DYN4M1X beta: opinions wanted

[bmanic]

Looks really interesting. I'll try it out tonight for sure.

Cheers!
bManic

[bmanic]

Well it doesn't take long to realize what a marvelous compressor this is! Good work man!

Now, some feature requests and ideas:

- Mid/Side encode and decode on input/output. You already have a very flexible stereo linking one knob solution so simply inserting an encoder/decoder at the i/o will give this compressor even more flexibility! For added control in Mid/Side mode the make-up gain and threshold should be split as well.

- Range feature needs 0.1dB resolution at the 0 to 6dB range for precise control of dynamics. This would help in a mastering situation. Actually, ALL knobs need finer resolution. 0.1dB resolution should be enough.

- Add a built in 1 or 2 band parametric EQ for the sidechain and add a LPF

- Oversampling button for the compression only would be nice. This way one doesn't have to use the soft limiting to get higher resolution

- Experiment with different "ballistics" for the attack and release behavior. What I mean is basically the amount of acceleration before the detector hits the desired db/ms rate. Adding this kind of lag or boost which is program dependent (for instance tied again to the amount of compression). Or perhaps I've understood you wrong how the current attack/release works?

Also, try experimenting with oversampling the control signal path only without touching the incoming audio itself.

EDIT: one more request, why not "zoom in" on the transfer curve when the range knob is used? Especially at low settings around 2dB it becomes nearly impossible to see the curve.

Cheers!
bManic

[IIRs]

Heh. Its a whole new plug you're asking for!

- Mid/Side encode and decode on input/output. You already have a very flexible stereo linking one knob solution so simply inserting an encoder/decoder at the i/o will give this compressor even more flexibility! For added control in Mid/Side mode the make-up gain and threshold should be split as well.

I'm not sure where to squeeze in the extra knobs! How about an MS stereo version with totally separate controls for each channel?

- Range feature needs 0.1dB resolution at the 0 to 6dB range for precise control of dynamics. This would help in a mastering situation.

Perhaps a log scale for that knob?

- Add a built in 1 or 2 band parametric EQ for the sidechain and add a LPF

I thought about that, but then decided the external sidechain inputs made it unnecessary..?

- Oversampling button for the compression only would be nice. This way one doesn't have to use the soft limiting to get higher resolution

That shouldn't be too hard.

- Experiment with different "ballistics" for the attack and release behavior. What I mean is basically the amount of acceleration before the detector hits the desired db/ms rate. Adding this kind of lag or boost which is program dependent (for instance tied again to the amount of compression). Or perhaps I've understood you wrong how the current attack/release works?

Yeah, I had a go at that actually. I was using a transient detector type algorithm to speed up the envelope on fast transients and slow it down otherwise... still not happy with the results I'm getting: will have to play around with it some more.

Also, try experimenting with oversampling the control signal path only without touching the incoming audio itself.

What benefit would that have? Aliasing occurs when then audio signal is multiplied by the gain modulation signal, right? What am I missing?

EDIT: one more request, why not "zoom in" on the transfer curve when the range knob is used? Especially at low settings around 2dB it becomes nearly impossible to see the curve.

That's an interesting idea, but I have no idea how to do it! I will give it some thought...

Thanks for the feedback

[CaptainMARC]

Sheffield? You an Owl or a Blade?

Anyway, the plugin sounds quite fab! Excellent!

***EDITED*** (found more stuff out...)

Strange, on my hopelessly outdated but almost legendary punk-rock-laptop, the CPU hit is around 8%, which is pretty respectable for a Synthmaker plug, but only as long as my host (Reaper) is running.

When host is stopped the CPU hit goes up to 24% when compressing and a whopping 48% when expanding! Crikey! Denormals? (I have a Pentium 4m.)

AS for bmanic:

Don't need any M/S stuff because I can do that with a freeware M/S encoder/decoder.

Range knob is ace. Love it. I do a fair bit of mastering but don't need 0.1dB resolution, although I can see others might like it.

I agree with you that I don't need any more EQ in the sidechain. If I really needed that, and I might, because this might turn out to be a really funky de-esser, I will just throw the low-CPU-hit EQ of my choice on the external sidechain. The HPF is dead useful though.

Oversampling on comp only would be nice. Perhaps automatically activated when rendering offline?

I like the ballistics just the way they are.

Don't need zoom on the transfer curve because, well, I'm sort of using my ears.

As for me:

I assume this is feed-forward (at least that's what it sounds like to me). It'd be smashing if you could have a switch to turn it to feed-back - that is that the ouput feeds the internal sidechain. That would make it dead sexy for emulating all kinds of classic hardware (I love all that Urei stuff).

Otherwise, it would be nice to sort that CPU business, but I think it might be a Synthmaker issue.

Even as it is, it would make my top three "Desert Island Compressors". Sounds great (absolutely outstanding to be honest) and very easy to use.

As for you lot:

Anybody who hasn't grabbed it yet, DO IT NOW!

Cheers & COYI!

PS
THERE'S ONLY ONE CARLOS TEVEZ! (hahahaha)

[Download SOphist]

thanks for this, will try! (really like your IQ4gui EQ, BTW)

Cheers!

[IIRs]

Sheffield? You an Owl or a Blade?

I'm a busy chap with better things to worry about than football.

Strange, on my hopelessly outdated but almost legendary punk-rock-laptop, the CPU hit is around 8%, which is pretty respectable for a Synthmaker plug, but only as long as my host (Reaper) is running.

When host is stopped the CPU hit goes up to 24% when compressing and a whopping 48% when expanding! Crikey! Denormals? (I have a Pentium 4m.)

Ouch! Try the update. I just dropped a standard SM denormal killer into the sidechain...

Oversampling on comp only would be nice. Perhaps automatically activated when rendering offline?

I added a separate option for the oversampling. I don't think doing it automatically when rendering would be wise however: even assuming the host realises there is now 4 samples of latency to compensate that wasn't there before (not a small assumption!) there is still some phase smear in the treble that might change the sound if used as a parallel compressor, or as individual channel comps in a multi-mic setup. Safer to leave it a manual choice I think!

I assume this is feed-forward (at least that's what it sounds like to me). It'd be smashing if you could have a switch to turn it to feed-back - that is that the ouput feeds the internal sidechain. That would make it dead sexy for emulating all kinds of classic hardware (I love all that Urei stuff).

Yes, feed forward. I thought about adding a feed back option, but that causes problems with the external sidechain option and with the expansion modes: obviously a feed-back external sidechain comp is not possible, so a logical solution would be a 3-way switch between FF, FB and Ext... but expanders get pretty unstable in a FB configuration, so I would need to exclude that option when expanding. It all seemed a bit messy to me, so I thought I would leave that for another plug, maybe more of a levelling amp style thingy with just the compression modes..?

bmanic: try the new Range knob. You were right I think, its much better now. I also increased the resolution on the Gain and Threshold knobs, though not quite as much as you requested: 0.2 dB resolution for Gain and 0.4dB for Threshold. Is that a good compromise?

Anything else will have to wait it little while: I have to do some real work now.

[s_t]

Hmm, I downloaded the 0.9.4 update, but the gui appears not be the same as in the screen shot. Oversampling button is missing, as well as 10db markers in the gain reduction meter. Knob resolutions are still 1dB also.
cheers

[IIRs]

Check the .dll file, it has the version number in the file name. I'm guessing you still have the old version: perhaps the zip file got cached somewhere and you downloaded the old one again. Or maybe you have old and new versions in your plugin folder and your host is getting confused because they have the same ID?

[s_t]

Yes, the plugin name says it's DYN4M1X_beta_0.9.4
I don't have older version in my computer

[istvankantor]

Hmm, I downloaded the 0.9.4 update, but the gui appears not be the same as in the screen shot. Oversampling button is missing, as well as 10db markers in the gain reduction meter. Knob resolutions are still 1dB also.
cheers

same here

[IIRs]

That's odd. I must have done something stupid in a hurry this morning.

Try downloading again now: http://platinumears.com/DYN4M1X_beta.zip

[A.M. Gold]

Well if bmanic says it's marvelous, it must be worth trying. I'll run it in Tracktion 2.

I still use the All Notes Off plug you made for me to stop notes when rewiring Giga 3. Thanks again for that, mate!

[IIRs]

No worries! Was that the one with the STFU button?

[A.M. Gold]

No worries! Was that the one with the STFU button?

Heh, yup! Still works.

[bootsie]

Yes, feed forward. I thought about adding a feed back option, but that causes problems with the external sidechain option and with the expansion modes: obviously a feed-back external sidechain comp is not possible,

It actually is, just have a look at the SSL bus comp schematics (or others). The simple trick is to use two VCA's and feeding the SC directly into the FB comp which's GR signal then controls the (feed forward) VCA with the actual audio input. If you have problems with that just drop me a note.

peace,
bootsy