Carve is in the Shop

[eusti]

Here you go: https://shop.propellerheads.se/product/carve-eq-ducker/

via Facebook.

D.

[VolatileCycle]

Yes yes yes!

[Kategra]

Just in time for me to finish my mix!

[Chango]

This is excellent!

[challism]

I was lucky enough to beta test it.  Great device.  $25 is a really nice price, too.  Thanks Ochen K!

[neoverse]

it's impressive to say the least

[Concep]

Saw it this morning and bought instantly.  Very good price for a great device.  This will make dynamic EQ a lot easier and cleaner, and more precise than my current process of using a combinators, EQs and Compressors.  Fantastic price at $25!  Compare this thing to DQ, which is $39 and only compresses a band evenly.

[Whoopdee]

Sounds like the carve uses a LPF for the ducking effect instead of amplitude? It sounds smooth.

[JNeffLind]

WOOT! This is the rack extension of my dreams. Next purchase for sure.

[TheGoodGoodMan]

Nice little tool. I like it!

[selig]

Just getting a chance to check this out, and my first impression is WOW, this is one sensitive beast! With Amount set to only 4% it's already cutting 6 dB. I'd probably only typically use 6-12 dB of cut (12 dB cut falls between 6 and 7 %) for an effect like this, which for me leaves 93% of the Amount knob unusable for me. The display barely shows anything happening at what would be the max I'd use it for in a typical setup. would be nice if the Amount read in dB IMO. Am I crazy, or missing something obvious here?

Here's what the default 66% "Amount" looks like with a sine wave @ 130.8 Hz fed into the REF inputs -  I don't think I'd ever need 40 dB of cut, and I wouldn't want it to look like this with a simple sine wave input - almost everything below 1 kHz being cut by 10 dB or more:
 

Also, there seems to be a bug with how the frequency is detected. When I play a low C (@ 65.4 Hz) I see TWO frequencies being cut, one around 44 Hz and the other around 88 Hz (amount set to 7%):


Compare that to raising the frequency one octave to 130.8 Hz, which looks ok even though it's wider than I'd expect:
 

Below that things get a bit off - here's what it looks like when lowering the octave down two (to 32.7 Hz), which as you can see is NOT 32 Hz but closer to 44 Hz (which we saw in the first graphic - related?):


I don't know if this is a known issue or not, but thought I'd point it out here to see if anyone else noticed anything odd. 

Also note that in all my examples I'm using a sine wave, as one might use for a sub bass patch. But the resultant cut is quite wide IMO, and I would assume this is the most narrow it would ever get. For comparison, here's Carve in purple and MClass EQ in yellow set to it's most narrow setting (Q=32 and with 15 dB of cut): (for reference, the two curves more closely match when MClass Q is at it's default Q setting of 5.7)
     

Hopefully I'm understanding this device correctly, and have things setup as directed. If anyone can point out any mistakes in my analysis, I'd be grateful!

[Whoopdee]

My only question is, why wasn't Selig hired to beta test this device, or any device.

[Arrant]

It feels buggy. And oversensitive. Not useful to me in its current state unfortunately.

[submonsterz]

I did not try this out as I only get RE`s that are sdk 1.
I`ve not upgraded my reason account with all RE`s in.
But I still prefer the dynamic Mclass route its more precise
and I like being able to use cv and audio triggers to drive the dynamic cuts via the Mclass method.
Also I never really have had the need to actually very often take more out than just the fundamental frequency.

[joeyluck]

I did not try this out as I only get RE`s that are sdk 1.
I`ve not upgraded my reason account with all RE`s in.
But I still prefer the dynamic Mclass route its more precise
and I like being able to use cv and audio triggers to drive the dynamic cuts via the Mclass method.

From what other folks around here have been saying, all REs from here on out are SDK2. 

[submonsterz]

I did not try this out as I only get RE`s that are sdk 1.
I`ve not upgraded my reason account with all RE`s in.
But I still prefer the dynamic Mclass route its more precise
and I like being able to use cv and audio triggers to drive the dynamic cuts via the Mclass method.

From what other folks around here have been saying, all REs from here on out are SDK2. 

well I got a problem not gonna keep upgrading different licenses I have no biggy. .

[ochenk]

Hey Selig. Thanks for the analysis.

There are a few things going on that are probably affecting your tests. 

First, the amplitude of any cut is a factor of both the amplitude of the input REF signal and the Amount knob. The Amount  knob scales the REF input from 0% to 300%, so to make a one-to-one cut, set the Amount to about a third. So, if there's not much headroom in the Amount knob, trim back the volume of the REF input signal.

Second, the more extreme the cut, the wider the width of the cut. And it's really easy to overblow the cut. So if you want something narrower, cut back on the Amount (or the input signal.)

Finally, when you overblow the cut, distortion can happen. Those distortions can introduce harmonics. That's probably why you're seeing the extra lobe with the sine wave.

Thanks again.

[kloeckno]

You could probably also use an M-Class Maximizer with the limiter and soft clip disabled, and just use the input gain knob to lower the volume of the reference signal if the amount knob isn't sensitive enough.

[selig]

Hey Selig. Thanks for the analysis.

There are a few things going on that are probably affecting your tests. 

First, the amplitude of any cut is a factor of both the amplitude of the input REF signal and the Amount knob. The Amount  knob scales the REF input from 0% to 300%, so to make a one-to-one cut, set the Amount to about a third. So, if there's not much headroom in the Amount knob, trim back the volume of the REF input signal.

Second, the more extreme the cut, the wider the width of the cut. And it's really easy to overblow the cut. So if you want something narrower, cut back on the Amount (or the input signal.)

Finally, when you overblow the cut, distortion can happen. Those distortions can introduce harmonics. That's probably why you're seeing the extra lobe with the sine wave.

Thanks again.

Hey Ochenk, thanks for responding here - I realize I should have probably posted this directly to you rather than here, sorry about that!

I fed it typical signal levels of peaking around -12 dBFS, fwiw. I would have to lower the levels quite a bit to get a useful amount of reduction for me (6-12 dB) since currently that amount of reduction happens with an Amount of around 7%. Any way to get this control to read in dB (would be infinitely more useful, since most EQs use that scale and this is essentially an EQ IMO)? With an input of -12 dBFS and the Amount set to 33% as you suggest, I'm getting over 40 dB of reduction - far more than I would ever use for "carving"! I would imagine most folks will present similar if not higher levels to the REF input, right?

BTW, what do you mean a "one-to-one cut"? I can't imagine someone needing anywhere near 30-40 dB of reduction in any situation, especially as you mention the cuts become wider as they get deeper. I am finding it difficult to dial in precise amounts of reduction because one percent change equals more than a dB of reduction (and there isn't a constant relationship between percent and dB that I can perceive). 

Finally, the bug I mention was noted when simply sending in a different frequency, and with only 12 dB of reduction. It was fine at other frequencies, meaning it's not distortion causing the double dip. Plus there the inaccuracy below 60 Hz that I also mentioned - above that range things are spot on!

Thanks for listening, and good luck with your new device!

[EnochLight]

My only question is, why wasn't Selig hired to beta test this device, or any device.

Selig is an RE dev.  I'm not sure how common it is for other devs to hire other devs, especially if a product being tested could be used to (ahem) "inspire" another dev?  

[Shokstar]

I´ve found a bug too!

I was playing a sinewave note a C5. on the spec you will see the sinewave at 1,05Khz. now I turned down the LP to 345Hz the REF input was connected with a rex drum loop. the amount of Carve was set to 100%! now you can see freq. below 1,05Khz are added to the main output. this is also visible on an amount of 2%!

take a look at this screenshot please:   

[Shokstar]

Jason you have filthy amount of RE's, bro!

yes, I have 162 RE´s now everyday I Discover a new RE

[ochenk]

Selig,

The Amount knob can't be db. It doesn't make sense for it to be, which makes me think that you might not quite get what this device does. I'll try to explain.

Let's say we have perfect white noise at unity, so every frequency at 0db. We're going to EQ that. Our second signal is two sine waves an octave apart. The first sine at unity at 440Hz, so 0 db. The second sine half of that. Let's say we're using a scale of -48db as silence and 0db as unity. So the second sine is at -24db at 880Hz. Without any scaling, if we used the frequency and amplitude signature of the sine waves to cut the white noise, the white noise would have a cut at 440Hz of 48db and a cut at 880Hz of 24db. So is that a 48db cut or a 24db cut? It's neither. It's 48 bands of cuts, each at unique dbs. That's with no scaling. With a scaler of 50%, those would be cuts of 24db and 12db respectively. The Amount knob is that scaler. It's not a db cut value. It's a percentage. It's range is 0% to 300%. That's why if you're playing a sine wave at unity (i.e. 48db) and you only want to cut 3db, your scaler would have to be at 6.25%. 

So I'm assuming the next question would be "Wouldn't a scale from 0% to 10% be better than a scale from 0% to 300%?" In my experience, I don't think so. I've been using Carve on a bunch of productions since November (when it was getting stable enough to be used in real work), and in practice, when I want to target something like a kick drum as a reference, peaks might float between -12db and -24db, but my bass instruments are usually much closer to unity. So if I have a -24db signal, a gentle cut of 6db requires Amount at 25%. A more aggressive cut of 12db needs 50%. Other parts of the signal, like hi hats, will sit even lower, requiring more extreme cutting.

So what to do when you have something like the white noise hypothetical and you need very gentle cut on a unity signal? Send the REF audio through a spider. Send one out of the spider through a gain stage (like your awesome Gain) or a compressor, take the level down, and feed it into Carve REF in. The lower the REF peak, the larger the usable range is on the Amount knob.