That's one of the reason you must have a approximate target of spl level you need at listening position.
This can be a daughting task for some people: what max output voltage from dac then adding voltage gain from amp and loss from distance... all that allowing some margin for headroom including eventual attenuation along the chain to optimise snr... and when i see what spl level some are targeting i sometimes don't understand choices made!
Last edited:
So how do you guys go about achieving your desired gain structure? I've used an oscilloscope while playing a few different frequencies of sine waves to find where outputs start to clip. But my goal was maximum output as long as I didn't start to hear hissing.
With the setup I'm trying to create, the analog out from my DACs will be feeding directly into amplifier channels to drive individual speakers. So I don't know that I'm going to have a lot of options for keeping volume low through amp gains. With my Crown amp I intend to use for subwoofers I have a choice of input sensitivity of 1.4 or .775 volts. But it has volume knobs for both channels on the front. I don't know how the volume works but I don't really care for sub bass frequencies.
But for the other 6 channels I have a dilemma. I would like to buy enough amplifier power to handle whatever I might throw at it down the road. And I also like the idea of having plenty of amp headroom. But will I be able to get my listening level down where I want it without digital attenuation if I have too much power?
I may not have much choice but to place something between the DAC and the amps to attenuate.
-Chris
With the setup I'm trying to create, the analog out from my DACs will be feeding directly into amplifier channels to drive individual speakers. So I don't know that I'm going to have a lot of options for keeping volume low through amp gains. With my Crown amp I intend to use for subwoofers I have a choice of input sensitivity of 1.4 or .775 volts. But it has volume knobs for both channels on the front. I don't know how the volume works but I don't really care for sub bass frequencies.
But for the other 6 channels I have a dilemma. I would like to buy enough amplifier power to handle whatever I might throw at it down the road. And I also like the idea of having plenty of amp headroom. But will I be able to get my listening level down where I want it without digital attenuation if I have too much power?
I may not have much choice but to place something between the DAC and the amps to attenuate.
-Chris
pano, im really surprised by digital attenuation. i have noticed obvious degradation in my headphone system and speaker system. subtle with very little att but very obvious with more attenuation. what is your system?
Last edited:
Iscream: first you have to choice a target spl at listening position.
This can not be an easy task because in a way this is dependent from your habits, limits given by choice you made about loudspeaker and the room*.
A nice point of reference imho is to go see a movie in a theater: why you ask? Because most of this place are calibrated to have a 85dbspl target for stereo reproduction (it varies between seats but given you are in the 'middle' of room it should be ok). If you think this is ok level for you when soundtrack is played back then let's start with that for example.
Which distance is your system located from your listening position? Let's say 2meters for our example: sound loose 6db each time you double distance (given you are not using line array for which it is -3db).
So, you are located 2m away from speakers you need 6 db more output from loudspeakers: 91dbspl.
Ok. You listen to jazz and blues maybe sometimes classical music. In such material it is wise to consider you'll have 20db crest factor. So we target 91dbspl+20db crestfactor: 111dbspl worst case (better case in fact as this mean you have dynamic range in music which is good).If you listen pop not heavily mastered crest factor is more on the 14db range, heavy mastered music and broadcast is 12db. Worst case (last Metallica lp) 6db!
If by chance your loudspeakers are 91dbspl/1w/1m you'll need 20db gain from your power amplifier. This equal to 128w used for your max output (each time you double power you gain +3db). 85dbspl target is given for stereo so for mono it's 83dbspl so with the given number you have 3db headroom already in chain. If you want 6db headroom at amplifier we have to double power so 256w per loudspeaker give you 6db headroom for your target. Nice and should be ok guaranty you amp wont clip and should'nt destroy tweeter.
Now we must see sensibility of amplifier for max output and voltage gain induced.
Lets say your amp have a voltage gain of +24db and max input voltage is 1.4v (rms) (approximate +5dbu) now what you want is to have a max input voltage -6db from this point (headroom from amp) which equal to -1dbu:0.7v (rms).
Now what you need to do is to find max level from your dac and restrict it's output level to 0.7v (either using a digital attenuation in using a value which in fact will be a maximum level allowed ) or take the maximum output voltage of dac (if i remember correctly and i correctly read the datasheet is +8dbu for the ones you have choosen) and you a pad to give it max allowed voltage for amp ( for this case you'll need 9db of attenuation from 8dbu to -1dbu). If you choose to pad in analog take care to use a kind of attenuator which give you same impedance input as the one from your amp and now maximum digital volume allowed is 0dbfs (max out in digital).
If you choose to use analog pad an other point to take into account is the input of the amp and try to keep it the same as converters: balanced with balanced, unbalannced with unbalanced. It will save on headache. Anyway if you have to mix and match there is solution: buffer. But it will add a stage and it will be active.
It's an example and other may do it differently. That's how i done mine.
*One thing to note: 85dbspl may be to high a target as volume of room play a role in the way our brain accept spl. If your room is small target can be lowered down by 10db to 75dbspl to have same sensation as in theater which obviously is a big room.
This can not be an easy task because in a way this is dependent from your habits, limits given by choice you made about loudspeaker and the room*.
A nice point of reference imho is to go see a movie in a theater: why you ask? Because most of this place are calibrated to have a 85dbspl target for stereo reproduction (it varies between seats but given you are in the 'middle' of room it should be ok). If you think this is ok level for you when soundtrack is played back then let's start with that for example.
Which distance is your system located from your listening position? Let's say 2meters for our example: sound loose 6db each time you double distance (given you are not using line array for which it is -3db).
So, you are located 2m away from speakers you need 6 db more output from loudspeakers: 91dbspl.
Ok. You listen to jazz and blues maybe sometimes classical music. In such material it is wise to consider you'll have 20db crest factor. So we target 91dbspl+20db crestfactor: 111dbspl worst case (better case in fact as this mean you have dynamic range in music which is good).If you listen pop not heavily mastered crest factor is more on the 14db range, heavy mastered music and broadcast is 12db. Worst case (last Metallica lp) 6db!
If by chance your loudspeakers are 91dbspl/1w/1m you'll need 20db gain from your power amplifier. This equal to 128w used for your max output (each time you double power you gain +3db). 85dbspl target is given for stereo so for mono it's 83dbspl so with the given number you have 3db headroom already in chain. If you want 6db headroom at amplifier we have to double power so 256w per loudspeaker give you 6db headroom for your target. Nice and should be ok guaranty you amp wont clip and should'nt destroy tweeter.
Now we must see sensibility of amplifier for max output and voltage gain induced.
Lets say your amp have a voltage gain of +24db and max input voltage is 1.4v (rms) (approximate +5dbu) now what you want is to have a max input voltage -6db from this point (headroom from amp) which equal to -1dbu:0.7v (rms).
Now what you need to do is to find max level from your dac and restrict it's output level to 0.7v (either using a digital attenuation in using a value which in fact will be a maximum level allowed ) or take the maximum output voltage of dac (if i remember correctly and i correctly read the datasheet is +8dbu for the ones you have choosen) and you a pad to give it max allowed voltage for amp ( for this case you'll need 9db of attenuation from 8dbu to -1dbu). If you choose to pad in analog take care to use a kind of attenuator which give you same impedance input as the one from your amp and now maximum digital volume allowed is 0dbfs (max out in digital).
If you choose to use analog pad an other point to take into account is the input of the amp and try to keep it the same as converters: balanced with balanced, unbalannced with unbalanced. It will save on headache. Anyway if you have to mix and match there is solution: buffer. But it will add a stage and it will be active.
It's an example and other may do it differently. That's how i done mine.
*One thing to note: 85dbspl may be to high a target as volume of room play a role in the way our brain accept spl. If your room is small target can be lowered down by 10db to 75dbspl to have same sensation as in theater which obviously is a big room.
Last edited:
@Krivium: You should have taken one step back as reference levels can be confusing, not least because movies actually obey a spec and recorded music doesn't (although I find the THX adverts, parodied by the simpsons as 'the audience is now deaf' as way too loud).
85dB at the listening position for 105dB peaks works for classical and films may not work for pop music that is compressed within an inch of its life. I have sized for 100dB peak at listening position, which is a tad high for a small room but gives me flexibility.
But if you don't have an AV receiver you need to generate the ref level at -20dbFS.
For pre-loudness wars pop I believe -12dBVU was used. post loudness wars I have seen stuff hanging around -6dBVU. so for either of those cases an extra 10dB or so of attenuation is needed (if SWMBO allows proper listening levels in the first place!)
85dB at the listening position for 105dB peaks works for classical and films may not work for pop music that is compressed within an inch of its life. I have sized for 100dB peak at listening position, which is a tad high for a small room but gives me flexibility.
But if you don't have an AV receiver you need to generate the ref level at -20dbFS.
For pre-loudness wars pop I believe -12dBVU was used. post loudness wars I have seen stuff hanging around -6dBVU. so for either of those cases an extra 10dB or so of attenuation is needed (if SWMBO allows proper listening levels in the first place!)
I did shoot for about 85 dB average listening level. It works pretty good but Pop songs of today wouldn't require the same amount of headroom to play at 85 dB average as the peaks would only be about ~6 dB louder.
Movies and classical music would require the headroom though. But it can get pretty loud in a room. If your speakers (and room) are up for it you wouldn't notice it though, if they are clean enough. Not unless you try to start a conversation.
Pop music with low dynamic range isn't always very pleasing turned up above that average. The compression makes me cringe at times. But at ~85 dB average it can be a lot of fun.
More dynamic music can be turned up way more and remain pleasing. Got to watch out for it though, you get too loud fast!
I keep a RadioShack SPL meter around to check and keep my levels sane.
With music that has a wider dynamic range I do have to turn my amps up in comparison to less dynamic recordings. So in a way I still use passive attenuation 😱. I'll admit I don't attenuate digitally for critical listening. I already use up a lot of digital space in the processing of my arrays.
But people should be aware of their room when you talk about this noise floor. I'm in a living room and was lucky to get my first reflections down to about -20 dB (or more) after the main peak. That's the start of the masking on the bottom. We are not talking about studio levels here.
This graph gives a basic view but does not tell the complete story. You're mainly looking at high frequency.
An APL_TDA view lets you see the reflections arriving in time:
In this plot you see the reflections coming out above the noise floor (and still somewhat coloring the sound).
This also shows why I use active FIR based DSP, as that was part of the subject here. But you can actually link the peaks in the Filtered IR plot to the bumps seen in the APL_TDA plot.
Look here for an example of a very clean room TDA plot, with a passive speaker: http://www.diyaudio.com/forums/full-range/273971-group-delay-questions-analysis-50.html#post4581047
The ridge you see in that plot at 24 ms is deliberate, a Haas kicker. But the difference compared to my room is obvious.
The same can be said about the processing, as my attempt to create time coherency finishes up the signal before the passive speaker does.
But my room will still be "singing along" after that first wave.
Both graphs were made with a stereo pair of speakers playing and measured at the listening position. The TDA plot was made with a demo version of APL_TDA. My own chain of processing was used to correct speaker behavior using JRiver, REW and DRC-FIR (custom template).
Last edited:
Ran out of time editing but I wanted to add: This post needs a disclaimer, the above results were made with use of 3 (rather large) damping panels at strategic positions. DSP cannot fix everything. It isn't a substitute for room treatment. The post I linked to is completely done with room treatment, but as a result does not resemble a living room anymore.
Billshurv: IScream gave us some clues about what kind of music he listen, so i took it as a 'worst case' scenario. I specified crest factor for other genre of music: pop,etc,etc...
To have the needed spl target you just need to replace 20db by the values given (14db,12db and if you listen to Metallica lastest lp: 6db). This safe as these values are all well below max dynamic range of 20db used for acoustic music: the demand on absolute spl is lessen (but it doesn't mean it's safer for your drivers...).
Yes i could happily give access to a wav file including pink noise at -20dbfs in different versions if needed.
This is the file to use to define target of 82dbspl (mono) at listening position.
-xxdbVU doesn't mean anything: vu meter are calibrated to 'inhouse' reference levels in studios. It can be +4dbu, or 0 dbu or whatever condidered useful for the application.
The target for digital in studios is usually -18dbfs. This means something because it's related to max digital level. It can or not be convenient for the use intended. I do prefer have some 20db crest factor. So the numbers given in previous post.
For pop music preloudness it's more in the 14db crest factor range (this is an average you have exceptions).
This is all about what we discuss for 2 or 3 days. For convenience and ease of use , i'll say you use digital volume control as Wesayso or Pano do and suggest.
If you care about same things as i am you use switched banks of passive pads.
You can even use both (pads for coarse volume control and digital for fine tuning) as it seems to me to give the 'cons' of the two way to do! 🙂
All the previous post in an example Billshurv. It's not given as an absolute reference. The fact that i give advice to go in a theatre is because this is one of the rare place where spl are calibrated (or should be). It's convenient because there's always a track played during a film and it can be a nice point of reference (It was obvious listening to Adele's 'Skyfall' in the movie of the same name: some artefacts of track producer's choice was enlightned at this level and in a big room with correctly calibrated monitors vs the track broadcasted by radio for example.)
I agree. You adress the issue of Early Reflections in your post so i won't do it again, but that's why i gave the advise to adjust target up to 10 db attenuation depending the size of room.
At least! Thruth in his naked nature! 😛 lol
I was sure that with an amplifier using a volume control as everyone on this earth you couldn't resist in using it! 🙂 Sane practice to have a spl meter at hand. I do the same! You'll end up working in a studio Ronald! Be careful!
Noise floor and ER are not exactly the same subject but yes results could be grouped as you do, well in a way i see what you mean.
Noise floor is nice subject on his own...
About yout ER you've managed some very nice results in your room! I would really love to have same results, even in some studios i used to work in! 🙂
Most place do. Except if you are in T. Hidley's zero environnment control room. In itself, as long at the ER arrive to listener with 20ms and 20db attenuation and that loudspeaker as a nice smooth power transfer this is not really a problem ( your brain is able to discern reverb of room from direct radiated loudspeaker sound, especially if your listening position is located at critical distance of room).
Obviously critaical distance is dependant from room's volume (the bigger the greater the critical distance) so that's why i said determining target level is not an easy task as it should vary with that parameter too so vary from room to room...
To have the needed spl target you just need to replace 20db by the values given (14db,12db and if you listen to Metallica lastest lp: 6db). This safe as these values are all well below max dynamic range of 20db used for acoustic music: the demand on absolute spl is lessen (but it doesn't mean it's safer for your drivers...).
Yes i could happily give access to a wav file including pink noise at -20dbfs in different versions if needed.
This is the file to use to define target of 82dbspl (mono) at listening position.
-xxdbVU doesn't mean anything: vu meter are calibrated to 'inhouse' reference levels in studios. It can be +4dbu, or 0 dbu or whatever condidered useful for the application.
The target for digital in studios is usually -18dbfs. This means something because it's related to max digital level. It can or not be convenient for the use intended. I do prefer have some 20db crest factor. So the numbers given in previous post.
For pop music preloudness it's more in the 14db crest factor range (this is an average you have exceptions).
This is all about what we discuss for 2 or 3 days. For convenience and ease of use , i'll say you use digital volume control as Wesayso or Pano do and suggest.
If you care about same things as i am you use switched banks of passive pads.
You can even use both (pads for coarse volume control and digital for fine tuning) as it seems to me to give the 'cons' of the two way to do! 🙂
All the previous post in an example Billshurv. It's not given as an absolute reference. The fact that i give advice to go in a theatre is because this is one of the rare place where spl are calibrated (or should be). It's convenient because there's always a track played during a film and it can be a nice point of reference (It was obvious listening to Adele's 'Skyfall' in the movie of the same name: some artefacts of track producer's choice was enlightned at this level and in a big room with correctly calibrated monitors vs the track broadcasted by radio for example.)
I agree. You adress the issue of Early Reflections in your post so i won't do it again, but that's why i gave the advise to adjust target up to 10 db attenuation depending the size of room.
At least! Thruth in his naked nature! 😛 lol
I was sure that with an amplifier using a volume control as everyone on this earth you couldn't resist in using it! 🙂 Sane practice to have a spl meter at hand. I do the same! You'll end up working in a studio Ronald! Be careful!
Noise floor and ER are not exactly the same subject but yes results could be grouped as you do, well in a way i see what you mean.
Noise floor is nice subject on his own...
About yout ER you've managed some very nice results in your room! I would really love to have same results, even in some studios i used to work in! 🙂
Most place do. Except if you are in T. Hidley's zero environnment control room. In itself, as long at the ER arrive to listener with 20ms and 20db attenuation and that loudspeaker as a nice smooth power transfer this is not really a problem ( your brain is able to discern reverb of room from direct radiated loudspeaker sound, especially if your listening position is located at critical distance of room).
Obviously critaical distance is dependant from room's volume (the bigger the greater the critical distance) so that's why i said determining target level is not an easy task as it should vary with that parameter too so vary from room to room...
I wish 😀
Yes, I do understand the concept of noise floor, but to maximize listening in our rooms we do have to deal with the room. Which in itself can mask the music. We need to be realistic about that too in these discussions.
I know! That was reason enough for me to work my way there. Don't count on DSP alone to get there though. But it's very rewarding and worth the troubles.
It's kind of sad that we talk this much about speakers and most of the time, we leave the room out of the discussions. In my opinion you really need to have both of them working closely together. I think that is key to get the most out of it. Do not only focus on building speakers but plan the room you intend to use them in and pick the speaker design that could work with your room. Keeping an eye on placement restrictions (if any).
In the example given there may be some wrong values given! Please apologize, it was a very late post (or very early depends on point of view) with baby crying in background...
I've alredady spot one:
this is not 83dbspl but 82dbspl for mono.
If you spot others just correct them! I don't have access to edit post anymore.
I've alredady spot one:
this is not 83dbspl but 82dbspl for mono.
If you spot others just correct them! I don't have access to edit post anymore.
I totally agree! This is a whole system and every link is 'interdependant' (interleaved). Room is almost everytime the weakest link.
But to take on the whole you have to discuss the each parts on it's own as loudspeakers are already difficult to understand/master. The problem is that we rarely step back about the whole system. And at home acoustic treatments can be difficult to deal with (estheticaly, waf,usability of room,...).
About the DSP yes this is not a Panacea! And why trying to deal with problems that acoustical and loudspeaker design have the ability to already fix gracefully?
Thanks a ton to all of you for the help and advice. In my previous house with a DIY home theater I set my five channels at 105 DB using test tones, or maybe it was pink noise, recorded with 0 DB attenuation.
Not sure if I will go with that for this music only system or not though. I have an SPL meter and one of the Dayton USB measurement microphones with the calibration file you download.
I'll probably crank up the spectrum analyzer in REW while I listen to some different types of music and just see where the average and peaks end up when listening at a level I enjoy.
I do listen to some pop, rock and metal as well as the blues and jazz. I have to be in the right mood for the classical stuff so it doesn't happen very often.
I have some questions about what I read over the last couple pages but I need to get logged into work and get my head into picking bugs out of code.
I really do appreciate the help.
-Chris
Not sure if I will go with that for this music only system or not though. I have an SPL meter and one of the Dayton USB measurement microphones with the calibration file you download.
I'll probably crank up the spectrum analyzer in REW while I listen to some different types of music and just see where the average and peaks end up when listening at a level I enjoy.
I do listen to some pop, rock and metal as well as the blues and jazz. I have to be in the right mood for the classical stuff so it doesn't happen very often.
I have some questions about what I read over the last couple pages but I need to get logged into work and get my head into picking bugs out of code.
I really do appreciate the help.
-Chris
That's a nice idea but you'll need some prerecorded pink noise to have a true idea of volume, it'll be very difficult to estimate level with a track (because they don't have same rms level all the time, being music freq change all the time, etc,etc,....).
To procede you'll have to choose some tracks you like (different genres if possible some quiet, some hot) play them at fixed volume changing settings until you feel confortable (with a quiet one) then once found the seet spot play a -20dbfs pink noise mono signal and then you use your spl meter to read your target (-20dbfs because it's way less painful than a 0dbfs pink noise! Safier for your ears and if you have earplug use them just in case of...).
When changing from genres to genres you'll have to attenuate the sound level (using digital volume control in player): this will be obvious when you'll switch from metal to jazz for example!
For metal you'll probably attenuate around -12/-14db range for jazz more in the -2/-6db range...
Anyway it'll be OBVIOUS! 🙂
Good morning from the middle of the Pacific Ocean. A lot of posts since I last checked - and a lot of really good discussion. 
As Bill wisely states, gain structure is easy to get right, but also easy to get wrong. And it's vitally important when using active crossovers. Target SPL? That's a tough one. How do you decide?
On my previous system I went for 109dB peaks at the listening position. That was not too difficult considering the size of the room and the efficiency of the speakers. That allowed good dynamics for classical music and way more than enough (too much) average level for rock/pop/jazz. In my present system, I'd be burning up drivers running that high. Peaks at 98dB may be all she will reasonable do.
"That's all she's got capt'n, I can no give her any more!"
How do you set that? I use pink noise with an RMS value of -14dB FS. Since the level is known in relation to Full Scale (FS), it's easy to calculate peak value. I run the pink noise with the digital volume at 100% and watching the SPL meter (slow average) bring up the amp volume controls until the target is met. If I want peaks of 98dB, I set the pink noise at 84dB. If I want peaks of 109dB the pink is set to 95dB. That's really loud, BTW. I rarely ran at that level.
As Bill wisely states, gain structure is easy to get right, but also easy to get wrong. And it's vitally important when using active crossovers. Target SPL? That's a tough one. How do you decide?
On my previous system I went for 109dB peaks at the listening position. That was not too difficult considering the size of the room and the efficiency of the speakers. That allowed good dynamics for classical music and way more than enough (too much) average level for rock/pop/jazz. In my present system, I'd be burning up drivers running that high. Peaks at 98dB may be all she will reasonable do.
"That's all she's got capt'n, I can no give her any more!"
How do you set that? I use pink noise with an RMS value of -14dB FS. Since the level is known in relation to Full Scale (FS), it's easy to calculate peak value. I run the pink noise with the digital volume at 100% and watching the SPL meter (slow average) bring up the amp volume controls until the target is met. If I want peaks of 98dB, I set the pink noise at 84dB. If I want peaks of 109dB the pink is set to 95dB. That's really loud, BTW. I rarely ran at that level.
How do you test this? I mean, how do you compare the same amount of analog and digital attenuation? Level matching is very important in AB tests.
Yes, that is a lot of attenuation. Do you normally go from 0dB attenuation down to -35dB? That's massive. On a normal analog volume knob, 12:00 is 20dB attenuation. How much would you move the volume knob?
Let's look at it this way. If your signal is being fed in 24 bits to your DAC, but your DAC has only a 20 bit equivalent S/N, what is that. That's 120dB. You could attenuate 24dB and still have the same S/N ratio as 16 bit CD. Think about that.
Another think to think about is "i'm absolutely in needs too that the character of the sound of the dac don't change". If you are changing your volume control by 20 to 35dB, the analog sound of your DAC is the least of your worries. Fletcher-Munson becomes the sticking point. What you need is a loudness control. And maybe some compression. You just don't hear the same at low levels as you do 20-30dB higher. It's your ear getting in the way more than the DAC.
Pano i get your point about digital attenuation no worry.
Normally (for entertainment) i usually don't need so much attenuation, i already said it i agree about the range of approximately 10db you gave.
But i use my system for mastering duties too.
In this particular case i need to hear what some dynamic and eq treatments do at my 'regular' listening level for work (around 82dbspl) but too at very low volume to hear if the change made which are sometimes very subtle are ok.
I'm aware of Fletcher-Munson curves and it is why i go so low in level sometimes.
The last things i want are a loudness control and a compressor or limiter because most of the time using so low volume is to check that the final brickwall limiter/optimizer are not audible in the chain. Or that the eq setting i've choosen over a compression still give the effect at lower volume. Or to be sure that after some time working at a constant level my brain is not fooling me about choice i make. 🙂
Sometimes eq are so subtles (shelf of +-0.5db) that they are not obviously heard, but long term and with change in level you get a 'feel' of the effect and if it is needed or not. There's something equivalent when you engage a treatment, listen to it and bypass it and listen at the same passage. When you do the inverse (bypass to on) you don't hear the same things... Some artefacts or effects are revealed one way not nescesserely the other (or other things are revealed the other way).
The nescessity i have to take dac out of the equation is here. Once you know how it sound 'full range' (including artefacts from analog part of the converter as you come close to 0dbfs :reconstruction of analog sometimes create peak between samples which are not present in the digital signal and are responsible of some attribute qualified as harsh sounding, kind of haze or transient smearing or other things which in some case may even be euphonics) once you get back some db it sound different and you can't rely on them to make objective choice. This is all about knowing how you converter sound (including artefacts) and be able to identify them when tey are presents and why.
But i agree it is very particular and not something everyone care about. But i'am. And i repeat i've no fear to perform digital volume control (i use level automation as much as fade in fade out and plug ins) just that in my chain they are not appropriate as a master level control as they change the sound character of dac.
Normally (for entertainment) i usually don't need so much attenuation, i already said it i agree about the range of approximately 10db you gave.
But i use my system for mastering duties too.
In this particular case i need to hear what some dynamic and eq treatments do at my 'regular' listening level for work (around 82dbspl) but too at very low volume to hear if the change made which are sometimes very subtle are ok.
I'm aware of Fletcher-Munson curves and it is why i go so low in level sometimes.
The last things i want are a loudness control and a compressor or limiter because most of the time using so low volume is to check that the final brickwall limiter/optimizer are not audible in the chain. Or that the eq setting i've choosen over a compression still give the effect at lower volume. Or to be sure that after some time working at a constant level my brain is not fooling me about choice i make. 🙂
Sometimes eq are so subtles (shelf of +-0.5db) that they are not obviously heard, but long term and with change in level you get a 'feel' of the effect and if it is needed or not. There's something equivalent when you engage a treatment, listen to it and bypass it and listen at the same passage. When you do the inverse (bypass to on) you don't hear the same things... Some artefacts or effects are revealed one way not nescesserely the other (or other things are revealed the other way).
The nescessity i have to take dac out of the equation is here. Once you know how it sound 'full range' (including artefacts from analog part of the converter as you come close to 0dbfs :reconstruction of analog sometimes create peak between samples which are not present in the digital signal and are responsible of some attribute qualified as harsh sounding, kind of haze or transient smearing or other things which in some case may even be euphonics) once you get back some db it sound different and you can't rely on them to make objective choice. This is all about knowing how you converter sound (including artefacts) and be able to identify them when tey are presents and why.
But i agree it is very particular and not something everyone care about. But i'am. And i repeat i've no fear to perform digital volume control (i use level automation as much as fade in fade out and plug ins) just that in my chain they are not appropriate as a master level control as they change the sound character of dac.
Last edited:
level matched of course.
I guess if you use active speakers, it makes no point in analog volume control, meanwhile with passive speakers, analog volume control is better.
One thing for sure, in my setup, dac direct into headphone amp or speaker amp and using digital attenuation degrades the sound.
Last edited:
- Status
- Not open for further replies.