Sonarworks Headphone Calibration software

[BIG POPPA]

@fjrabon you just confused a ton of peeps. I get your point. Try not to get that winded. With tubes it is a safety net. Yes I am over simplifying

 

[phoenixdogfan]

That's one of the reason's why I run my IFI IDSD Micro in Turbo mode when running The AB0000 correction curve with my HD800's. BTW, no issues whatsoever with clipping, digital or otherwise.

 

[fjrabon]

@fjrabon
you just confused a ton of peeps. I get your point. Try not to get that winded. With tubes it is a safety net. Yes I am over simplifying

How is it a safety net? Your tubes don't care if there's digital distortion or not. To them, signal is signal. I don't think anybody seems particularly confused. I think saying "only use clipping protection if you're pushing your amp" like you did is confusing since that's exactly backwards.

 

[phoenixdogfan]

For RudeWolf:
 
I've been looking over my control panel and I've noticed that SW has the ability to enable your EQ'd phones to simulate the sonic profile of other phones.  Only thing is, it's limited to only four phones: Beats Studio, Sennheiser HD 650,  AKG 712, and Grado GS1000's.
 
SW undoubtedly has the sonic profiles of dozens of headphones and could, in all probability, convincingly simulate the sound of every one of them--especially if the Sonarworks user was in the possession of an extremely capable phone that you've already published the EQ profile of. 
 
I think it might be very interesting for your audiophile clientele to be able to dial up the sound of the Audeze LCD 2's, for example, on their HD 800's or the Oppo PM 3's on their Beyerdynamic T1's.  I realize this would be a premium feature that not everyone would be interested in, but research by Sean Olive seems to point to the conclusion that if one headphone's FR is precisely matched to another, it will be sonically indistinguishable from the phone it's matched to. I have a pair of ultra premium phones, and I think this feature would add a lot of value to your product, and I know I would pay extra for it.  So have you ever explored the expansion of your list of simulated components?

 

[MattTCG]

Can someone chime in with impressions on Sonarworks with the pm-3?
 
thanks! 

 

[fjrabon]

research by Sean Olive seems to point to the conclusion that if one headphone's FR is precisely matched to another, it will be sonically indistinguishable from the phone it's matched to. 

 
I do agree with you that it would be cool if you had more options in the simulate menu, for sure.  
 
However, one thing sonarworks has pointed out to me is that frequency response isn't everything, or for me personally, even the primary thing.  For example, when I have my HD800 and my SRH840 both 100% wet on sonarworks, they sound very different, although they should at that point have basically the same exact frequency response.  Yes, they sound more similar than without sonarworks. But they are still really quite easily noticeably different.  The differences in transient response, transparency and soundstage are immediately recognizable.  And just so that it isn't open vs. closed, I've heard the HD650 through sonarworks, and again, the HD800's superiority in speed, soundstage and detail are certainly there even after the frequency responses are matched, even if it's less obvious than between the SRH840.  
 
I like the simulate as a cheap way to get flavors.  And I actually think if you have a HD800 it's cool because it allows you to have those different flavors while still keeping the technical superiority of the HD800.  Want a HD650's smooth sound, but hesitant to give up the HD800's soundstage and transparency?  Hit that simulate button and you got it.  Want that exciting color of the GS1000, but don't want its distortion in the bass?  you can have it.  Want to hear how crappy NS10s sound?  Just push that button dog.  
 
(side note, it always amazes me that home consumers buy NS10s.  The whole point of NS10s is that they sound roughly like the most average consumer system, particularly an average car stereo, and thus allows producers and engineers to hear how the music will sound as most consumers listen to it)

 

[castleofargh]

  For RudeWolf:
 
I've been looking over my control panel and I've noticed that SW has the ability to enable your EQ'd phones to simulate the sonic profile of other phones.  Only thing is, it's limited to only four phones: Beats Studio, Sennheiser HD 650,  AKG 712, and Grado GS1000's.
 
SW undoubtedly has the sonic profiles of dozens of headphones and could, in all probability, convincingly simulate the sound of every one of them--especially if the Sonarworks user was in the possession of an extremely capable phone that you've already published the EQ profile of. 
 
I think it might be very interesting for your audiophile clientele to be able to dial up the sound of the Audeze LCD 2's, for example, on their HD 800's or the Oppo PM 3's on their Beyerdynamic T1's.  I realize this would be a premium feature that not everyone would be interested in, but research by Sean Olive seems to point to the conclusion that if one headphone's FR is precisely matched to another, it will be sonically indistinguishable from the phone it's matched to. I have a pair of ultra premium phones, and I think this feature would add a lot of value to your product, and I know I would pay extra for it.  So have you ever explored the expansion of your list of simulated components?

bold part:  that's really not what they intended to show or research. they tested with different headphones, and in other tests they went for one headphone EQed to different signatures, and people said what they preferred. at no moment did they have to try and tell headphones apart. if I've forgotten a paper where they mention it, and you're right, I'm sorry in advance, but it would surprise me if I had forgotten such an important point was proved.
 
it might come very very close in several cases but the 2 headphones would already need to have exemplary specs, not get to audible distortion levels after the EQ is applied, and they also must have controlled enough resonant frequencies because else simply applying EQ isn't guaranteed to be able to deal with it.
I'm fairly confident that EQ can improve about all the headphones ever made, but mimicking other headphones is another matter. the results will really depend on the 2 headphones picked and the frequency range we're trying to replicate.
 

  @fjrabon you just confused a ton of peeps. I get your point. Try not to get that winded. With tubes it is a safety net. Yes I am over simplifying

he really didn't IMO. tube amps won't hard clip themselves contrary to SS amps, so the clipping will occur but won't sound remotely as bad. it's still clipping and it's still not a good idea. but more important, that's only dealing with clipping of the amp. this has little to do with digital clipping.
to prevent clipping you can lower the digital gain at a value to make sure nothing will ever reach 0db or above(going crazy low would reduce the dynamic so not an ideal choice). or you can use a limiter, that will do nothing when all is good and digital headroom is big, but will compress the top to avoid or reduce the amount of clipping and attenuate the damage done to the signal. just letting things clip without a care because it doesn't sound too bad on a tube amp is your choice of course, but let's not pretend that it's the better choice.

 

[RudeWolf]

  [...] So have you ever explored the expansion of your list of simulated components?

 
Sort of. We could go all out and simulate every headphone we've measured, but so far it's limited to choices interesting for engineers who do mix double checking.
 
 
 

   
[...] (side note, it always amazes me that home consumers buy NS10s.  The whole point of NS10s is that they sound roughly like the most average consumer system, particularly an average car stereo, and thus allows producers and engineers to hear how the music will sound as most consumers listen to it)

 
Initially the NS-10 was chosen because it's a limited bandwidth sealed speaker. You get a very decent impulse response and the lack of low end means that it won't introduce masking in critical mid-frequencies. It was common practice to tweak the tweeters, because stock they sounded unbearably bright.
 
Because some knowledgeable engineers used the NS-10's, a myth was created that they're some magic device which right away tells you if your mixes suck. As they say - if it sounds good on the NS-10, then it sounds good anywhere. What's funny - this has been extended to headphones as well, so if they're unbearably bright, then surely they must magically make mixing mistakes more apparent.

 

[Yoga]

...

 
As much as I adore this plugin (and constantly—deservedly—big it up on this forum), for the love of all things holy, please add an auto makeup gain when using clipping.
 
Incredibly simple to do, and would make using this *effortless* in a production environment. As it stands, it's an absolute workflow pain having to gain match different audio sources and apps in various scenarios.

System wide is A) unnecessary and B) not possible when using a DAW.
 
I'm constantly having to compensate the gain manually in Ableton also. Sonarworks is never without a companion gain utility.
 
A fundamental bit of functionality in this day and age. Adding a simple auto makeup gain would perfect this plugin and completely optimise (remove!) the workflow required to use it.
 
Peace!

 

[DivineCurrent]

   
I do agree with you that it would be cool if you had more options in the simulate menu, for sure.  
 
However, one thing sonarworks has pointed out to me is that frequency response isn't everything, or for me personally, even the primary thing.  For example, when I have my HD800 and my SRH840 both 100% wet on sonarworks, they sound very different, although they should at that point have basically the same exact frequency response.  Yes, they sound more similar than without sonarworks. But they are still really quite easily noticeably different.  The differences in transient response, transparency and soundstage are immediately recognizable.  And just so that it isn't open vs. closed, I've heard the HD650 through sonarworks, and again, the HD800's superiority in speed, soundstage and detail are certainly there even after the frequency responses are matched, even if it's less obvious than between the SRH840.  

I have the same experience with my HD650 and DT880 using Sonarworks; they have a very similar frequency response, but the DT880 still has a hazy treble texture.
 
However, the main reason, I believe, they sound different is because I purchased a pre-calibrated pair of HD650s from Sonarworks, which gives you the most accurate calibration within +/- 1 dB, instead of +/- 3 dB due to production variation. If I had also gotten the individually calibrated pair of DT880s from them, I'm almost positive the frequency response would be so similar, they would be nearly indistinguishable. The main differences like you said, would most likely be soundstage, imaging, and limitations of the drivers themselves (ringing, decay, harmonic distortion). You may not think 3 dB of variation makes a difference, but it really does, especially in the mids and treble. If there is anyone here who has gotten 2 or more individual calibrations done for their headphones, please chime in. Or @RudeWolf if you can confirm if I am correct in assuming this.
 
I also definitely notice a difference in the low end bass distortion with my HD650s compared to the DT880s, as the DT880 have a little less of it and therefore sound tighter. So, distortion plays a big role too. 

 

[fjrabon]

I'm almost positive the frequency response would be so similar, they would be nearly indistinguishable. 

I agreed with everything else you wrote, but just not this.  At all.  There are too many other easily audible aspects to a headphone's sound to say "nearly indistinguishable" or even close.  Now, for mixing purposes, the ways in which they would then differ aren't all that important.  Which was the original purpose of SonarWorks anyway, so people could get consistent mixes on different gear.  While two different headphones both individually calibrated would have, functionally, the same frequency response, they wouldn't, for example, have the same impulse response.  Impulse response is something people can very readily hear, even though they often mistakenly attribute it to frequency response.  For whatever reason, people on here like to attribute a quick decay with a sharp response to "being bright."  It's like head-fiers must have everything converted into terms of frequency response, even for things that have nothing to do with frequency response.  Same for how the distortion breaks down within a given amount of THD.  Sonarworks won't do anything directly about the fact that a HD800s has more 2nd harmonic distortion than the HD800, relative to the 3rd and 4th harmonics.  For whatever reason head-fiers always talk about a headphone with more distortion in the 4th harmonic as though it had elevated treble, and more distortion in the 2nd harmonic as if it had more bass.  I get that is how they're often perceived, but understanding the difference is important, because then you understand how SonarWorks won't address most of the differences between the HD800 and HD800s, because it doesn't address differences in how the distortion profiles.  Note here that I'm not even talking about *amount* of distortion (measured as THD) but rather what harmonics that distortion goes into within a given amount of THD.
 
For one more example, every moderately decent amp ever made has the same ***frequency response*** to within a tenth of a decibel.  Yet, I think we all agree, people can pretty readily discern different amps.  There's much more to it than frequency response, and even with the same frequency response, different headphones are often nowhere close to "nearly indistinguishable"
 
As far as i know, nobody who has done research has ever claimed that (the research you cited earlier certainly doesn't).
 
The distortion you're noticing in the HD650 vs the DT880 has more to do with the fact that the HD650 has to have its bass boosted more, and thus there is greater distortion after SonarWorks, it's not because the HD650 has more distorted bass to begin with.  SonarWorks can put fairly significant demands on a headphone if you get aggressive with the wet mix and the headphone isn't a great technical performer.  
 
(I get that some parts of the above were fairly technically dense, and most people can just ignore it, but unfortunately there is no way around that if we are talking about the ability of one transducer to mimic another SOLELY via EQ).

 

[RudeWolf]

@fjrabon makes a really good point. Tonal correction can't address everything about what makes headphones sound like they do. But it addresses the most audible and currently worst performing aspect of headphones. We have successfully done phase correction for headphones and might implement it further down the road, but the funny thing is the effect wasn't something we'd readily call "a better sound". Hence more research and so on...
 
Same goes for harmonic distortion; it can introduce psychoacoustic effects like "more bass", "richer mids", "more resolution". I know many of you are aware that there are some rogue calibration profiles floating around for HD800/650/600 with no-limits correction. There's a reason why these are not officially supported and bundled with the plug-in. First of all - they require a heroic amount of analog gain for compensation. Secondly - calibration of such magnitude at low frequencies will introduce large amounts of THD. I'm fairly positive that most of the reports that these curves give more bass are from people who have actually heard 20Hz being supplemented by a lot of 40Hz and 60Hz rumble.
 
There is also a bright side to tonal correction from THD perspective. Any subtractive correction will reduce THD for the respective region. Resonant FR peaks usually are accompanied by harmonics and by reducing incoming signal, we also reduce the crud.
 
Hence - we can improve many if not all headphones, but we can't turn everything into an HD800 with HD650 tonal profile.

 

[514077]

I'd love to try this now that I have Windows 7.  I already blew my non-functional trial on XP.  Having trouble buying the download, can't figure out why.
How can this be fixed?  And as a Canadian, do I use the world or US site?
Thanks

 

[MattTCG]

Okay, I've got a weird issue with SW. For the past 2-3 days I'm getting some pops/clicks every few seconds ONLY when using Sonarworks. Using musicbee or any other player the noise is gone. I use SW with jriver and I've never had a single problem. 
 
Any suggestions? 

 

[LajostheHun]

For one more example, every moderately decent amp ever made has the same ***frequency response*** to within a tenth of a decibel. Yet, I think we all agree, people can pretty readily discern different amps.

Must have missed the empirical data[not anecdotes] on that. Can you provide it?
Having said that amps can "react" differently to different loads, but that's not intentional it's more of a" side effect" and it's audibility is still extremely debatable even subjectively.