is, by definition, a very short-term event, and since power is energy over time, peaks dissipate very little power. Current sourcing abilities of an amplifier are therefore typically informed by the RMS voltage it will need to sustain through the load, and the load's impedance, and then making sure the voltage rails of the amplifier can accommodate the peaks with the desired crest factor. In a live sound reinforcement situation, where headroom is extremely expensive, the mixes tend to be a bit more limited and compressed, and the occasional peak limiting (6 dB or so) on transients is not considered to be extremely detrimental. For lots more on this, check out some of the work done by SynAudCon and Pat Brown with the CAF Viewer program. Thanks a lot! Again a lot of new things to read and learn. With crest factor I now have a good direction.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 21, 2020 19:02:10 GMT
"First a lot of people says they listen at 85 db. Personnaly I listen at 75 db, C weighted, when I want it loud. Usually it will as low as 65 db."
Many people definitely do, especially on studio environments. Many people listen to quite loud in studio enviroments (100dBc+).
The mix or even more the mastering(loudness war) have a lot of do(At least for me) if I can listen to something quite loud and enjoy it. If the master is dynamic and not limited to death to be loud, I can easily listen to at around 95dBc(small room) and really feel the music and enjoy it.
|
|
|
Post by patate91 on Aug 21, 2020 21:22:53 GMT
"First a lot of people says they listen at 85 db. Personnaly I listen at 75 db, C weighted, when I want it loud. Usually it will as low as 65 db." Many people definitely do, especially on studio environments. Many people listen to quite loud in studio enviroments (100dBc+). The mix or even more the mastering(loudness war) have a lot of do(At least for me) if I can listen to something quite loud and enjoy it. If the master is dynamic and not limited to death to be loud, I can easily listen to at around 95dBc(small room) and really feel the music and enjoy it. What puzzle me about those SPL 90 + is that they become harmful pretty quickly. Long rest at lie spl is needed to désert time to 0. dangerousdecibels.org/education/information-center/decibel-exposure-time-guidelines/
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 22, 2020 8:25:47 GMT
"First a lot of people says they listen at 85 db. Personnaly I listen at 75 db, C weighted, when I want it loud. Usually it will as low as 65 db." Many people definitely do, especially on studio environments. Many people listen to quite loud in studio enviroments (100dBc+). The mix or even more the mastering(loudness war) have a lot of do(At least for me) if I can listen to something quite loud and enjoy it. If the master is dynamic and not limited to death to be loud, I can easily listen to at around 95dBc(small room) and really feel the music and enjoy it. What puzzle me about those SPL 90 + is that they become harmful pretty quickly. Long rest at lie spl is needed to désert time to 0. dangerousdecibels.org/education/information-center/decibel-exposure-time-guidelines/Yep. The modern world is anyways, everywhere too noisy for human ears. I live in a really small city and the noise pollution even here is way too much. Must be really a nightmare in big cities...
|
|
|
Post by patate91 on Aug 22, 2020 12:50:28 GMT
For lots more on this, check out some of the work done by SynAudCon and Pat Brown with the CAF Viewer program. I did not pay for SynAudCon course but I found this article from Pat Brown www.prosoundweb.com/understanding-the-nuances-of-crest-factor/Can I use Audacity to get crest factor from my library? I'm doing this for fun and learn so I don't need to be extremely precise. It appears that I need the Leq ( C weighting to get the crest factor, is it correct?
|
|
|
Post by Michael Lawrence on Aug 22, 2020 13:39:21 GMT
You can use MeldaProduction's free MLoudnessAnalyzer plugin to meter crest factor, hosted in the DAW of your choice. I use Reaper and Logic Pro. www.meldaproduction.com/MLoudnessAnalyzerCrest factor should use the full spectrum, so unweighted (Z weighting). For typical broadband content, Z and C give answers within about a dB. Sound exposure, regardless of whether you're using OSHA, NIOSH or WHO guidance, is always A weighted. OSHA is what is legally permissible in the US, whereas NIOSH is what is actually safe. Smaart SPL can calculate both OSHA and NIOSH exposure directly in percent dose: Which takes out a lot of the guesswork and futzing around with TWA tables. When the meter hits 100%, you've been exposed to your dose for the day. In my industry (live sound) we are seeing more and more regulations that specify an audience exposure limit as a % dose rather than as a more typical LAeq 15 or so.
|
|
|
Post by patate91 on Aug 22, 2020 13:50:02 GMT
You can use MeldaProduction's free MLoudnessAnalyzer plugin to meter crest factor, hosted in the DAW of your choice. I use Reaper and Logic Pro. www.meldaproduction.com/MLoudnessAnalyzerCrest factor should use the full spectrum, so unweighted (Z weighting). For typical broadband content, Z and C give answers within about a dB. Sound exposure, regardless of whether you're using OSHA, NIOSH or WHO guidance, is always A weighted. OSHA is what is legally permissible in the US, whereas NIOSH is what is actually safe. Smaart SPL can calculate both OSHA and NIOSH exposure directly in percent dose: View AttachmentWhich takes out a lot of the guesswork and futzing around with TWA tables. When the meter hits 100%, you've been exposed to your dose for the day. In my industry (live sound) we are seeing more and more regulations that specify an audience exposure limit as a % dose rather than as a more typical LAeq 15 or so. Again : thanks a lot!
|
|
|
Post by patate91 on Aug 22, 2020 16:16:03 GMT
Sound exposure, regardless of whether you're using OSHA, NIOSH or WHO guidance, is always A weighted. OSHA is what is legally permissible in the US, whereas NIOSH is what is actually safe. Smaart SPL can calculate both OSHA and NIOSH exposure directly in percent dose: View AttachmentWhich takes out a lot of the guesswork and futzing around with TWA tables. When the meter hits 100%, you've been exposed to your dose for the day. In my industry (live sound) we are seeing more and more regulations that specify an audience exposure limit as a % dose rather than as a more typical LAeq 15 or so. Just for fun how long it takes for a professionnal like you to hit 100%? Do you also calculate your exposure when you are not working? In the past I worked in the aluminum industry and I had to wear earing protections, but I was exposed to high db anyway. I was taking care to not expose me during my hobby time.
|
|
|
Post by Michael Lawrence on Aug 22, 2020 16:26:05 GMT
I try to absolutely minimize my exposure, both on and off the clock. I always wear hearing protection when I'm operating a lawn mower or a vacuum or any similarly loud device. I also wear ear protection at any live events where I'm not mixing or system tech. Loud concerts are, in general, a problem for the industry, and a typical live event will give the audience far more than 100% exposure dose. I actually teach a class targeted at live sound engineers that teaches them techniques for making their mix sound loud and impactful but at a safe SPL. Those techniques allow me to run my mixes a little lower than a typical concert, and leave the audience with less exposure dose. I have an article about it here if you'd like to read more: www.prosoundweb.com/how-loud-how-long-observations-questions-about-sound-exposure-issues/
|
|
|
Post by Hexspa on Aug 25, 2020 4:52:44 GMT
Just make sure to take your earplugs out onstage when they call out the key of the song. Only F and A don't rhyme so 2/3rds of the time, you're in the wrong key.
|
|
|
Post by Michael Lawrence on Aug 26, 2020 14:08:28 GMT
Just make sure to take your earplugs out onstage when they call out the key of the song. Only F and A don't rhyme so 2/3rds of the time, you're in the wrong key. That sounds like something spoken from experience
|
|
|
Post by Hexspa on Aug 27, 2020 1:55:18 GMT
Ya, not the first time I embarassed myself onstage. I want to echo that 85dB is too loud for most small rooms. This is corroborated by this SoS article:
However, although this 83dB SPL reference level (with 103dB peaks) is perfectly acceptable when listening in a big space, like a cinema or a film dubbing theatre, or even a very large and well‑treated commercial studio control room, it will be completely overwhelming in a smaller space, because the listener is inevitably sitting much closer to both the speakers and the room boundaries. The very different nature of early reflections in these conditions makes the level seem, psychoacoustically, much higher than it would be in a larger room.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 27, 2020 14:52:01 GMT
Ya, not the first time I embarassed myself onstage. I want to echo that 85dB is too loud for most small rooms. This is corroborated by this SoS article:
However, although this 83dB SPL reference level (with 103dB peaks) is perfectly acceptable when listening in a big space, like a cinema or a film dubbing theatre, or even a very large and well‑treated commercial studio control room, it will be completely overwhelming in a smaller space, because the listener is inevitably sitting much closer to both the speakers and the room boundaries. The very different nature of early reflections in these conditions makes the level seem, psychoacoustically, much higher than it would be in a larger room.
If its "NE type" room, this isn't true even in smaller rooms.
|
|