You are here

Stuck in the Glue

For everything after the recording stage: hardware/software and how you use it.

Re: Stuck in the Glue

Postby ManFromGlass » Wed Aug 22, 2018 11:53 am

Following this with great interest. Installed the demo version. Read the SOS review from a few years ago. Liking this plug. New tool for the arsenal!
The photo of the hardware unit in the review reminded me of mixing a project 20 years ago when the studio owner just bought one for what felt like an obscene amount of money. (Could it have been $8k Cdn? Memory unsure)
We played with it, pushing the unit hard, almost pinning the meters and it still sounded wonderful.
Anything that makes me sound like I can actually play well is always welcome!
User avatar
ManFromGlass
Frequent Poster
Posts: 1410
Joined: Sat Jul 23, 2011 11:00 pm
Location: In the woods in Canada
 

Re: Stuck in the Glue

Postby andrewsimper » Mon Aug 27, 2018 2:42 am

Hugh Robjohns wrote:EDIT -- apologies -- I thought I was responding to the Op, not the designer.. but I think the points still relate appropriately:

andrewsimper wrote:In summary: The Glue is a feedback compressor.

That's not what the block diagram in the manual shows. There it is drawn as a feed-forward compressor -- the side chain takes it's input from the input signal, not the output.

Sorry for the late reply, I need to update my notification settings.

It actually does show the compressor as a feedback one in the diagram, please have another look, and take your time. There are two amps, which may be the confusing bit.

Look at the bottom signal flow, and think about just that. Think of the output of the bottom amp as being the actual output, then look what happens. The output gets ratio shaped, then envelope followed, then that control voltage signal is "fed back" into the bottom amp, so there is a feedback loop from output to input and it is indeed a feedback compressor.

Now lets just say you wanted independent control of the actual output level. For that you would need a second amp, then you can do things like fade ins and fade outs as well as makeup gain. The top amp is this "second" amp, it takes the same amp control voltage as the bottom amp, so is identical to a regular feedback compressor if you send the same audio to both amps and sent the threshold gain to the makeup gain.

Please let me know if this insight helps, then we can move on to what parallel compression actually does compared to the range knob for steady state fixed frequency sine waves (ie a very contrived but basic illustrative case).
User avatar
andrewsimper
New here
Posts: 7
Joined: Sun Sep 19, 2010 11:00 pm

Re: Stuck in the Glue

Postby andrewsimper » Mon Aug 27, 2018 2:58 am

Martin Walker wrote:Indeed - welcome Andrew to the discussion, and to the SOS Forums! 8-)

There are some sticky arguments here, but you should be able to shed some truth on the proceedings, as you created the Glue to great acclaim ;)


Martin
Thanks Martin! I didn't have my "notify me when a reply is posted" enabled, but it is now :)

It's great that people are interested in understanding more, that is the most important thing. A few misunderstandings are actually good since it gets people thinking and usually there are many people reading threads that could possibly have the same misunderstanding, so many people can be helped at once.

The Glue is coming up to 10 years old now, and I'm very proud how well it stands up to modern offerings considering the core DSP hasn't changed in all that time, only the oversampling options and user interface. It's about to receive a major update, with main signal chain non-linear modelling and in depth DBX Blackmer amp modelling. I just need to finish The Scream first :)
User avatar
andrewsimper
New here
Posts: 7
Joined: Sun Sep 19, 2010 11:00 pm

Re: Stuck in the Glue

Postby andrewsimper » Fri Aug 31, 2018 12:50 am

Hugh Robjohns: are you still participating in this thread? Please let me know either way :)
User avatar
andrewsimper
New here
Posts: 7
Joined: Sun Sep 19, 2010 11:00 pm

Re: Stuck in the Glue

Postby Hugh Robjohns » Fri Aug 31, 2018 12:35 pm

Yes, still hanging around in the darker recesses of the forum ;-) ...but this thread had fallen off my radar...

andrewsimper wrote:Look at the bottom signal flow, and think about just that. Think of the output of the bottom amp as being the actual output, then look what happens.

Ah ha! I see what you're getting at! :smirk: I misunderstood what you were trying to illustrate in the sidechain section of the diagram -- running two gain control elements in parallel with the same control voltage, cunning! So I apologise -- it is indeed a feed-back compressor. :thumbup:

Please let me know if this insight helps

It has clarified the diagram enormously. Thanks.

H
User avatar
Hugh Robjohns
Moderator
Posts: 21535
Joined: Thu Jul 24, 2003 11:00 pm
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Stuck in the Glue

Postby Martin Walker » Fri Aug 31, 2018 4:35 pm

Phew - when you get two such experts as the likes of Andrew Simper and Hugh Robjohns disagreeing, the world seems to be closer to ending :?

I'm SO glad you're both back on the same page - my faith has been restored :clap:


Martin
User avatar
Martin Walker
Moderator
Posts: 12509
Joined: Wed Jan 13, 2010 8:44 am
Location: Cornwall, UK

Re: Stuck in the Glue

Postby Gone To Lunch » Fri Aug 31, 2018 8:33 pm

Martin Walker wrote:Phew - when you get two such experts as the likes of Andrew Simper and Hugh Robjohns disagreeing, the world seems to be closer to ending :?
Martin

I did wonder what have I started !

Mind you, I got my question answered pretty thoroughly, which is great so thanks everyone.

Which brings another question - Andrew, why not put all this in the manual ?
Gone To Lunch
Frequent Poster
Posts: 707
Joined: Thu Jun 10, 2004 11:00 pm
Location: London

Re: Stuck in the Glue

Postby Hugh Robjohns » Fri Aug 31, 2018 11:19 pm

There's always something new and interesting to learn, what ever level you're at! :thumbup:
User avatar
Hugh Robjohns
Moderator
Posts: 21535
Joined: Thu Jul 24, 2003 11:00 pm
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Stuck in the Glue

Postby andrewsimper » Mon Sep 03, 2018 4:32 am

Hugh Robjohns wrote:Yes, still hanging around in the darker recesses of the forum ;-) ...but this thread had fallen off my radar...

andrewsimper wrote:Look at the bottom signal flow, and think about just that. Think of the output of the bottom amp as being the actual output, then look what happens.

Ah ha! I see what you're getting at! :smirk: I misunderstood what you were trying to illustrate in the sidechain section of the diagram -- running two gain control elements in parallel with the same control voltage, cunning! So I apologise -- it is indeed a feed-back compressor. :thumbup:

Please let me know if this insight helps

It has clarified the diagram enormously. Thanks.

H

Ok, great :) Now that is covered let's move onto the differences between the range knob and the mix knob by taking a look at some static compression plots generated by DDMF's Plugin Doctor: https://ddmf.eu/plugindoctor/

First is the regular compression curve 100% wet with the range knob on full. Second is 100% wet with the range knob at -12 dB, third up is 79% wet and range knob on full. You can see that with the range knob you keep the ratio you want and then back off once you hit a certain point, but with parallel compression the effective ratio is reduced the more dry signal you add. This is one key difference between the two, but there is also a non-linear difference which isn't shown here to do with the compression action always being there with the range knob, but partially bypassed with parallel compression. It doesn't really matter the exact details, it's more important to listen to both and decide which works best in a particular context.

Image
Image
Image

Please remember that this isn't the full picture, but it does give you an idea of what to expect with when using the mix knob vs the range knob. These types of plots don't show accurately what is actually going on to produce the "knee", and also don't show what happens with different frequencies, or what happens dynamically with content dependant compression.
User avatar
andrewsimper
New here
Posts: 7
Joined: Sun Sep 19, 2010 11:00 pm

Re: Stuck in the Glue

Postby andrewsimper » Mon Sep 03, 2018 4:43 am

Gone To Lunch wrote:
Martin Walker wrote:Phew - when you get two such experts as the likes of Andrew Simper and Hugh Robjohns disagreeing, the world seems to be closer to ending :?
Martin

I did wonder what have I started !

Mind you, I got my question answered pretty thoroughly, which is great so thanks everyone.

Which brings another question - Andrew, why not put all this in the manual ?
I try to keep the manual as brief as possible as very people actually read them! I prefer to educate via Q&A on forums as well as writing technical articles on my web page. Currently the technical articles are all heavily DSP related, and mainly benefit other developers, but I will be writing a bunch of more production focussed articles in the future so keep an eye out for them.
User avatar
andrewsimper
New here
Posts: 7
Joined: Sun Sep 19, 2010 11:00 pm

Re: Stuck in the Glue

Postby andrewsimper » Mon Sep 03, 2018 12:09 pm

PS: i forgot to mention that the way you blend in the dry signal doesn't really matter too much, just the relative amplitudes of dry vs wet. In The Glue, because the dry and wet are in phase I use a regular -6dB crossfade, so at 50% you get 0.5*dry + 0.5*wet. If the overall gain is higher or lower then this will offset the ratio curve upwards or downwards, but it will not change the slope, which is the important bit. More specifically, in the section below the threshold it will always be parallel to the 1:1 "no compression" dotted line.
User avatar
andrewsimper
New here
Posts: 7
Joined: Sun Sep 19, 2010 11:00 pm

Previous

Who is online

Users browsing this forum: No registered users