EQ+ issue

[askewd]

Hi, firstly sorry if this has already been discussed already.

I'm posting this more as a warning really, incase anybody else runs in to the same problem i did.

So EQ+ is decent for sure - but it adds latency, without reporting the latency nor being properly compensated! This is super dangerous imo, although the devs do not see this as an issue because the latency is 'so small' In real world terms, it means that if you have duplicate tracks, some with the EQ+ and some without, it causes big phasing issue.

I was working on a mix with a tonne of stems and noticed something was off and couldn't figure it out initially, then isolated the issue down to this EQ+ problem.

So yeah, be careful.

[antic604]

Yes, it's a known issue

[askewd]

Ok thanks, that's kinda good to know. Hope they fix it. It's a big thing imo.

[Dionysos]

Is filtering duplicate tracks with non-linear-phase filters ever a safe thing to do?

It's probably caused by the EQ's oversampling, so might be quite easy to avoid by placing another instance of EQ+ on the parallel track.

[billcarroll]

Is filtering duplicate tracks with non-linear-phase filters ever a safe thing to do?

It's probably caused by the EQ's oversampling, so might be quite easy to avoid by placing another instance of EQ+ on the parallel track.

Exactly. I'd like to understand exactly what the OP is doing? Is the OP simply experiencing normal phase shift using EQ? This has been discussed here numerous times.

viewtopic.php?f=259&t=561412&p=8041337& ... Q#p8041337

Are you saying there is a bug and the entire signal is delayed due to latency introduced by oversampling, and this is not being compensated? If so, have you tested this and reported it as a bug?

From the response above from Bitwig it appears they are saying certain harmonics get delayed with EQ+ because it is not linear phase, and as expected this causes “smearing” across the sound. Therefore not good for parallel processing.

When one sound is heard across multiple tracks, a regular EQ can change the phase relationship between those tracks. This is one example when you want linear phase.

I'm not aware that there is a known issue with EQ+ latency not being compensated. Does anyone have other info?

[anoise]

I was not aware of this and used EQ+ withing mid/side splitter on the master. I could not hear a difference, but now knowing that it did shift the sides by 10 samples or so, bugs me. Anyway, the tracks are released now.
I did some tests on a kick. It not only shifts the samples in time, it also "distorts" them. Look (in the oscilloscopes) at the long ringing in the transients. Probably of the aggressive antialising filter they use.
EQ+:


EQ5 and the other EQ's:

[antic604]

I was not aware of this and used EQ+ withing mid/side splitter on the master. I could not hear a difference, but now knowing that it did shift the sides by 10 samples or so, bugs me.

I don't think it matters a lot for mid/side processing, especially for electronic music, as side - stereo, basically - information is delayed anyway. I do this all the time using EQ+ and don't hear a difference other than the mix being much clearer after high-passing the lows on the side channel...

...Anyway, the tracks are released now.

Hopefully the releases were succesful!

If not, you can now blame it on the audience disliking the phase shift on side channel

[anoise]

Damn the sabotaging phase shifters.

[Chameleonman]

Is filtering duplicate tracks with non-linear-phase filters ever a safe thing to do?

Yes it is. All the minimum phase Eq a.k.a normal Eq change the phase around the filter shape that is boosted, but it's just a color and it can be zero latency just like any other plugin where there's no time shifting going on. There shouldnt be such a thing in a Eq. I had only once a problem with Eq+. The effect chain was quite complex and the problem is not so easy to reproduce but I then tested the eq5 was ok. I love the sound of the Eq+. You can make very narrow boost and it still sounds good unlike eq5, but i don't trust the Eq+ and I don't use it anymore.

[coolblinger]

Yeah this has been a known issue with EQ+ since day one. If you want to use it to do parallel processing you need to put an instance in each chain. The pitch shifter device also doesn't report its latency (it's really small, but it's there).

[askewd]

It surprises me Bitwig Devs allows this on their own plugins. All it achieves is that I end up reaching for 3rd party EQ instead like Fab or CraveEQ because I "trust" them more. Crazy. I'd love to use EQ+ but can not be worrying about these kind of issues creeping in.

The official reply & workaround is "you can compensate the 4 samples base latency EQ+ introduces manually by putting an EQ+ on both of the parallel channels or by using a timeshift device."

[Dionysos]

I end up reaching for 3rd party EQ instead like Fab or CraveEQ because I "trust" them more

Not sure how trust factors into it... You're aware of an issue that's only relevant in a very specific situation, why not just use it when safe, and use something else (or the workaround) when it's not?

Personally I learned at some point that filtering parallel tracks isn't safe in general, so I'd never expect this to work without caveats anyway. Even if apparently it can be safe, and is, with some plugins.

[billcarroll]

It surprises me Bitwig Devs allows this on their own plugins. All it achieves is that I end up reaching for 3rd party EQ instead like Fab or CraveEQ because I "trust" them more. Crazy. I'd love to use EQ+ but can not be worrying about these kind of issues creeping in.

The official reply & workaround is "you can compensate the 4 samples base latency EQ+ introduces manually by putting an EQ+ on both of the parallel channels or by using a timeshift device."

I have yet to see someone post a bug report and acknowledgement reply on an issue from Bitwig.

Lat thread on this Bitwig replied to someone saying this is expected behavior.

If there is an issue, let's report it and get a response from Bitwig.

[askewd]

They have replied further and says it is a conscious choice, prioritising live use (so that EQ+ can be inserted into a live playing track without a 'click' issue).

As someone who mixes full time and constantly using parallel processes - and not doing live work at all - I have suggested a workaround is needed, in that it should maybe use intelligence to know if the song is playing and if so, does not compensate the latency until the user stops the song. This would solve all issues. I await their reply on this

Personally i feel that device latency has to be either zero OR needs to be compensated. One or the other. Workarounds and 'avoiding certain uses' is not ideal at all. I'll stick to 3rd party EQ's for now but it's a shame as EQ+ seems lovely.

[stamp]

They have replied further and says it is a conscious choice, prioritising live use (so that EQ+ can be inserted into a live playing track without a 'click' issue).

Who the heck inserts eqs while performing live?