Latest Thread Images

Amps, Volume, and Transients
Jan 5, 2016 at 11:55 AM Thread Starter Post #1 of 18
Maconi

Maconi

100+ Head-Fier
Joined
Jul 27, 2015
Posts
192
Likes
37
I'm new to "sound science" but I love to learn and a topic I've been trying to pin down better are transients and how they relate to power/volume. I know that in order to produce transients you need the required power. My question is how does volume actually correspond to that power requirement?
 
Lets say I have an Audeze headphone (below is an image snagged from the LCD-3 page). It's happy with anything between 1W and 4W of power. It also appears to be able to handle a 15W transient.
 

 
Lets say I have an amp that can easily supply 15W for the given impedance (110 Ohms). So it should be able to push the headphone to its max ability.
 
My question is how does power relate to volume? Obviously the volume will become overwhelming way before the "optimal power requirement" range. Are you expected to use external volume control (preamp) so that the volume is still comfortable while the headphone is being supplied with the "optimal power" range? Also, how do transients play in (if I'm pushing 4W of power into the headphones, how do I know the amp isn't going to throw a 30W transient at the headphone causing damage)?
 
Excuse me if I'm completely off-base/misunderstanding. Still trying to wrap my head around everything. 
beerchug.gif
 
 
Jan 5, 2016 at 12:51 PM Post #2 of 18
The sound pressure the headphone produces is directly related to delivered power by the sensitivity or efficiency spec. For the LCD-3, the sensitivity is 103dB @ 1mW, meaning if you deliver a continuous 1mW of power to it, it will produce a continuous 103dB of sound pressure. If you delivered 1W of continuous power, that would be 133dB, and 15W would be about 145dB. 145dB is extremely high, I would be surprised if it could actually go that loud. It would heavily distort long before that. But in theory, that's what it would do if it was completely linear.
 
Most people do not listen to continuous tones, they listen to music which has transient peaks which go higher than the average or perceived loudness. Say your music has transients which peak 10dB above the average level. You would need to be able to produce an extra 10dB of SPL, and your amp would have to supply 10x more power. So if you want to listen to that music at 80dB SPL, you would need the power capability to reach 90dB. The EBU R128 broadcast standard for loudness normalizes all content to -23dB, allowing for 23dB of room for transient peaks. That number was chosen because it has enough headroom for practically anything. I doubt you would ever actually need that much headroom, but if you wanted to listen to something with transients that high, you would need 200x more power than the average level.
 
The power capability of an amp does not matter unless it is actually delivering that much power. So if the LCD-3 is producing 103dB, that means the amp is delivering, and the headphone is receiving, 1mW of power. If you had a preamp volume control and reduced it by 3dB, the amp would no longer be delivering 1mW. It would deliver 0.5mW, which causes the headphone to now produce only 100dB. If you're playing music at average level of 103dB, and it peaks at 10dB above that, the amp would momentarily deliver 10mW of power.
 
Your amp is not going to suddenly deliver 30W of power unless the audio it is playing has transients that would go that high. If the amp delivered a 30W transient, that would cause the LCD-d to produce 148dB. As I said above, the highest transients I would ever expect to see would be 23dB. That means to be able to produce that 30W transient spike, you would have had to be listening at an average SPL of over 125dB, which would mean that you are an insane person who is also deaf now. If you have the volume turned way down somewhere and then somehow accidentally turn it up a lot, that might cause you to deliver a lot of power to the headphone. That probably won't happen though. And at that point I would be more worried about damaging your ears then your headphone. As long as it's not too loud for you to listen to, your headphone will be absolutely fine.
 
Jan 5, 2016 at 4:49 PM Post #3 of 18
mindsmirror said:
  The power capability of an amp does not matter unless it is actually delivering that much power. So if the LCD-3 is producing 103dB, that means the amp is delivering, and the headphone is receiving, 1mW of power. If you had a preamp volume control and reduced it by 3dB, the amp would no longer be delivering 1mW. It would deliver 0.5mW, which causes the headphone to now produce only 100dB. If you're playing music at average level of 103dB, and it peaks at 10dB above that, the amp would momentarily deliver 10mW of power.
Click to expand...

 
Thanks for the response. 
beerchug.gif

 
So you're saying that a 0dB preamp with the amp's volume knob @ 9 o'clock and a  -30dB preamp with the amp's volume knob @ 3 o'clock (resulting in the same volume in this hypothetical) results in the same amount of Power (Watts) being delivered into the headphones?
 
What about Current (Amperage)? I've heard using a negative preamp to push your amp harder (while maintaining the same volume) results in higher current which seems to improve the sound (be it soundstage, transients, etc.). Is there any truth to that?
 
Jan 5, 2016 at 5:32 PM Post #4 of 18
Yes, you are correct in that hypothetical.
 
Power, current, and voltage are all related by Ohm's Laws which say V = I*R, and P = I^2*R. R is resistance, but it is interchangeable with impedance which has the symbol Z.
 
The impedance of your headphone does not change. That means if you increase voltage, current must also be increased, and vice versa. Increasing current also must increases power. So there is no way to increase current, voltage, or power while maintaining the same volume, since these things are all related to each other.
 
I'm not sure what a negative preamp is. You would typically use a preamp if you have a low signal like from a turntable, to increase the signal so you will get less noise and so the amp doesn't require as much gain. But in the end if the amp is outputting the same volume/voltage, it's shouldn't be driven more or less by doing this. I know of using a preamp to increase gain of the signal going into an amplifier to drive the input stage harder and produce distortion. This is very common in electric guitar playing where distortion in the amplifier's input stage is desirable, because to get distortion in the output stage of the amp it would need to be very loud.
 
Jan 5, 2016 at 5:48 PM Post #5 of 18
mindsmirror said:
  I'm not sure what a negative preamp is. You would typically use a preamp if you have a low signal like from a turntable, to increase the signal so you will get less noise and so the amp doesn't require as much gain. But in the end if the amp is outputting the same volume/voltage, it's shouldn't be driven more or less by doing this. I know of using a preamp to increase gain of the signal going into an amplifier to drive the input stage harder and produce distortion. This is very common in electric guitar playing where distortion in the amplifier's input stage is desirable, because to get distortion in the output stage of the amp it would need to be very loud.
Click to expand...

 
I'm referring to the opposite. Using an negative equalizer preamp (like on a PC) to reduce the source signal gain so the amp has to be driven "harder" to get the signal to the same volume level as it would be without the negative preamp. I was curious as to the science behind it. Some people say they feel like doing so opens everything up, but I'm curious as to if they're just creating pleasing distortions (similar to tube amps) rather than truly "unlocking" any potential in the headphones/amp by putting more power/current through them (if doing so even results in increased power/current, as you say it should still be the same voltage @ the same volume)?
 
Jan 5, 2016 at 7:46 PM Post #6 of 18
It seems to me like there is more placebo then science behind that negative preamp idea. If the amp is already being over driven like what I mentioned with electric guitar amps, then reducing the preamp gain would help clean that up. Otherwise it doesn't make sense, it would only reduce the signal strength, i.e. reduce the signal to noise ratio which is not what you want. Can you show me where someone said that? I'd be curious to see their explanation.
 
Jan 5, 2016 at 8:13 PM Post #7 of 18
one e-cookie to @MindsMirror for being such a good host again. didn't check the numbers, but they certainly give the right ideas.
 
reducing the volume on the computer can be done withing reason. you just tell the computer to output the signal in 24bit and that gives you plenty of headroom to set your volume from the computer without crippling the signal's bit depth. but to do it to improve the sound of the amp, it's unlikely. unless the amp is clipping like said above, or if you really have the worst time setting the loudness on the amp(not enough precision or channel imbalance at low volume). otherwise just use your amp as it should be used and don't concern yourself too much with special stuff that are supposedly better but used only by a handful of self proclaimed enlightened people ^_^.
 
the 1-4W optimal power requirement always cracks me up on audeze's website. even funnier is that they seem to go along with it for headphones with very different sensitivities.
confused.gif

 
Jan 5, 2016 at 8:45 PM Post #8 of 18
What first led me into amp volume knob position vs eq preamp were comments by iFi saying the best performance for their amp was between 12 o'clock and 3 o'clock (either using the built-in attenuation or eq preamp manipulation was recommended so that you could get the amp to that knob position while avoiding unhealthy dB levels). I believe that was due to imbalance with their potentiometer or distortion/noise/Class A limit issues.
 
ifi audio said:
 
iancraig10 said:
Is there any audible effect when power is changed? This may be placebo, but for some reason, I think my hd650 sounds slightly better on turbo than normal. Perhaps a little more dynamic; especially with hi resolution files.

Of course it's totally subjective and could be due perhaps to volume change, although (obviously) I have to lower it a great deal.

For some reason, I feel that the Senn is more lively on the higher setting.
Click to expand...

 
Hi,
 
Yes for sure because it is about the micro DSD & headphone matching.
 
Take for example the Remiyo PAT-777 which is a 300B power amplifier. If it was paired with a pair of Magico Q1s, then this combination would sound dull and lifeless.  Neither the PAT-777 nor the Q1s are poor products. Far from it in fact.
 
It is just a simple matter of 7W+7W into 86dB sensitivity speakers being a poor match.
 
The same principle applies to any amplifier & heaphones.
 
 
For example, the HD650 being an average-ish sensitivity headphone would not "sing" if it ran in Eco mode. We would recommend trying
 
1) Normal + iEMatch off
or
2) Turbo + High-Sensitivity
 
Ideally get the listening level to 3 o clock is when the matching is for us (the least resistance, hence the most resolution/dynamics), the most ideal. For us, anything <12 o clock is not ideal. Where the volume position is, does not reflect the power.
 
A lot of people say when the loudness of an amplifier goes to 12 o clock, it has a lot of power. This just isn't the case.
Click to expand...

ifi audio said:
  Last week, we received this from a very reputable customer (who shall remain annoymous) and thought interesting questions were raised. And so here is the Q&A for some of you who may find useful:
 
 
 
Hardware:
 
> I read your explanaition about amping the hd650 on head.fi. And it worked well with Eco 3o clock and IEM match off.
> Since i should use it on 3o clock to get the minimum distortion. What about <50 ohm HPs? (Fidelio X2 and Senn Momentum)
 
As a rule, the power mode for the iDSD micro which places the volume control at around 12 o'Clock to 3 o'Clock with a given headphone offers:

# Maximum Class A power, before switching to Class B
# Lowest measured distortion
# Lowest noise
Click to expand...

 
So that's the first issue. Is it generally accepted that 12-3 o'clock is the best knob position for amps or is that an iFi design issue?
 
Second was the line of thought that the EQ preamp could be used as trimpots for gain staging. I can't find some of the posts I'm looking for but this guy goes into it a bit.
 
tbritton said:
I own Ultrasone Proline 750 headphones, and they have great, tight bass, but draw a ton of current, even though they are only 40-ohm impedance phones.

Driving these from the built-in Via audio deck sound card in my Averatec laptop is a challenge. With everything "normal" the bass was clearly topping out, becoming flabby and distorted, compressing itself as a result of its running out of current.


Many years ago I was an apprentice in sound reinforcement systems, and the professional methodology everywhere I went was always to push the mixing boards' instrument channels and output mains all the way up, initially doing what they called "trimming" at the inputs by adjusting their signal levels to optimize their levels for the board as well as set their initial maximums, using dedicated "trim pots" (really preamp sections) on the signals as they came into the board, and they maxed-out the power-amps gain to the speakers as well, to take advantage of the highest current flow possible and thus create the maximum "headroom." (Mixing was then accomplished by "pulling back" on the instrument and sub-mix sliders in relationship to the louder-sounding parts kept up at maximum.)

This was called "gain-staging."

I just tried a variety of this with my laptop's amp, using iTunes and its equalizer. First, I pushed the iTunes volume all the way up, as well as the volume settings in Windows' volume controls panel - both the main and the "Wav" volume controls. Back in iTunes, I set the EQ to "Flat". Then I cut the preamp slider (the one on the far left side) to minus-3 (or even to minus-12), which is where the compression of the bass seemed to magically just "go away."

Following the sound reinforcement line of thinking, the EQ preamp slider is acting as the "input trim pot" in this case (and I think I can predict it is acting on a digital level, being part of the EQ software tool), while the other volume sliders are being treated "as if" they worked directly on the amplification stages.

With those volume sliders following the EQ all set at full, the digital output on the software level after the EQ is assured to be allowing the full 16-bit "word" to be delivered to the amps and so push the most dynamic range out of them, thus the most current. I am trimming away at that "word" from the EQ's preamp slider, but at least there is no confusion that this is genuinely what is happening, and that I'm working upon the signal in the digital realm there.

You see, I cannot predict how the three "volume" sliders are interacting (the iTunes volume control, Windows' "Wav" volume slider, and Windows' Master volume slider) to affect the amplifier gain - one of them may directly affect analog gain stages while the others "trim" away at the digital word, but which is which. and where, etc?

Now that question isn't important - I'm pushing max digital and analog dynamic range, and thus maximum power through my amplifier stages to my power hungry headphones, and the compression is no longer happening. How nice! Back to music listening now!

Try this for yourself and see what sort of result you get from your system!

Terry 
750prolinebx3.png
Click to expand...
Quote:


tbritton said:
  Quote:

Originally Posted by Happy Camper /img/forum/go_quote.gif
Are there any guidelines or reasons (other than preference) for setting volumes?

I have listed options and would like to read your reasoning for your preference. If you do it differently, share.

Volume out maxxed on all programs. Amp/pre volume control for listening levels.

Source volume varied and turning up the amp/pre volume.

Of course the last option is amp/pre at full volume and control from the pc.

My thoughts are dropping the program volume to 50-80% and turning my amp up to at least 12-2 will give me the most from my components. I adjust this to get my best sound with lowest noise. My reasoning is the amp performance curve would be better at 30-80% output range. Running anything at maximum would not be the best criteria.

I have had so many of these attitudes about this hobby that I have been wrong about. I am again in a learning mode to benefit from my new equipment.

Open to discussion.


I wrote a brief article a while ago:
Optimizing a laptop output for high-current drawing headphones
It is mostly about gain-staging and using an EQ's preamp slider with the EQ at a flat or otherwise setting to control signal levels through the chain so that max power flows through the final amp stages.

In another thread:
Why your awesome IEMs sound bad from your iPod and what YOU can do about it!
The grand-finale in the final pages is mostly about using a program called mp3gain and its sister component, aacgain to control the gain of the music file itself at the decoding level (when it is decompressed from the lossy format) to also control the gain-staging process - before it even reaches the preamp stages!

Neither article received much interest, but there may be something you can use there to support your own experimentation and ideas.

The excellent article by EnOYiN
ASIO4All Explanation
covers using that plugin with foobar and winamp, and other programs ultimately which can use asio, which gets around the issues of bit reduction. However, we should understand that lower volumes are represented by fewer bits being needed in the maximum levels, since we've eliminated those maximum levels, so there is never a need for them, theoretically! But if you reduce the bit-depth, you are reducing the dynamic range of softest to loudest sound, so that's where the trade off occurs and where using the analog volume control keeps the bit-depth and dynamic range, just at quieter levels. Where bit depth becomes most important is on the recording end, for dynamic range, so the very quiet parts like piano fades and reverb tails can exist pure without being dithered with (very quiet) noise to sound natural. It is also important for DSP's so they have more numbers for performing precision calculations, so often a DSP will upconvert to 20 or 24 bit precision to do its thing and then down-size the bit-depth back to 16 bits for passing through the rest of the digital systems.

An old rule-of-thumb with computer audio was to keep the mixer levels at 80% so their gain would never overwhelm the input stages of the final output amplifier. As the first article above discusses, we are now faced with too darn many volume controls - mp3gain/aacgain, EQ preamp gain, kmixer gain, final amp gain. How many volume controls do we adjust? That article suggest adjusting only one of them - the first at the EQ preamp stage. Otherwise, you could only adjust the one in the mp3/aac file itself using mp3gain or aacgain, which would keep bit levels intact as they are going to be, with EQ preamp volume left alone (not using EQ, in other words) and kmixer levels up all the way. On my laptop I'm using a combination of mp3gain/aacgain and the EQ preamp now, with all other kmixer levels all the way up, to get good results from Proline 750s. Try it! Also using the Asio4All plugin with Winamp on my home PC following EnOYiN's directions, and it works flawlessly for using the AC'97 on-board audio card, but truthfully I mostly pipe through my digital Aardvark 24/96 Direct Pro, though not when computer stability issues for other programs become an issue, as its drivers are a little dodgy.

Let us know what your own findings are, please!

Terry 
750prolinebx3.png
Click to expand...
 

So was he incorrect in his thinking or is there some truth/logic to it?
 
Thanks again for your time/input on all this. 
biggrin.gif


 
Jan 5, 2016 at 10:08 PM Post #9 of 18
The iFi iDSD micro seems like a complicated device, I'm not sure what the IEM match setting does, or what exactly he meant when by "Maximum Class A power, before switching to Class B," I guess it switches modes with one of those settings. In general there is not a reason that a position between 12 and 3 o'clock on an amp would be best, but it sounds like there is some specific aspect of the iDSD micro's operation that makes it so.
 
In general on an amp that has both a volume control and a gain switch, you want to use the lowest gain setting, and if you can't get enough volume on that setting, switch to the next one up. As castleofargh said, if there is some specific reason such as distortion or a sensitive or imbalanced volume knob, you can reduce the volume in software if that will fix it. I didn't quite follow the part about gain-staging entirely, but I think this is what that was about.
 
I think gain-staging is relevant in the analog realm, where if you have too much or too little gain at any point in your analog gear, it could add unnecessary noise or distortion. If I'm not mistaken, your software and OS will process audio in 32 bit floating point, meaning you won't lose any precision or cause clipping until the end when it converted to 24 bit integer and sent to the DAC. That means that you can control volume with an EQ preamp slider, your software's volume settings, the windows volume slider, it doesn't really matter which, as long as it's below the clipping point in the end when it is converted to 24 bit int.
 
Jan 9, 2016 at 9:39 AM Post #10 of 18
maconi said:
   
I'm referring to the opposite. Using an negative equalizer preamp (like on a PC) to reduce the source signal gain so the amp has to be driven "harder" to get the signal to the same volume level as it would be without the negative preamp. I was curious as to the science behind it. Some people say they feel like doing so opens everything up, but I'm curious as to if they're just creating pleasing distortions (similar to tube amps) rather than truly "unlocking" any potential in the headphones/amp by putting more power/current through them (if doing so even results in increased power/current, as you say it should still be the same voltage @ the same volume)?
Click to expand...

 
You generally don't want to do this in software, at least not for critical listening and not very much.  Digital volume control in software throws bits away, the lower you go, the more it chucks out.
 
Jan 9, 2016 at 9:49 AM Post #11 of 18
mindsmirror said:
  If I'm not mistaken, your software and OS will process audio in 32 bit floating point, meaning you won't lose any precision or cause clipping until the end when it converted to 24 bit integer and sent to the DAC. That means that you can control volume with an EQ preamp slider, your software's volume settings, the windows volume slider, it doesn't really matter which, as long as it's below the clipping point in the end when it is converted to 24 bit int.
Click to expand...

 
This varies a huge amount from platform to platform and playback software to playback software.
 
Not all OS/HW combinations have great core audio implementations.
 
Jan 9, 2016 at 3:12 PM Post #12 of 18
watchnerd said:
   
You generally don't want to do this in software, at least not for critical listening and not very much.  Digital volume control in software throws bits away, the lower you go, the more it chucks out.
Click to expand...


This really isn't a problem if you're using 24 bit, rather than 16 bit.
 
Jan 9, 2016 at 3:15 PM Post #13 of 18
cjl said:
 
This really isn't a problem if you're using 24 bit, rather than 16 bit.
Click to expand...

 
If you're only doing basic playback, sure.
 
If you're combining it with editing it is.
 
If you're combining it with lots of DSP/DRC, it can be depending upon the FIR/PEQ settings you're using.
 
Jan 10, 2016 at 11:52 AM Post #14 of 18
The "gain staging" way of managing volume makes sense if you have a lot of real-time DSP work going on, especially non-linear DSP work that relies on a known relationship between the digital signal's signal level and the actual output volume.

I'll give a few examples here:


1st example (top): I've set up the Multidynamics dynamic compressor to compress sibilance-related frequencies above a certain level. For this to work consistently, the digital signal levels should relate to the actual volume output by loudspeakers / headphones in a consistent way. E.g. if you turn up the volume knob of the amplifier without re-adjusting the compressor, it lowers the digital signal level in comparison to the same headphone output volume, making the compressor compress less. Vice versa for turning down the volume knob. So what I do is set the amplifier to a fixed level that guarantees ear-splitting volume from the headphones (if not exactly pegging the amplifier to maximum) and adjust the volume via software. But I don't need to fix my music player or system volume, because these plugins I've placed in a system-level VST host downstream of those volume controls. If, instead, your compressor were inserted as e.g. a foobar2000 plugin, it would be upstream of all the volume controls in your system and you would have to fix all the volume sliders / knobs in advance, and adjust volume via a preamp plugin placed in front of the compressor in the DSP chain.

2nd example (lower left): An aggressive EQ for my current CIEMs, which are going out in the treble I'm afraid, but which I still thoroughly enjoy especially out and about. Here, fixed volume knobs a la gain staging isn't as important as simply having plenty of gain downstream and adequate negative preamp--the exact amount of negative preamp isn't important as long as it's negative enough to avoid digital clipping and not negative enough to cause quantization noise or noise in general to be heard from a lifted noise floor. The latter is easy enough to determine--run a clip of dithered silence through your system and crank up the amplifier until you start "hearing" the amplifier's background, or hearing it more than usual.

3rd example (lower right): When running a system with limited dynamic range it can be difficult to trade off between avoiding clipping and raising the noise floor. Crank all the volume knobs to max and you hear a lot of background hissing; turn them down enough and clipping becomes a possibility. One good solution is to put a limiter at the end of your DSP chain and set it up so that it doesn't do anything to the sound when signal levels are ok, but to limit the signal when clipping is happening.
 
HiBy Stay updated on HiBy at their facebook, website or email (icons below). Stay updated on HiBy at their sponsor profile on Head-Fi.
 
https://www.facebook.com/hibycom https://store.hiby.com/ service@hiby.com
Jan 10, 2016 at 12:39 PM Post #15 of 18
off topic super hidden so the modo won't see it:
I find vsthost to lag visually(audio is fine when I let it be) as soon as I start using a few stuff, or simply changing my EQ with some serious FIR settings( like when I play with left and right separately). but as I don't get 99.9% of the settings of vsthost, I was wondering if maybe it was my fault(between that and VAC, I've been randomly doing stuff until it worked ^_^).  is it a smooth experience for you?
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top