DSPplug announces: DSPplug boardstation

[TIMT]

King OZ; the manager of DSPplug and VSTplug has invented a new algorithm

[kingozrecords]

King OZ; the manager of DSPplug and VSTplug has invented a new algorithm

King OZ; the manager of DSPplug and VSTplug has invented a new algorithm

Thankyou, that's kind to say. I'm sorry I didn't respond earlier; I've been coming up with a way to convert the textual name of folders to a float 16 point decimal value so as to be saved in the default float format. It means I've been fiddling with a reg_replace method but it's very annoying. However, it will lead to My ability to make folder patches for the upcoming free DSPplug beat drum machine sampler.

The idea behind that is that there is folders or zip files within a folder and that unlike an .sfz there is no text file designed to control it. Instead, the default settings can optionally be specified in the sample's name. But, if you think about it, how will the preset remember the folder without a string saved? because VST only likes to save 0-1 as a rule. So I need to truncate all of the text to ascii and then make it into numbers somehow so that it remembers the location and that way if you add a new file it won't instead depend on the order. I mean, if you made your own factory patches, you could just number them; but what if people make patches and folders themselves... You see the problem. Making things that are "easier" to use is very hard.

OK, that said and out of the way; updates:

30/05/2020 - Updated! VSTplug compressor & compressor 2

Information
Forum users had asked why the compressors were the same color scheme, so I've made the compressor and compressor 2 new colors!

VSTplug compressor:
In harvest gold, like a seventies touch.

demo download
https://vstplug.com/?download=1&kccpid= ... 6-DEMO.zip

marketplace link
https://www.kvraudio.com/marketplace/vs ... oz-records

screenshots




VSTplug compressor 2:
In hot red/pink like an older corvette, my favorite.

demo download
https://vstplug.com/?download=1&kccpid= ... 1-DEMO.zip

marketplace link
https://www.kvraudio.com/marketplace/vs ... oz-records

screenshots

[swatwork]

But, if you think about it, how will the preset remember the folder without a string saved? because VST only likes to save 0-1 as a rule.

Flowstone can save text strings as part of a preset. Look up "VST Preset String"
in the component reference.

[BlitBit]

[...]
Thankyou, that's kind to say. I'm sorry I didn't respond earlier; I've been coming up with a way to convert the textual name of folders to a float 16 point decimal value so as to be saved in the default float format. It means I've been fiddling with a reg_replace method but it's very annoying. However, it will lead to My ability to make folder patches for the upcoming free DSPplug beat drum machine sampler.

The idea behind that is that there is folders or zip files within a folder and that unlike an .sfz there is no text file designed to control it. Instead, the default settings can optionally be specified in the sample's name. But, if you think about it, how will the preset remember the folder without a string saved? because VST only likes to save 0-1 as a rule. So I need to truncate all of the text to ascii and then make it into numbers somehow so that it remembers the location and that way if you add a new file it won't instead depend on the order. I mean, if you made your own factory patches, you could just number them; but what if people make patches and folders themselves... You see the problem. Making things that are "easier" to use is very hard.
[...]

No, you don't need to do that and to be frank it sounds like a nightmare. I am pretty sure that the VST interface also provides means to save your own representation of a patch. If you use that interface then you can store the paths as simple strings and don't need to do these brittle transformations.

HTH

[kingozrecords]

[...]
Thankyou, that's kind to say. I'm sorry I didn't respond earlier; I've been coming up with a way to convert the textual name of folders to a float 16 point decimal value so as to be saved in the default float format. It means I've been fiddling with a reg_replace method but it's very annoying. However, it will lead to My ability to make folder patches for the upcoming free DSPplug beat drum machine sampler.

The idea behind that is that there is folders or zip files within a folder and that unlike an .sfz there is no text file designed to control it. Instead, the default settings can optionally be specified in the sample's name. But, if you think about it, how will the preset remember the folder without a string saved? because VST only likes to save 0-1 as a rule. So I need to truncate all of the text to ascii and then make it into numbers somehow so that it remembers the location and that way if you add a new file it won't instead depend on the order. I mean, if you made your own factory patches, you could just number them; but what if people make patches and folders themselves... You see the problem. Making things that are "easier" to use is very hard.
[...]

No, you don't need to do that and to be frank it sounds like a nightmare. I am pretty sure that the VST interface also provides means to save your own representation of a patch. If you use that interface then you can store the paths as simple strings and don't need to do these brittle transformations.

HTH

Certain DAW's give errors as far as I know. I used to work with php and once coded bbcode; so really, no less complicated than that. It comes down to skill set . Just what I'm used to.

[BlitBit]

[...]
Certain DAW's give errors as far as I know. I used to work with php and once coded bbcode; so really, no less complicated than that. It comes down to skill set . Just what I'm used to.

If certain DAWs do not support this part of the VST interface then they are severely broken and lack very basic functionality.

In software development it's also an important skill to use the right tool for the job and to find out what's supported by certain libraries/interfaces and what's not. This usually means reading documentation and existing code.

You are currently taking a needlessly complicated path which is very bug prone while two people have already told you that there's an easier and "officially" supported way. All the time that's spent fiddling bits could be spent for much more interesting things, e.g. actual algorithms, DSP, etc.

Perhaps you want to do this as a challenge. However, I would not take that challenge in a software that I'd want to sell.

[kingozrecords]

[...]
Certain DAW's give errors as far as I know. I used to work with php and once coded bbcode; so really, no less complicated than that. It comes down to skill set . Just what I'm used to.

If certain DAWs do not support this part of the VST interface then they are severely broken and lack very basic functionality.

In software development it's also an important skill to use the right tool for the job and to find out what's supported by certain libraries/interfaces and what's not. This usually means reading documentation and existing code.

You are currently taking a needlessly complicated path which is very bug prone while two people have already told you that there's an easier and "officially" supported way. All the time that's spent fiddling bits could be spent for much more interesting things, e.g. actual algorithms, DSP, etc.

Perhaps you want to do this as a challenge. However, I would not take that challenge in a software that I'd want to sell.

Hmm, maybe you're right - well to be honest I just wanted to err on the side of caution due to various reports by other developers I've read of various companies suggesting they were switching all of their presets over to float rather than string.

But, yeah I guess maybe I could do that. Though in the end I may have to do the same. Thanks for the advice.

If, in the end I'll still be able to switch it over to a float in the end anyways, at least the preset id# will be the same and should allow that bug fix. It could speed things along. I'm used to bugs being pervasive however seemingly implausible like only recently realizing that Acoustica Mixcraft 9 and Steinberg Cubase 10 have less font support for VST. I guess it came down to fonts being listed like in the way of a string: "Arial,2,Normal,Alternate Font". Funny thing sometimes.

[kingozrecords]

"I have invented an even newer algorithm"

Here's a video showing My new metering vs older metering styles. And I show how I use double precision etc.

https://youtu.be/f4dSLT2O_WA

[BlitBit]

"I have invented an even newer algorithm"

And it seems that this new algorithm just fudges some numbers that you don't understand (you even say so in the video at around 04:34). Around the 5:00 mark one can see that the algorithm seems to do the following:

1. Take a constant value of -0.31 and feed it into a low pass filter. This means that after some time the low pass filter will also start to spit out something around -0.31 for every sample.
2. You then convert that value to a double. This does not really make sense because in the following it's not used as part of a computation that requires double precision.
3. The constant value of -0.31 is then subtracted from an averaged signal.

To put it short: the algorithm uses a very complicated way to subtract a constant from a signal. I hope that this is not the actual audio stream that comes out of the plugin because in that case it would introduce DC to the signal, i.e. it would worsen it.

Also please note that the other three plugins all show -20.2 dB LUFS because they have obviously understood the EBU loudness specification and implemented it correctly. It's a test signal that's intended to result in -20.2 dB LUFS if the algorithm is implemented correctly. If an algorithm shows something different it's not compliant with the specification.

[Ploki]

jesus christ. LUFS is weighted and has very specific ballistics.
you're not supposed to just randomly add and remove constants from metering.

you have invented a horrible way to meter shit that nobody at all needs.

instead of being obsessed with doing things "differently" for the f**k of it, learn the basics.

[kingozrecords]

Well, technically it's just A and a C weighted version of dBFS and actually it's quite useful if you know what to look for, but I guess everyone doesn't study it. So, I gave in and learned how to do it the official way. And assuming that youlean hasn't added inter-sample peak detection even moved ahead in accuracy over his by -3 LU.

Not to worry updating soon.

I'll add the option of using the older or newer EBU spec, 2010 - 2016 EBU option, one without oversampling, the other with (inter-sample peak detection)

Here's the first video where I was learning about the averaging:
https://www.youtube.com/watch?v=f4dSLT2O_WA

and the new video where I've succeeded in meeting the newer EBU test, but I suspect Youlean and others don't like the newer recommendation as much and default to the older model.
https://youtu.be/b4fRN6Lr70M

OK Ploki? Not to worry I'm a busy guy. But really an average is not all that useful when you gain the years of skill. Unless you over-sample anyways. hard to beat that kind of accuracy.

[gentleclockdivider]

Comedy Gold

[TB-ProAudio]

"I have invented an even newer algorithm"

Here's a video showing My new metering vs older metering styles. And I show how I use double precision etc.

https://youtu.be/f4dSLT2O_WA

 

The spec you are referring (R128) is from EBU and a very serious thing: https://en.wikipedia.org/wiki/EBU_R_128
The algorithms are well known and easy to implement (freely available on the net).
BTW: all R128 loudness values are based on K-weighting (not A or C). The spec from EBU defines everything. No need to add constants...

[kingozrecords]

"I have invented an even newer algorithm"

Here's a video showing My new metering vs older metering styles. And I show how I use double precision etc.

https://youtu.be/f4dSLT2O_WA

 

The spec you are referring (R128) is from EBU and a very serious thing: https://en.wikipedia.org/wiki/EBU_R_128
The algorithms are well known and easy to implement (freely available on the net).
BTW: all R128 loudness values are based on K-weighting (not A or C). The spec from EBU defines everything. No need to add constants...

Hello TBPro, I like the color of your software; our color in some applications is a bit similar, you have a lot of skill.

but yes, I'd thrown together a schematic that can accomplish this (which requires flowstone, will work in the free fl studio compatible version).

http://dsprobotics.com/support/viewtopi ... 55#p144632

In finer detail Bob Katz combined A and C due to his own experience with weighting; I know because I was the one who'd emulated the Linkwitz Riley filters using a transposed direct form assembly made by Martin Vicanek, and I'd been reading about in school. It is his own combinations which became known as K-weighting.

I'd then followed recommendations and added oversampling as well as double precision. However, there is one instance in My schematic in which double precision is un-necessary.

So, in regards to it being serious I very much agree and was frankly rather surprised that quite a few products have fallen back on the 2010 recommendation and are not employing either double precision and oversampling. It makes Me wonder which is the rationale for this falling on old designs and recommendations:
Convenience and ease of creation, and popularity of the 2010 recommendation being employed. So maybe it is best to leave the double precision in and then in two points, to allow the over sampling to be turned off.

Once you do merely that; it seems to meet the 2010 spec, so I've provided that schematic which could quite easily be converted to C for anyone's use should they wish to accomplish r128.

[vurt]

can it do stereo to mono?