- Joined
- Feb 23, 2016
- Messages
- 20,769
- Likes
- 37,634
Can you record a performance, play it back in your home, and get 120 db of dynamic range in the playback?
Let me start at the beginning, where the recorded performance takes place. I'll ignore electronic soundscape creations where obviously the answer is yes.
I'll steal this from Amir's postings. Who stole it from Dolby I believe.
We see that our hearing thresholds actually are not too far from noise levels in some concert halls like the Davies Hall in the graph. So those can be quiet enough or nearly so that we'd hear silence from the venue itself. At our most sensitive mid-band we get down to 0 db SPL and a little below. The noise is a little above, but we can hear into it some anyway. So let us say this is close enough to say yes. If some orchestral or other recordings can peak at 120 db SPL we could hear down to 0 db SPL in quiet portions. Though not while music is playing.
Next step is the microphones. Can they be quiet enough? They have self noise. There are some that get close to 0 db SPL equivalent self noise (4-7 db). Some that reach that with A-weighting. So on the low end yes with A-weighting. Or very close to it as far as what we could hear. I'll will say not many do, and most used for recordings will fall shy by 10 db or more. Plenty of microphones can cleanly record SPL levels above 120 db. Some well above it. So I'd say yes if you use the right mike, but usually we've now lost 10 db on the low level end of the scale. So we are usually looking at 110 db of range now being available.
So we are at 110 db left or maybe, maybe if the stars align 120 db. The microphones have to feed a microphone preamp. What happens there. Well if you use gain in the preamp, the gain amplifies the noise as well as the signal. An EIN spec, Equivalent Input Noise is pertinent here. A good one might have -128 dbu EIN with 150 ohm output impedance microphone (which most condensers are). Sounds good. The limit in all this at this point is the thermal noise which is determined by the output impedance of the microphone. At 150 ohms that will be -131 dbu which would be theoretical perfection. However this spec is usually given with 60 db of preamp gain. And it doesn't mean we have 128 db of dynamic range. It means we have a noise level of -68 dbu, and if you subtracted the 60 db gain that is what you'd get if the input without gain were this quiet. You might think if the microphone is sensitive enough, and you could turn down all the gain to zero then we'd get 128 db to work with. But usually microphone preamps don't have their best dynamic range with 0 gain. So what can we get?
To keep from being too confusing I'll skip over some stuff, that if there is interest can be discussed further. With just the right microphone, with just the right sensitivity with some attenuation and low level gain you'll be hard pressed to exceed 110 db dynamic range between the preamp noise, microphone noise and gain levels that amplify that noise. And you'll have to be very, very careful to extract that much. It actually would be very, very rare and difficult to exceed 100-105 db here. More than 100 db is rarely seen. You know when you hear a minimally miked recording, and it starts where you hear the sound of the space of the hall open up before you hear any music? If we had kept our noise levels low enough you wouldn't hear that. It would be effectively silence or close to it. I've made recordings and hear the space prior to the music and thought, "good you can hear where it was, the size of it". But actually it wasn't that noticeable in person, and the musicians often react with the opinion the recording has noise in it. And they were right.
So at this point with all our lucky stars I'm saying you might manage 110 db, probably not above 100 db. So at this point I'd answer the question of the thread topic with No you can't get 120 db into your room. Can you really get 110 db in your room with peaks at 120 db? Well now we have to play all this on a DAC. Let us say we have one of the excellent DACs Amir has tested which has 110 db of dynamic range and SINAD as our source. Alright. Actually a 110 db DAC and 110 db recording would give us 107 db. So we'll need one of those DACs pushing 120 db for us to get close to our mythical 110 db range for our recording.
And then we send the signal to the power amp. Which will provide something like 20-26 db of gain. Which pushes up the noise floor by that much. So how much is left. Looking at something like the Benchmark or some of the Hypex or the new one from Bruno et al, we might actually have something between max power and noise of approaching 130 db, and if we subtract the gain of the input noise level we'd have maybe 105, maybe, maybe, maybe 110 db which preserves our source signal range. And this would be if max power coincides with max output of our connected speaker and it could do 120 db SPL at this level. Those are all possible, but very few combinations will manage it. And this still only gets us 110 db of dynamic range, not the 120 db we were shooting for. Sure in movies or synthetic creations gain riding and such we might get the 120 db, but in the mythical minimally miked super quality stereo recording we can't get there from here. And how many have a speaker that can do 120 db? I find most top out around 105 to 110 db if they are clean. These are pretty darn good speaker that will manage that.
My own Soundlabs might in the midrange get you to 105 db, but will need 500 watts or so to do that. Of course we don't usually need that much in the midrange. Usually big power is need in the lower frequencies. So is this entire exercise a bit daft? With our hearing thresholds at lower frequencies being some 20 db higher maybe with 120 db SPL levels we really don't need more than 100 db anyway. Rarely will music have those levels in the 3-5 khz range where our hearing dynamic range is greatest.
I'll add a final bit about recordings. Do we have any that reach even 100 db range within them? Just as an example, 2L which actually isn't super minimally miked, they use at least several microphones, analysis of their recordings indicate in the silence just prior to the music, the level is only about - 60 dbFS. Remember the sound of the hall? In the moments before music, you hear those halls, and they ramp up the level, but the hall is at - 60 dbFS. They are likely using more than low gain levels on the microphones and this is where the noise of the hall is coming from. In some of my own recordings this is pretty typical for hall sound. I've a few that do a little better than this, but nothing that exceeds about 70 to 75 db.
So, I don't claim to know all the answers, and maybe I've looked at some of this wrong. I hope this post will kick off some interesting discussion among those knowledgeable and experienced with such things. And I hope those which aren't can also post and raise some interesting questions, maybe some interesting and embarrassing questions. So don't hesitate to post an opinion, or a crazy question or to point out where I've messed up in my thinking.
Let me start at the beginning, where the recorded performance takes place. I'll ignore electronic soundscape creations where obviously the answer is yes.
I'll steal this from Amir's postings. Who stole it from Dolby I believe.
We see that our hearing thresholds actually are not too far from noise levels in some concert halls like the Davies Hall in the graph. So those can be quiet enough or nearly so that we'd hear silence from the venue itself. At our most sensitive mid-band we get down to 0 db SPL and a little below. The noise is a little above, but we can hear into it some anyway. So let us say this is close enough to say yes. If some orchestral or other recordings can peak at 120 db SPL we could hear down to 0 db SPL in quiet portions. Though not while music is playing.
Next step is the microphones. Can they be quiet enough? They have self noise. There are some that get close to 0 db SPL equivalent self noise (4-7 db). Some that reach that with A-weighting. So on the low end yes with A-weighting. Or very close to it as far as what we could hear. I'll will say not many do, and most used for recordings will fall shy by 10 db or more. Plenty of microphones can cleanly record SPL levels above 120 db. Some well above it. So I'd say yes if you use the right mike, but usually we've now lost 10 db on the low level end of the scale. So we are usually looking at 110 db of range now being available.
So we are at 110 db left or maybe, maybe if the stars align 120 db. The microphones have to feed a microphone preamp. What happens there. Well if you use gain in the preamp, the gain amplifies the noise as well as the signal. An EIN spec, Equivalent Input Noise is pertinent here. A good one might have -128 dbu EIN with 150 ohm output impedance microphone (which most condensers are). Sounds good. The limit in all this at this point is the thermal noise which is determined by the output impedance of the microphone. At 150 ohms that will be -131 dbu which would be theoretical perfection. However this spec is usually given with 60 db of preamp gain. And it doesn't mean we have 128 db of dynamic range. It means we have a noise level of -68 dbu, and if you subtracted the 60 db gain that is what you'd get if the input without gain were this quiet. You might think if the microphone is sensitive enough, and you could turn down all the gain to zero then we'd get 128 db to work with. But usually microphone preamps don't have their best dynamic range with 0 gain. So what can we get?
To keep from being too confusing I'll skip over some stuff, that if there is interest can be discussed further. With just the right microphone, with just the right sensitivity with some attenuation and low level gain you'll be hard pressed to exceed 110 db dynamic range between the preamp noise, microphone noise and gain levels that amplify that noise. And you'll have to be very, very careful to extract that much. It actually would be very, very rare and difficult to exceed 100-105 db here. More than 100 db is rarely seen. You know when you hear a minimally miked recording, and it starts where you hear the sound of the space of the hall open up before you hear any music? If we had kept our noise levels low enough you wouldn't hear that. It would be effectively silence or close to it. I've made recordings and hear the space prior to the music and thought, "good you can hear where it was, the size of it". But actually it wasn't that noticeable in person, and the musicians often react with the opinion the recording has noise in it. And they were right.
So at this point with all our lucky stars I'm saying you might manage 110 db, probably not above 100 db. So at this point I'd answer the question of the thread topic with No you can't get 120 db into your room. Can you really get 110 db in your room with peaks at 120 db? Well now we have to play all this on a DAC. Let us say we have one of the excellent DACs Amir has tested which has 110 db of dynamic range and SINAD as our source. Alright. Actually a 110 db DAC and 110 db recording would give us 107 db. So we'll need one of those DACs pushing 120 db for us to get close to our mythical 110 db range for our recording.
And then we send the signal to the power amp. Which will provide something like 20-26 db of gain. Which pushes up the noise floor by that much. So how much is left. Looking at something like the Benchmark or some of the Hypex or the new one from Bruno et al, we might actually have something between max power and noise of approaching 130 db, and if we subtract the gain of the input noise level we'd have maybe 105, maybe, maybe, maybe 110 db which preserves our source signal range. And this would be if max power coincides with max output of our connected speaker and it could do 120 db SPL at this level. Those are all possible, but very few combinations will manage it. And this still only gets us 110 db of dynamic range, not the 120 db we were shooting for. Sure in movies or synthetic creations gain riding and such we might get the 120 db, but in the mythical minimally miked super quality stereo recording we can't get there from here. And how many have a speaker that can do 120 db? I find most top out around 105 to 110 db if they are clean. These are pretty darn good speaker that will manage that.
My own Soundlabs might in the midrange get you to 105 db, but will need 500 watts or so to do that. Of course we don't usually need that much in the midrange. Usually big power is need in the lower frequencies. So is this entire exercise a bit daft? With our hearing thresholds at lower frequencies being some 20 db higher maybe with 120 db SPL levels we really don't need more than 100 db anyway. Rarely will music have those levels in the 3-5 khz range where our hearing dynamic range is greatest.
I'll add a final bit about recordings. Do we have any that reach even 100 db range within them? Just as an example, 2L which actually isn't super minimally miked, they use at least several microphones, analysis of their recordings indicate in the silence just prior to the music, the level is only about - 60 dbFS. Remember the sound of the hall? In the moments before music, you hear those halls, and they ramp up the level, but the hall is at - 60 dbFS. They are likely using more than low gain levels on the microphones and this is where the noise of the hall is coming from. In some of my own recordings this is pretty typical for hall sound. I've a few that do a little better than this, but nothing that exceeds about 70 to 75 db.
So, I don't claim to know all the answers, and maybe I've looked at some of this wrong. I hope this post will kick off some interesting discussion among those knowledgeable and experienced with such things. And I hope those which aren't can also post and raise some interesting questions, maybe some interesting and embarrassing questions. So don't hesitate to post an opinion, or a crazy question or to point out where I've messed up in my thinking.