Big Companies That Get It Right

[AnX]

Honorary mention to Microsoft

without them, i prob couldnt afford to make music, and most of the other companies wouldnt exist

[kingozrecords]

You know, I recently made a EQ that uses trigonometric functions. But I wonder if trig is'nt the very reason why some EQ's fall short. I mean, like in the case of the alloy 2 eq, it seems You need to use the de-esser in tandem; And in all fairness could that not be due to the fact that a multi-threaded algorithm is being iterated 44,100 or 48,000 times a second? People often forget that it was designed as a visual math, allowing You to decide the plane of ellipses; Normally the most prudent in architecture.

For instance something like a round doorway or a porticullis. But the math involved uses more processor than mere float multiplication alone. I think the real issue is that this math takes half that 44'100'th of a second to catch up meaning that half the original sound You are trying to filter out bleeds through unhindered for split seconds at a time. So like half of 44,100.

Given that; It'd be nice to see the big guys use math knowingly, realizing what the impact of it is.
This is the EQ without trigonometry:
https://www.kvraudio.com/product/la-ban ... ds/details

and this is the one with:
https://www.kvraudio.com/product/la-ban ... oz-records

See I like both, but there really should be the choice of such a filter; And if low or high passes are applied, they should be done before hand, and innately linear to prevent the phenomenon.

[Burillo]

You know, I recently made a EQ that uses trigonometric functions. But I wonder if trig is'nt the very reason why some EQ's fall short. I mean, like in the case of the alloy 2 eq, it seems You need to use the de-esser in tandem; And in all fairness could that not be due to the fact that a multi-threaded algorithm is being iterated 44,100 or 48,000 times a second? People often forget that it was designed as a visual math, allowing You to decide the plane of ellipses; Normally the most prudent in architecture.

For instance something like a round doorway or a porticullis. But the math involved uses more processor than mere float multiplication alone. I think the real issue is that this math takes half that 44'100'th of a second to catch up meaning that half the original sound You are trying to filter out bleeds through unhindered for split seconds at a time. So like half of 44,100.

Given that; It'd be nice to see the big guys use math knowingly, realizing what the impact of it is.
This is the EQ without trigonometry:
https://www.kvraudio.com/product/la-ban ... ds/details

and this is the one with:
https://www.kvraudio.com/product/la-ban ... oz-records

See I like both, but there really should be the choice of such a filter; And if low or high passes are applied, they should be done before hand, and innately linear to prevent the phenomenon.

i'm not really a DSP wizard, but what you said made absolutely no sense whatsoever. not to mention it's woefully off topic, to the point of me wondering if you have posted in the wrong thread.

[kingozrecords]

I was pointing out that big name companies are using technologies that are unreliable. If they were built more intuitively, there wouldn't be so many disappointments in the end result.

I guess My take is merely from the technical end; But that in turn is relevant. If one is to use a filter to be a pro with audio then I guess that person should know what it is doing.

And I went on to say doing it right means by using linear filters rather than multi stage or trigonometric based for high-passes and low-passes as they are far more reliable.

[Burillo]

I was pointing out that big name companies are using technologies that are unreliable.

that's a very bold claim to make.

If they were built more intuitively, there wouldn't be so many disappointments in the end result.

that... is bordering on pseudoscience peddling. quite often the tools that are the most reliable are the least intuitive. building precise tools "intuitively" is a recipe for disaster.

And I went on to say doing it right means by using linear filters rather than multi stage or trigonometric based for high-passes and low-passes as they are far more reliable.

define "reliable".

[kingozrecords]

Like I used My stuff on the DB here to master My latest track: