The No.1 Website for Pro Audio
Goodhertz CanOpener Studio v3
Old 7th April 2020
  #361
Lives for gear
 
🎧 5 years
Thanx ! I tried this and indeed it's totally flat.

Nevertheless i notice i really don't like how angle 0° sounds on my HD 650 :o

The angle give more depth to the music and bring more clarity. But i notice with this test i much prefer 45° rather than 60°. 60 is a bit too wide for my taste.
Old 7th April 2020 | Show parent
  #362
Lives for gear
 
blaugruen7's Avatar
 
🎧 15 years
Angle == doesnt make much sense for me. then you can almost press bypass

canopener is great. i use it always.
Old 17th June 2020
  #363
Lives for gear
 
🎧 5 years
I notice when I reduce the output of CanOpener it's changing the frequency response. Is it something normal ??

Screenshot :

0 db : https://ibb.co/vVBy9HB

- 10 db : https://ibb.co/rFnVz3R

Edit : Okay sorry I understood, i need to put the compensation to 0 % if I want to keep the flat frequency response.
Old 26th June 2020
  #364
Here for the gear
 
In the image in a Sonarworks blog post, you can see in the Master channel image, that they are using Reference 4 before CanOpener

Post: https://www.sonarworks.com/blog/lear...with-canopener
Old 18th July 2020
  #365
Gear Maniac
 
Verified Member
I'm just download the 15 day demo. My first impression: Very nice but the low end disappear, without the CanOpener the sub-low-end sounds booming in my AKG 702. My mix suddenly sounds so nice. The low end is way better when this plugin is on ... but is it true, am I fooling my self?
Is the low end more realistic with or without this plugin?
Old 5th August 2020 | Show parent
  #366
Gear Nut
 
Bongo Spank's Avatar
 
...

Last edited by Bongo Spank; 5th August 2020 at 06:28 PM.. Reason: dbl post
Old 5th August 2020 | Show parent
  #367
Gear Nut
 
Bongo Spank's Avatar
 
Correct me if I'm wrong, but:

It is my understanding that each Sonarworks SPEAKER calibration curve has one singular stated purpose:

1) Achieve a flat response curve by negating any derivations from that ideal.

It is further my understanding that the headphone curves do this as well, but then also do the following:

2) Bend the flattened curve to a proprietary Harman-esque reference designed to approximate the perception of a flat monitors in a tuned studio environment.

While both these ends are achieved in the case of headphones by applying a single curve, it is easier for the sake of discussion to refer them as separate curves being applied since they could just as easily have been applied as separate curves, their objectives are distinct, and only the second relates directly for the purposes of this discussion to Can Opener.

Once the single Sonarworks headphone curve is applied, BOTH objectives are already applied. Tonality is fully dialed, headphones are ALREADY adjusted to match perception of flat monitors in a tuned studio, and any futher EQ is NOT wanted.

So if you want crosstalk, just add Can Opener, right?

Except:

Quote:
Originally Posted by devinkerr ➡️
CanOpener was tuned in our mastering room, and on decently-flat headphones, it sounds (and measures) remarkably similar to high-quality speakers in a highly treated room.
Ignoring that "decently flat" is precisely the kind of vaguery Sonarworks users have spent time, money and effort to eliminate...

Since the quote is out of context, I'll clarify that it's my understanding that the dev is referring to imposing an EQ curve in addition to the crosstalk for the purpose of bridging the gap between flat monitors in a tuned studio and a matched perception on headphones. Which would be fine... if Sonarworks users hadn't ALREADY applied another curve for that same purpose since it's baked into the correction curves for headphones.

The entire POINT of using Sonarworks in the first place is to leave nothing to chance by doing the most accurate possible system specific tuning to bring the monitoring as close as possible to absolute reference curves. The monitoring chain is, after all, the absolute last place anyone should ever be putting anything that changes the tonality in ways that are not very deliberate and measured as it's the most problematic spot for the one thing that ultimately matters: How your mixes translate.

Imagine for a moment that you have a very specific problem with your guitar track where you identify that you need a boost of precisely 3db at precisely 1500hz. No problem. Add EQ, and done. What if I came in behind you every time you did that, though, and added ANOTHER eq on top of that that you didn't need or want.

Sure, it's irritating and counterproductive, but how much moreso if the only remedy I gave you was to add MORE processing on top of the thing you didn't want in the first place to undo it. Oh... and I'm not going to tell you how much EQ I applied, or at what frequencies, so you're going to have to guess.

If that doesn't sound bonkers, then read it again... and if it sounds like a terrible idea on a guitar track, consider that the same concept applied to your monitoring chain screws up not just your guitar tone in ways you CAN hear, but absolutely everything you put out in ways you CAN'T hear.

Quote:
Originally Posted by devinkerr ➡️
Generally, this is a desirable feature unless the headphone manufacturer has already emphasized the bass...
Or unless your ALREADY imposing curve with the same stated purpose... which is the case for literally EVERY Sonarworks user. For them, and anyone in a similar situation, it's an objectively undesirable feature.... like dealbreaker undesirable.

Quote:
Originally Posted by devinkerr ➡️
.... which is why we provide a super transparent shelving EQ when compensation is needed.
To do what exactly? Cut by an unspecified amount at an unspecified frequency? The only thing crazier than putting a curve you don't want or need on top of already painstakingly dialed tonality is doing so, then also trying to UNDO it with vague and unspecified additional processing. We are still talking about audio engineers making critical decisions on calibrated systems, right... not kindergarteners trying to mix crayons to make brown? (or more accurately being forced to make brown, then being told to add even more colors to make it UN-brown.)

I ask because many Sonarworks users are professionals who rely on the accuracy of their monitoring to make a living. They spent money for a calibrated microphone, shipping off headphones, etc. They spent the time and effort to run the diagnostics, and get their room and headphones PERFECTLY dialed with every single nudge having a precisely measured and clearly intended purpose leading to the optimal tonal balance... and now they're just going to toss an vague and unspecified EQ curve on top of that and blindly try to undo the thing they never wanted in the first place?

It's not like this hasn't been mentioned for years now, so why is there no way to simply disable the shift in tonality when AFAIK, a substantial percentage of users are also using Sonarworks?

Quote:
Originally Posted by devinkerr ➡️
We could make it flatter, but it would be less accurate.
... For NON Sonarworks users. It seems to me that this is missing the point entirely, and that what Sonarworks users need is not "flatter", but rather complete defeat of ANY tonality shift since it's entirely redundant.

Am I missing something, or should Sonarworks users just not be using this plug?
Old 18th August 2020 | Show parent
  #368
Gear Head
 
🎧 10 years
I don't think you tried Can Opener.

1st, it doesn't color the sound that much and to me it clearly doesn't try to. I don't see it trying to apply any Harman curve on top of Sonarworks. As far as this kind of plugins go, this one is as transparent as it gets. It looks like the slight increase in bottom end is a by-product of the crosstalk process on itself, as it aglutinates the lower frequencies in the center of the stereo image more than it does for the high frequencies. It's probably not easy to apply such a dramatic change in sound stage without some compromise.

2nd, the shelving in the plugin is not an "unspecified ammount at an unspecified frequency". It is totally adjustable. It defaults the center frequency of the low shelf at 120hz, but you can freely change it, and you also can select how many db of attenuation you want (in 0.5db steps, so not perfect).
Old 20th August 2020 | Show parent
  #369
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by Gonira ➡️
I don't think you tried Can Opener.

1st, it doesn't color the sound that much and to me it clearly doesn't try to. I don't see it trying to apply any Harman curve on top of Sonarworks. As far as this kind of plugins go, this one is as transparent as it gets. It looks like the slight increase in bottom end is a by-product of the crosstalk process on itself, as it aglutinates the lower frequencies in the center of the stereo image more than it does for the high frequencies. It's probably not easy to apply such a dramatic change in sound stage without some compromise.

2nd, the shelving in the plugin is not an "unspecified ammount at an unspecified frequency". It is totally adjustable. It defaults the center frequency of the low shelf at 120hz, but you can freely change it, and you also can select how many db of attenuation you want (in 0.5db steps, so not perfect).
Except that it definitely is changing tonality.

It's frankly hard to miss.

The dev has stated this is being added on purpose, and is a separate consideration from the mere mechanics of mixing signal from the two channels.

I've confirmed this with various types of signals as input. There just absolutely positively IS eq being added. And yes, it IS of unspecified frequency and amount.

It's entirely irrelevant how accurate the tool to fix something is if it's unclear how it's broken.

More to the point, it shouldn't need to be fixed in the first place.

Again, the tonal shift may be desired if you're NOT using Sonarworks. If you are, it is entirely redundant and not only unwanted, but specifically counterproductive. The CONCEPT behind the shift is entirely valid... but that's why Sonarworks already did it on their end.

There's really no valid reason that accuracy in the monitoring chain should even be a controversial topic.
Old 20th August 2020 | Show parent
  #370
Gear Head
 
🎧 10 years
Quote:
Originally Posted by Bongo Spank ➡️
Except that it definitely is changing tonality.

It's frankly hard to miss.

The dev has stated this is being added on purpose, and is a separate consideration from the mere mechanics of mixing signal from the two channels.

I've confirmed this with various types of signals as input. There just absolutely positively IS eq being added. And yes, it IS of unspecified frequency and amount.

It's entirely irrelevant how accurate the tool to fix something is if it's unclear how it's broken.

More to the point, it shouldn't need to be fixed in the first place.

Again, the tonal shift may be desired if you're NOT using Sonarworks. If you are, it is entirely redundant and not only unwanted, but specifically counterproductive. The CONCEPT behind the shift is entirely valid... but that's why Sonarworks already did it on their end.

There's really no valid reason that accuracy in the monitoring chain should even be a controversial topic.
I never said there is no tonal change, I only said it's slight and it isn't something intentionally added but a natural side effect of what this plugin does (crossfeed).

If I happen to be wrong, I completely misunderstood the dev's explanation.

And I definitely misunderstood you when you said about "unspecified frequency, etc", I thought you were talking about the effect of the shelving filter inside the software, not the initial tonal shift. Alright then.
Old 20th August 2020 | Show parent
  #371
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by Gonira ➡️
I never said there is no tonal change, I only said it's slight and it isn't something intentionally added but a natural side effect of what this plugin does (crossfeed).

If I happen to be wrong, I completely misunderstood the dev's explanation.

And I definitely misunderstood you when you said about "unspecified frequency, etc", I thought you were talking about the effect of the shelving filter inside the software, not the initial tonal shift. Alright then.
It is added intentionally... but by both Can Opener AND Sonarworks.

Either by itself is probably a good thing. I take no issue with the concept. I rather like it... but stacking two corrective EQ's on top of each other for the same purpose makes no more sense than stacking two crossfeed algos on top of each other. Actually, it's quite a bit worse since the crossfeed alone if screwed up would likely only cause panning issues.

The tonal shift was actually the very first thing I noticed about the plug. A quick search of discussion on the topic showed I wasn't alone. If I weren't using Sonarworks, I'd probably be happy about it, but I am, so it's just a problem... and one that a great many users likely share.... all of which could be easily fixed by adding a defeat for those already using other software that does a similar tonal shift making this one redundant.

I just want crosstalk. Most of the other GoodHertz plugs are well thought out, and I've probably turned on dozens of users to them, but this is something that should be fixed. As is, it's effectively a bug for Sonarworks users.
Old 23rd January 2021 | Show parent
  #372
Gear Addict
 
devinkerr's Avatar
 
🎧 10 years
Quote:
Originally Posted by Bongo Spank ➡️
It is added intentionally... but by both Can Opener AND Sonarworks.

Either by itself is probably a good thing. I take no issue with the concept. I rather like it... but stacking two corrective EQ's on top of each other for the same purpose makes no more sense than stacking two crossfeed algos on top of each other. Actually, it's quite a bit worse since the crossfeed alone if screwed up would likely only cause panning issues.

The tonal shift was actually the very first thing I noticed about the plug. A quick search of discussion on the topic showed I wasn't alone. If I weren't using Sonarworks, I'd probably be happy about it, but I am, so it's just a problem... and one that a great many users likely share.... all of which could be easily fixed by adding a defeat for those already using other software that does a similar tonal shift making this one redundant.

I just want crosstalk. Most of the other GoodHertz plugs are well thought out, and I've probably turned on dozens of users to them, but this is something that should be fixed. As is, it's effectively a bug for Sonarworks users.
Hi there, I know this post is pretty old, but we’ve gotten some support messages related to it, so I thought it’d be good to respond on the forum for anyone else with similar questions. There seems to be confusion about what exactly CanOpener (and Sonarworks) does, how crossfeed works, and the type of “coloration” that CanOpener might introduce. I’ll try to address all of those, and if anyone has any other questions, please feel free to ask.

While Sonarworks aims to correct or flatten the frequency responses of different makes & models of headphones, CanOpener does not, currently. Instead, CanOpener uses crossfeed to make the spatial response (of any set of headphones) as similar to working on stereo loudspeakers as possible.

CanOpener’s crossfeed is not a corrective EQ — we don’t artificially “color” or “correct” the frequency response. You can use CanOpener’s crossfeed with Sonarworks EQ correction without issue (& many mixing and mastering engineers do) — they’re complimentary, not overlapping. Assuming your headphones are pretty good, I personally find crossfeed to be more important than EQ correction for headphone mixing. With HD650’s or just about any Audeze headphones + crossfeed, it feels incredibly similar to working on the ATC’s in our mastering room, and my results are nearly identical — which is wild! The fact that one plugin & some $500 headphones can produce similar results to speakers and room treatment that cost more than 50x that amount is a testament to how good headphones and DSP have gotten.

Crossfeed, while a simple concept, is quite difficult to implement well. It’s not as a simple as just adding “crosstalk” between the left/right channels. That would just be equivalent to a “stereo width narrower” and would sound nothing like real speakers. If you want that, though, I have a Midside Matrix plugin to sell you (it’s free!)

On the other hand, if you try to go a more complex route and directly use spatial ear measurements or HRTF’s (like Waves NX) you’ll be relying on a set of gnarly filters that: a) only work for certain ear/head sizes/shapes b) drastically change the frequency response & introduce comb filtering, and c) are actually unnecessary for faithfully recreating the stereo field of studio monitors.

The approach we take is to do as little as is needed to accurately represent the stereo field on headphones. We model the crossfeed head shadowing, interaural time delays, etc, but we do it in a bunch of fancy ways that are more transparent and minimize the impact on the frequency/phase response.

CanOpener’s crossfeed introduces essentially the same coloration that would occur naturally with loudspeakers and is just inherent to crossfeed. I hesitate to even call it coloration because it’s just what having crossfeed at all sounds like, and there’s no “defeat” we could offer for this. Although there is not actually more bass in the frequency response of CanOpener, crossfeed can enhance the perception of bass. In previous versions of CanOpener there was actually some added low end (for certain types of settings/signals) as a result of the crossfeed algorithm — I think that’s where most of the quotes you’ve pulled in the previous posts come from — but we’ve tweaked the algorithm over the years to be even more transparent and to avoid this altogether. Here’s a quote about it from the CanOpener v3.5 blog post:
“In v3.5, we’ve improved the overall flatness and loudness-matching a bit in the crossfeed algorithm: CanOpener now produces consistent loudness and frequency response for any crossfeed setting. Crossfeed does naturally enhance bass perception — in fact, that’s one of the main reasons to use crossfeed in the first place — but we’ve improved the algorithms so that mono elements, in particular, will have as little coloration as possible.”
Here’s the frequency response of CanOpener v3.5 on the “Mix Engineer, Flat” preset, for example:



If you’re not using v3.5 you should definitely update (it’s a free upgrade).

One thing to be aware of is that some headphone manufacturers try to compensate for the lack of crossfeed in headphones by increasing the bass response (...Harman👀). This is an unfortunate trend in headphone design because it doesn’t address the root cause of the discrepancy between headphones and speakers, and it just decreases overall headphone fidelity/transparency. I don’t know how Sonarworks calibrates their headphone measurement, and it’s possible that they are also trying to achieve some non-flat “ideal” headphone response. Headphone measurement itself is an art, the calibration matters a lot, and not all headphone measurements are equally good.

A plea to headphone makers: please, please don’t do this — at least not for “pro” headphones. Just make the headphones as close to flat as possible and leave any post processing or deviation from flat up to the user or developers who have spent their entire careers working on crossfeed/headphone algorithms. At Goodhertz, we’ve gotten to work on several licensing/consulting projects with headphone manufacturers, and it’s a little alarming how often they are inclined to intentionally boost areas of the frequency response when crossfeed would address their perceived issues more directly.

OK! This is a longer post than I’d imagined, but I hope that it at least explains our approach with CanOpener and what it does/does not do. CanOpener is really just a tool to make working on headphones easier (& more like speakers). Many of our users find it indispensable, and it’s been one of our most popular plugins for years. If you’re not one of those people, that’s totally okay!
Old 23rd January 2021 | Show parent
  #373
Lives for gear
 
blaugruen7's Avatar
 
🎧 15 years
Thank you for your Input.
Very much appreciated.
Pls state again what is your official perspective.
Where should canopener be placed on the masterbus?
Before or after Sonarworks?
Old 23rd January 2021 | Show parent
  #374
Gear Addict
 
devinkerr's Avatar
 
🎧 10 years
Quote:
Originally Posted by Gonira ➡️
I don't think you tried Can Opener.

1st, it doesn't color the sound that much and to me it clearly doesn't try to. I don't see it trying to apply any Harman curve on top of Sonarworks. As far as this kind of plugins go, this one is as transparent as it gets. It looks like the slight increase in bottom end is a by-product of the crosstalk process on itself, as it aglutinates the lower frequencies in the center of the stereo image more than it does for the high frequencies. It's probably not easy to apply such a dramatic change in sound stage without some compromise.

2nd, the shelving in the plugin is not an "unspecified ammount at an unspecified frequency". It is totally adjustable. It defaults the center frequency of the low shelf at 120hz, but you can freely change it, and you also can select how many db of attenuation you want (in 0.5db steps, so not perfect).
This post is spot on — thank you for posting! — except for one tiny thing (literally): in CanOpener you can actually adjust the EQ in 0.1 dB steps, you just have to hold the Command key (on Mac) or Control (on Windows) for fine control. This is true for the sliders in all Goodhertz plugins. Conversely, holding Shift will move the control in coarse steps. You can also use the arrow keys or just type values into the sliders as well.
Old 23rd January 2021 | Show parent
  #375
Gear Addict
 
devinkerr's Avatar
 
🎧 10 years
Quote:
Originally Posted by blaugruen7 ➡️
Thank you for your Input.
Very much appreciated.
Pls state again what is your official perspective.
Where should canopener be placed on the masterbus?
Before or after Sonarworks?
Short answer: We recommend using CanOpener last on the 2-bus.


Longer answer: CanOpener and Sonarworks are both effectively linear processors, so it doesn't much matter which order they process in. One exception is that CanOpener includes our mastering-grade dither that can provide increased fidelity when used with 24-bit DAC's. If you intend to use the dither (which you can turn on in the Advanced pane), you must put CanOpener last in the processing chain for it to work properly.
Old 23rd January 2021 | Show parent
  #376
Lives for gear
 
blaugruen7's Avatar
 
🎧 15 years
Quote:
Originally Posted by devinkerr ➡️
Short answer: We recommend using CanOpener last on the 2-bus.


Longer answer: CanOpener and Sonarworks are both effectively linear processors, so it doesn't much matter which order they process in. One exception is that CanOpener includes our mastering-grade dither that can provide increased fidelity when used with 24-bit DAC's. If you intend to use the dither (which you can turn on in the Advanced pane), you must put CanOpener last in the processing chain for it to work properly.
Thank you very much.
I did tests recently and in in this order it sounded more "right"
Old 23rd January 2021 | Show parent
  #377
Gear Nut
 
Quote:
Originally Posted by devinkerr ➡️
Hi there, I know this post is pretty old, but we’ve gotten some support messages related to it, so I thought it’d be good to respond on the forum for anyone else with similar questions. There seems to be confusion about what exactly CanOpener (and Sonarworks) does, how crossfeed works, and the type of “coloration” that CanOpener might introduce. I’ll try to address all of those, and if anyone has any other questions, please feel free to ask.

While Sonarworks aims to correct or flatten the frequency responses of different makes & models of headphones, CanOpener does not, currently. Instead, CanOpener uses crossfeed to make the spatial response (of any set of headphones) as similar to working on stereo loudspeakers as possible.

CanOpener’s crossfeed is not a corrective EQ — we don’t artificially “color” or “correct” the frequency response. You can use CanOpener’s crossfeed with Sonarworks EQ correction without issue (& many mixing and mastering engineers do) — they’re complimentary, not overlapping. Assuming your headphones are pretty good, I personally find crossfeed to be more important than EQ correction for headphone mixing. With HD650’s or just about any Audeze headphones + crossfeed, it feels incredibly similar to working on the ATC’s in our mastering room, and my results are nearly identical — which is wild! The fact that one plugin & some $500 headphones can produce similar results to speakers and room treatment that cost more than 50x that amount is a testament to how good headphones and DSP have gotten.

Crossfeed, while a simple concept, is quite difficult to implement well. It’s not as a simple as just adding “crosstalk” between the left/right channels. That would just be equivalent to a “stereo width narrower” and would sound nothing like real speakers. If you want that, though, I have a Midside Matrix plugin to sell you (it’s free!)

On the other hand, if you try to go a more complex route and directly use spatial ear measurements or HRTF’s (like Waves NX) you’ll be relying on a set of gnarly filters that: a) only work for certain ear/head sizes/shapes b) drastically change the frequency response & introduce comb filtering, and c) are actually unnecessary for faithfully recreating the stereo field of studio monitors.

The approach we take is to do as little as is needed to accurately represent the stereo field on headphones. We model the crossfeed head shadowing, interaural time delays, etc, but we do it in a bunch of fancy ways that are more transparent and minimize the impact on the frequency/phase response.

CanOpener’s crossfeed introduces essentially the same coloration that would occur naturally with loudspeakers and is just inherent to crossfeed. I hesitate to even call it coloration because it’s just what having crossfeed at all sounds like, and there’s no “defeat” we could offer for this. Although there is not actually more bass in the frequency response of CanOpener, crossfeed can enhance the perception of bass. In previous versions of CanOpener there was actually some added low end (for certain types of settings/signals) as a result of the crossfeed algorithm — I think that’s where most of the quotes you’ve pulled in the previous posts come from — but we’ve tweaked the algorithm over the years to be even more transparent and to avoid this altogether. Here’s a quote about it from the CanOpener v3.5 blog post:
“In v3.5, we’ve improved the overall flatness and loudness-matching a bit in the crossfeed algorithm: CanOpener now produces consistent loudness and frequency response for any crossfeed setting. Crossfeed does naturally enhance bass perception — in fact, that’s one of the main reasons to use crossfeed in the first place — but we’ve improved the algorithms so that mono elements, in particular, will have as little coloration as possible.”
Here’s the frequency response of CanOpener v3.5 on the “Mix Engineer, Flat” preset, for example:



If you’re not using v3.5 you should definitely update (it’s a free upgrade).

One thing to be aware of is that some headphone manufacturers try to compensate for the lack of crossfeed in headphones by increasing the bass response (...Harman👀). This is an unfortunate trend in headphone design because it doesn’t address the root cause of the discrepancy between headphones and speakers, and it just decreases overall headphone fidelity/transparency. I don’t know how Sonarworks calibrates their headphone measurement, and it’s possible that they are also trying to achieve some non-flat “ideal” headphone response. Headphone measurement itself is an art, the calibration matters a lot, and not all headphone measurements are equally good.

A plea to headphone makers: please, please don’t do this — at least not for “pro” headphones. Just make the headphones as close to flat as possible and leave any post processing or deviation from flat up to the user or developers who have spent their entire careers working on crossfeed/headphone algorithms. At Goodhertz, we’ve gotten to work on several licensing/consulting projects with headphone manufacturers, and it’s a little alarming how often they are inclined to intentionally boost areas of the frequency response when crossfeed would address their perceived issues more directly.

OK! This is a longer post than I’d imagined, but I hope that it at least explains our approach with CanOpener and what it does/does not do. CanOpener is really just a tool to make working on headphones easier (& more like speakers). Many of our users find it indispensable, and it’s been one of our most popular plugins for years. If you’re not one of those people, that’s totally okay!
What are your thoughts on the virtual acoustic environment aspect of this entire pursuit? Say compared to something like Slate VSX or Realphones, if we assume the headphones used are perfectly neutral or calibrated with Sonarworks, and the system (VSX/Realphones) is emulating speakers that measure perfectly neutral and is therefore not attempting to impart any coloration of either the speaker's own frequency response or headphone frequency response,

So the only variable now is the presentation of loudspeakers projecting sound through air to the ears, and the natural effect of neutral loudspeakers in the typical, professional and treated control room. What sort of differences might one expect to hear and what variables would affect the sound compared to the pure crossfeed of Canopener? And how do you feel about such products, assuming they succeed in doing exactly what they try to?
Old 24th January 2021 | Show parent
  #378
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by devinkerr ➡️
In previous versions of CanOpener there was actually some added low end (for certain types of settings/signals) as a result of the crossfeed algorithm — I think that’s where most of the quotes you’ve pulled in the previous posts come from —
I got Can Opener early on, and my system is offline for security, so I wouldn't have updated it since at least the last time I bought a new GHz plugin which may have been a few years ago now, so it's possible I was using a previous version.

Will check version numbers next time I fire it up. If it is no longer true that Can Opener is changing the spectrum, then that may be that, but it's very clear from the version I was using that it was doing just exactly that and that:


One of two things is true.

1) Sonarworks curves ONLY flatten the response. There is no additional adjustment baked in to emulate the PERCEPTION of listening to those "flat" monitors in an idealized listening environment. Can Opener DOES emulate the perception of listening in an idealized environment, so there is no redundancy.

2) Sonarworks curves flatten the response, then ALSO bake into their curves a spectrum shift that emulates the PERCEPTION of listening to those "flat" monitors in an idealized listening environment. Can Opener provides other functions, but ALSO emulates the perception of listening in an idealized environment, so ONE of the functions of each plugin is redundant, and both are assuming they need to do 100% of the lifting in that department.



So which is it? Can Opener can't be both optimized to be used with an input that's adjusted to compensate for perception of listening in a physical room AND for an input that's not.

I'm not assigning blame here. Maybe the real issue is that Sonarworks should have a defeat switch for the emulation part of what's baked into the curves which would then allow it to be used with third party tools for that purpose. Just because two tools are well designed doesn't mean they integrate well. Companies merge all the time because they have a list of great synergies. The very FIRST thing they always do is to eliminate the redundancies so they can function together.
Old 24th January 2021
  #379
Lives for gear
 
blaugruen7's Avatar
 
🎧 15 years
Have you read post 372?
Old 24th January 2021 | Show parent
  #380
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by blaugruen7 ➡️
Have you read post 372?
Did you actually read what I wrote? If you had, you might have noticed that the post you're responding to starts by quoting post 372. A detailed description of how a plugin changes the sound to achieve a desired outcome doesn't change the fact that it IS changing the sound... and that stacking it on top of another plugin that ALSO changes the sound for that exact same purpose is redundant.

The details may be complex, but the concept isn't.

Flip it around if you like (as I already said). Maybe it's not a Can Opener issue. Maybe Sonarworks simply was not designed to be followed by another plugin that ALSO adds it's own emulation of the acoustic listening environment. That's fine, but it doesn't change the fact that despite the different things each is doing, they are BOTH doing their own version of that one particular task as if the other isn't.... because neither was designed to be used specifically with the other.
Old 24th January 2021 | Show parent
  #381
Lives for gear
 
blaugruen7's Avatar
 
🎧 15 years
Quote:
Originally Posted by Bongo Spank ➡️
Did you actually read what I wrote? If you had, you might have noticed that the post you're responding to starts by quoting post 372. A detailed description of how a plugin changes the sound to achieve a desired outcome doesn't change the fact that it IS changing the sound... and that stacking it on top of another plugin that ALSO changes the sound for that exact same purpose is redundant.

The details may be complex, but the concept isn't.

Flip it around if you like (as I already said). Maybe it's not a Can Opener issue. Maybe Sonarworks simply was not designed to be followed by another plugin that ALSO adds it's own emulation of the acoustic listening environment. That's fine, but it doesn't change the fact that despite the different things each is doing, they are BOTH doing their own version of that one particular task as if the other isn't.... because neither was designed to be used specifically with the other.
I am.sorry, I didn't fully read you post
Old 25th January 2021 | Show parent
  #382
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by blaugruen7 ➡️
I am.sorry, I didn't fully read you post
No worries. I wasn't intending to be rude. It just seems to have been a waste of time in general explaining the conflict to folks in this thread. I thought the stacked EQ analogy was pretty spot on personally, but perhaps I should have picked another effect since I'm not necessarily talking about EQ specifically, but rather the idea that the same issue is being addressed twice by two different effects in serial where both are assuming they need to do the entire job.

Haven't had a chance to check version numbers, etc, but I do know for a fact that the version I first used WAS adding bass.. In fact, back then, the first discussion I dug up on it was folks discussing the same issue.

Now in 2021 I see the dev acknowledging it.

If there is a years old response now from anyone at all on here verifying that what I was said was correct, but that the plugin had been changed, I must have missed it.

Regardless, if Can Opener is designed to function as a standalone that emulates the acoustic difference between headphones and listening to speakers in a room, then it is by definition not designed to follow another plugin that already purports to do a portion of that however different the approaches may be.

We're not talking about touchy feely stuff here or "vibe" plugins to warm up a bass. This isn't a case of stacking multiple compressors because you dig the feel. It's the monitoring chain. Accuracy matters above all else. This is not an insignificant detail.
Old 25th January 2021 | Show parent
  #383
Lives for gear
 
DownSideUp's Avatar
Quote:
Originally Posted by Bongo Spank ➡️
No worries. I wasn't intending to be rude. It just seems to have been a waste of time in general explaining the conflict to folks in this thread. I thought the stacked EQ analogy was pretty spot on personally, but perhaps I should have picked another effect since I'm not necessarily talking about EQ specifically, but rather the idea that the same issue is being addressed twice by two different effects in serial where both are assuming they need to do the entire job.

Haven't had a chance to check version numbers, etc, but I do know for a fact that the version I first used WAS adding bass.. In fact, back then, the first discussion I dug up on it was folks discussing the same issue.

Now in 2021 I see the dev acknowledging it.

If there is a years old response now from anyone at all on here verifying that what I was said was correct, but that the plugin had been changed, I must have missed it.

Regardless, if Can Opener is designed to function as a standalone that emulates the acoustic difference between headphones and listening to speakers in a room, then it is by definition not designed to follow another plugin that already purports to do a portion of that however different the approaches may be.

We're not talking about touchy feely stuff here or "vibe" plugins to warm up a bass. This isn't a case of stacking multiple compressors because you dig the feel. It's the monitoring chain. Accuracy matters above all else. This is not an insignificant detail.
CanOpener is Crossfeed. Goodhertz doesn't feel that room resonance or more room simulation are needed.
quote:

On the other hand, if you try to go a more complex route and directly use spatial ear measurements or HRTF’s (like Waves NX) you’ll be relying on a set of gnarly filters that: a) only work for certain ear/head sizes/shapes b) drastically change the frequency response & introduce comb filtering, and c) are actually unnecessary for faithfully recreating the stereo field of studio monitors.

Crossfeed makes the headphones a little more in front of you, and natural. I can adjust the width, and have a some sort of experience reminding me of speakers, left right and "in front", whereas without CO, it's 180 left, 180 right and center in my head.

That's all I need, it's just a bridge between to extremes: the weirdness of headphones imaging, and two speakers in front of you.

I also use that "bridge" EQ to massage my headphones response to my taste ( Audeze usually need a lilttle bite and a lilttle less lows).

subtle tool, simple but smart.
it's NOT a room sim. to me it's a "front the headphones" trick which gets you close to real life, I wouldn't need room sim or stack room sim+ can opener.
Old 25th January 2021 | Show parent
  #384
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by DownSideUp ➡️
CanOpener is Crossfeed. Goodhertz doesn't feel that room resonance....
Except that:

Quote:
... We model the crossfeed head shadowing, interaural time delays, etc
Who wants to model room resonances? The whole point is to emulate an idealized space from the perspective of a listener in that space with idealized monitors... which is exactly what head shadowing and the rest is all about.

Sonarworks is, in fact, emulating THE SAME THING... except that it does so solely through a shift in the frequency response.

I'm really not seeing what's so difficult to grasp about this.

I'll repeat once again that it's entirely possible they pair better now if Sonarworks is emulating it from a frequency perspective, and Can Opener from other perspectives that don't shift the spectrum.

If so, great. It seems a bit kludgey having two different plugins approaching the same task from different perspectives, but if they're not overlapping in terms of their actual processing, then it may functionally be a non-issue now. If that's the case, then given that they're from different developers, this is probably about as good a resolution as one could hope for.

That absolutely positively was NOT the case with the version of Can Opener I was using, and it frankly wasn't all that subtle.

Sure seems like this conversation could have been a LOT simpler YEARS ago with just a basic:





"You said the plug was purposefully adding a spectrum shift.

It was.

You said this was a problem.

It was. In fact, we felt strongly enough about that that we have rewritten the plug to eliminate this issue from future releases.

You said there should be a way to enable the crosstalk with no spectrum shift whatsoever.

There should. In fact we designed a setting that does just exactly that and made a graph to prove it.

Problem recognized.

Problem addressed.

Problem solved.

Enjoy."
Old 25th January 2021 | Show parent
  #385
Lives for gear
 
DownSideUp's Avatar
Quote:
Originally Posted by Bongo Spank ➡️

Sonarworks is, in fact, emulating THE SAME THING... except that it does so solely through a shift in the frequency response.
To my knowledge it's absolutly not "emulating" the "same thing". Sonarworks is just fighting hardware limitations ( as they put) to make things flatter, more correct. no crossfeed or anything else.

https://www.sonarworks.com/whitepaper

For headphones (way simpler than their loudspeaker system) they aim to make the response flatter and balance left and right, SW needs the music to massage it for your unperfect headphones.

Once your headphones have been made more correct by SW, I see no problem using that with can opener. plus they are different brands so they'd probably never develop something in common.

from all I have gathered they don't do the same thing, at all. and CO should be last.
Old 25th January 2021 | Show parent
  #386
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by DownSideUp ➡️
Sonarworks is just fighting hardware limitations ( as they put) to make things flatter, more correct....
You're describing how the monitor curves work. The headphone curves are more nuanced. I laid out the difference in my first post along with the remedy. In their own words:

Quote:
The Sonarworks SR standard sets the frequency response target to be completely neutral for speaker-based playback systems, i.e. a flat FR curve across all audible frequencies AS PERCEIVED BY THE LISTENER IN THE LISTENING POSITION."
While that doesn't require addressing anything but "flattening" the hardware in the case of monitors, you can't address the listener's perspective on a headphone curve without accounting for all the things that happen up to the point where the listener hears it... which includes the sound in a physical room. It really couldn't be any other way. Their stated goal is a sort of universal idealized listening experience, and they're very clear that that ideal experience is one of speakers in a room, not headphones, therefore, they MUST adjust the headphones TOWARD that ideal, and not simply "flatten" the hardware as they do with the monitors. Without that, they would fail at their goal, and it would be difficult to switch back and forth with confidence... which is the whole point.

They're not attempting to create "flat" headphones. They say so. They're attempting to create on headphones the listener perception of flat monitors in a properly tuned listening environment. SW really only does one thing thoug... EQ. The lack of crosstalk opens an excellent opportunity for synergy with a third party plugin.

The early version of C.O, however, was ALSO adding bass. Again, fine in isolation, but absolutely redundant for SW users when their own frequency manipulation is already baked into the curve.

What I said originally was so on the nose it's really not even an analogy at all. Using CO with SW was quite literally stacking two EQ's on top of each other to deal with the same issue. Was each plug ALSO doing other things? Sure... but THAT bit was completely redundant.

I stated all of this very clearly in my first post along with exactly how the plug needed to be changed to fix the issue.

... which is exactly how the plug WAS changed.

Quite a coincidence that random change was made for no particular reason.
Old 26th January 2021 | Show parent
  #387
Lives for gear
 
DownSideUp's Avatar
Quote:
Originally Posted by Bongo Spank ➡️
No. You're describing what they do with SPEAKER curves. What they do with HEADPHONE curves is more nuanced. I laid it out clearly in my first post, but if you're not clear on this, then you can start here:



Whereas the speaker curves simply "flatten" the the response (i.e. "fixing" the hardware), that is NOT what the headphone curves are doing. They are specifically designed to address the listener's perception... which occurs after the audio has left the speaker and traveled through the room to the listener's ears.

The fact that it is limited to EQ as it's only remedy means it doesn't delve into crosstalk creating an excellent opportunity for synergy, but it IS adjusting the EQ curve to reflect the perspective of the listener... not the speaker itself.

The early version of C.O that was ALSO adding bass FOR THE EXACT SAME REASON (per dev... not my opinion) absolutely WAS double dipping on that task and while it may make sense without SW, it was redundant when used in conjunction.

I stated all of this very clearly in my first post along with the remedy.... which is exactly how the plug was changed. Quite a coincidence that random change was made for no particular reason.
Yep you quoted what SW do for speakers. That's why they talk about "listening position".

For speakers SW is fighting the the speakers response in the room ( the whole room, all areas) + time domain etc and it's got presets, from what i remember of simulated speakers.

For headphones, it is correcting the response of the membranes, measured with their special system. that's it. They never talk about listening position, or room position.
use a preset of your cans, or get your membranes measured. done, corrective EQ.
no simulation of anything. NOT mentioned in the white paper or anywhere in any videos I found.
And btw, the user guide from the headphone edition, is full of speakers sim infos and has no info about the headphone dsp. yes the link shows "SW4-headphones-manual" but no headphones infos.



then again anyway, CO can be adjusted so it's all good.
Old 26th January 2021 | Show parent
  #388
Gear Nut
 
Bongo Spank's Avatar
 
Quote:
Originally Posted by DownSideUp ➡️
Yep you quoted what SW do for speakers. That's why they talk about "listening position".
What you seem to be missing is that the target FR for Headphones and speakers is the SAME. That's the entire point of the endeavor. Maybe it's easier to see it from the opposite perspective and say it's the speakers that need the EXTRA calibration since they need to fight the room acoustics. It's all just a matter of perspective, though. What's really happening whichever way you view it is that there is ONE and ONLY one FR. The FR the headphones are calibrated to is the very same FR the speakers are calibrated to... as measured from the listening position in the room.

You can't calibrate headphones as if they're speakers with no regard for the listening position.. because that's not how speakers are calibrated. They have to match. That's the whole deal.

Forget about the details for a moment and just think through the common sense. The whole goal here in plain English is that you should be able to sit in the listening position with calibrated monitors and calibrated Headphones, and switch between them and have them sound exactly the same.

In order to meet that goal, the headphone FR absolutely HAS TO match the FR of the speakers specifically from the listener's perspective in the room.

This isn't a minor detail. It's the whole concept in a nutshell. The single unified FR is literally the whole reason SR exists. It's the reason the white paper was written.
Old 26th January 2021 | Show parent
  #389
Gear Addict
 
devinkerr's Avatar
 
🎧 10 years
Quote:
Originally Posted by Phosphenetre ➡️
What are your thoughts on the virtual acoustic environment aspect of this entire pursuit? Say compared to something like Slate VSX or Realphones, if we assume the headphones used are perfectly neutral or calibrated with Sonarworks, and the system (VSX/Realphones) is emulating speakers that measure perfectly neutral and is therefore not attempting to impart any coloration of either the speaker's own frequency response or headphone frequency response,

So the only variable now is the presentation of loudspeakers projecting sound through air to the ears, and the natural effect of neutral loudspeakers in the typical, professional and treated control room. What sort of differences might one expect to hear and what variables would affect the sound compared to the pure crossfeed of Canopener? And how do you feel about such products, assuming they succeed in doing exactly what they try to?
Good questions!

CanOpener effectively tackles the “projecting sound through air to the ears” part with crossfeed, which is the biggest difference between speakers and headphones, but our feeling is that modeling the in-room response or the sound of the control room (or speakers) itself is counterproductive for a couple reasons:

1) Many of the best mixing/mastering rooms in the world are designed to have minimal interaction between the speakers and the room. The overhead diffusers in a Northward Acoustics room, for example, are there mostly for self-noises created by the engineer, not the speakers, and the speakers essentially see the room as anechoic (which is an extremely difficult/costly thing to do in a real acoustic space, but quite easy with headphones!). This is similar with CanOpener: we model the journey from speaker to ears but nothing more — no reverberation, speaker/room coloration or distortion.

2) We’ve done listening tests at Goodhertz with virtual control rooms and they all mostly sounded like cheap tricks on headphones, and we found they clouded mixing judgements (especially reverb-related judgements). I think part of the reason for this is that in a real room, your own natural HRTF’s allow you to focus on the sounds emitted from the speakers while de-prioritizing room sound and reflections from the ceiling, rear, etc. On headphones, this is not the case as that 3D information is largely lost and it becomes much more difficult to separate the music from the room response. Trying to replicate the full 3D environment on headphones would be extremely difficult: it would require a ton of processing power to model correctly, as well as individualized 3D measurements of your ear responses convolved with a high-channel count 3D reverb (sort of like Dolby Atmos). While all of this would be possible (though quite expensive), I don’t think it would provide any better tools for mixing/mastering than what we currently have with crossfeed. The room I work in, for example, contributes almost nothing to the speaker response and anything it does contribute is a flaw in the acoustics.

One exception to all this stuff above is the physical sensation of bass frequencies that happens in a control room, which can be important for mixing and judging bass punch. We’ve seen lots of CanOpener users either mixing with light subwoofers or something like a SubPac for that purpose.
Old 26th January 2021 | Show parent
  #390
Here for the gear
Hey Bongo Spank, I was agreeing with your point of view, but now I think there's something missing.

Quote:
Originally Posted by Bongo Spank ➡️
It is further my understanding that the headphone curves do this as well, but then also do the following:
2) Bend the flattened curve to a proprietary Harman-esque reference designed to approximate the perception of a flat monitors in a tuned studio environment.
First, you've said that it is your understanding and asked to correct you if you're wrong.

Quote:
Originally Posted by Bongo Spank ➡️
Sonarworks is, in fact, emulating THE SAME THING... except that it does so solely through a shift in the frequency response.
Then, suddenly, it became a fact. No proof though.

Quote:
Originally Posted by Bongo Spank ➡️
While that doesn't require addressing anything but "flattening" the hardware in the case of monitors, you can't address the listener's perspective on a headphone curve without accounting for all the things that happen up to the point where the listener hears it... which includes the sound in a physical room.
Then there's a reasoning on how it should work in your understanding (not how it really works, though, unless you're the engineer lead behind the Sonarworks Reference).

What I think is missing here is that the "sound of a speaker in a physical room from a listener position" is actually just a flat response in case of using Sonarworks. All that measuring is made to aim for a flat FR, with all the speakers inaccuracies and room resonanses compensated.

Quote:
Originally Posted by Bongo Spank ➡️
What you seem to be missing is that the target FR for Headphones and speakers is the SAME. That's the entire point of the endeavor. Maybe it's easier to see it from the opposite perspective and say it's the speakers that need the EXTRA calibration since they need to fight the room acoustics. It's all just a matter of perspective, though. What's really happening whichever way you view it is that there is ONE and ONLY one FR. The FR the headphones are calibrated to is the very same FR the speakers are calibrated to... as measured from the listening position in the room.
Exactly. The target FR is a FLAT response for the whole frequency spectrum.

Here's the simple illustration:

In case of the speakers both the speakers and the room will color the sound. For the headphones the sound is being colored only by a membrane.

Let's say we have a pair of speakers in the accoustically ideal room, and a pair of headphones. In this room both the speakers and the headphones sound the same, let's say they both add 2db on 60Hz and otherwise they have an absolutely flat FR.

We place the speakers in the not-so-ideal room and measure the FR with Sonarworks. If the speaker adds 2db on 60Hz and also the room resonates on the same frequency and adds 3db more, then we'll need to apply -5db on 60Hz to get it flat.

With the headphones there's NO room involved, so all we need to do is to apply -2db on 60Hz.

Thus, with the speakers in a room and -5db 60Hz compensation we get the same sound as with the headphones and -2db 60Hz compensation. They sound flat.

What I do not understand is where does this come from:
Quote:
Originally Posted by Bongo Spank ➡️
proprietary Harman-esque reference designed to approximate the perception of a flat monitors in a tuned studio environment.
📝 Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 1005 views: 119176
Avatar for MogwaiBoy
MogwaiBoy 1 week ago
replies: 1343 views: 192050
Avatar for Zmago Šmon
Zmago Šmon 4 weeks ago
replies: 0 views: 324
Avatar for traveler984
traveler984 28th January 2021
Topic:
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump