Tilt EQ options: Boz T-bone vs SlickEQ

[jens]

Something like this happened more that just once to me, so I usually try to avoid plugins which introduce latency,

Then where is the problem to simply deactivate it or nudge the track once you realized this? Instead of wasting a good take, I'd go this round to be honest.


Also... a lot of plugins introduce latency. Guitar Amps or Cabinets included. And most of the time, you can work around it.


Then again, my 2c.

You can't deactivate it after the recording has already been made - but yes, maybe you can nudge the track (which is what I tend to do in such a case), but first of all the performance will typically suffer to some varying degree and then it's of course a hassle I prefer to avoid.

Normally not using amp-sims or cab sims here... I prefer the real deal - Using amp sims is so 00's


On topic: just purchased a T-Bone license, despite its eye-cancer inducing GUI - what the f**k was I thinking?

[Compyfox]

Well... or you could turn off the "monitor" modes of your host and still record in real time. Er... yeah.

Actually, using amp sims is so 2010's.
But I get the point.


Still... those who are on Windows/Mac, and have suitable bridge tools, could use TAL USEq for this purpose. Especially while tracking. Or... maybe utilize the tone dials at the module in question prior to recording... *cough*

[jens]

Yesyesyes... all kind of valid...

but still, while we're at it: I mean, I'm a huge Bootise-fan, however - maybe just because I am not a coder - I can't see a reason why there should be any plugin in the entire universe that doesn't offer the option to only use oversampling during offline-rendering... I mean: really - WTF? (and it's of course not just the latency, it's cpu hit as well)

Quinto and Aleksey, they show us how it's done - most other developers though... and I just don't get it.

[Compyfox]

SKnote is NOT CPU friendly with his "forced OS" schemes with his last couple of compressor releases. The idea that a lot of users had with RoundTone and StripBus regarding offline rendering sadly never found it's way to the other hardware emulations.

Voxengo is a world for itself, mostly digital though.


Christian Budde once told me in a mail conversation that you can only have two out of three things:
- low latency
- Great Sound without cut ins
- Low CPU usage

To me, SlickEQ is low on CPU, sounds great but has latency.
SKnote's GTS-39 sounds great, has no latency (IIRC, haven't used it in a while due to reasons), but has an insane high ASIO/CPU load due to the fixed 8x OS.


It really depends on if you want to go the extra step to actually support a lot of hosts with letting your plugin recognize "wait, now I'm in realtime mode, now I'm in rendering mode". Most developers won't go that route and rather say "CPU's are cheap and fast evolving".

YMMV.

[antithesist]

I'd imagine Paul Wolff probably got the idea from Quad's decades earlier preamps.

That's where I first remember seeing tilt tone controls when I was a kid in the hi-fi shop.

[overhishead]

'tilteq' by Liteon works great, with negligible CPU consumption.

[jens]

Christian Budde once told me in a mail conversation that you can only have two out of three things:
- low latency
- Great Sound without cut ins
- Low CPU usage

well, that's going sideways now - actually it doesn't really answer my point at all.

Maybe the lack of oversampling prevents the "great sound" - well then I would typically prefer to not have the "great sound" during mixing but instead only during offline rendering. That was my point.
Why do most plugins don't offer the option to use oversampling only during offline rendering?
Coders tend to be very bright people (and that goes especially for Herbert and Fabien), so there should be a valid reason for this. I just can't see it as I know some plugins do it and it works perfectly well (of course only if the host reports the offline state to the plugin).

[Compyfox]

I seriously can't answer your question on that behalf, jens.

But I think we both hinted at one important issue to the equation:

I just can't see it as I know some plugins do it and it works perfectly well (of course only if the host reports the offline state to the plugin).

That, and extra work on actually implementing that feature. So devs save on some time and say "its good right out of the box" rather than implementing something that is not necessarily needed.


In case of SKnote again, the GTS-39, STA-Limit and C165a VCA all have fixed 8x OS without a possibility to use a lower mode in real time and a higher mode while rendering. Quinto told me several times on KVR that he insists on the 8x OS due to the sound. So here we also have the other extreme: developers who don't give a funk. So we have great sound and low latency, but high CPU/ASIO usage.

Also... SlickEQ has no latency anymore as soon as you exceed 176.4kHz. Else we talk about 4,14ms on 44kHz, and 1,94ms at 88.2kHz. Others drop the latency once you exceed 96kHz. On top of that... one has to see for what purpose this plugin was made.



And now we're kind of walking in circles.

[vladg]

I can imagine 2 cases:

1. A plugin sounds like crap without OS and sounds good with OS
2. A plugin sounds good without OS and sounds excellent with OS

OS for offline render is useful for 2nd case only.

BTW, some time ago I was surprised, Reaper by default didn't report offline state to a plugin (there's a checkbox in settings). I think they did it for some meaningful reason...

Another interesting question is how OS degrades audio quality by itself. We (in SlickEQ) did great OS implementation but at cost of latency & CPU. Lower-quality oversampling would seriously degrade audio quality.

A good quote from Melda MDynamicEQ manual:

Upsampling also has several disadvantages you should be aware of before you start using it. Firstly, upsampled processing induces latency (at least in high-quality mode, although you can select low-quality mode in the plugin settings), which is not very usable in real time applications. Secondly, upsampling also takes much more CPU power, due to both the processing being performed at a higher sampling rate (for 16x upsampling at 44kHz, this equates to 705kHz!), and the complex filtering. Finally and most importantly, upsampling creates some artifacts of its own and for some algorithms processing at higher sampling rates can actually lower the audio quality, or at least change the sound character. Your ears should always be the final judge.

So if a developer says "use your ears", it's better to provide optional oversampling support. If a developer says "I know what I'm doing, my plugin sounds crap without OS but sounds great with 8x" he uses mandatory OS then.

[Obsolete236871]

It is my feeling that some of the VOS plugins (before TDR VOS) had this thing going on, where the plugin only make use of oversampling when using them with lower sampling rates (eg 44.100). So Bootsy and others encouraged using higher sampling rates (which can increase the general CPU-consumption, if I'm not mistaken). Any thoughts on this?

[pscorbett]

And looking at how many Tilt EQ's exist these days, and when they all originated, it makes me wonder who "jumped on the bandwagon" and who not. Because before the flood, nobody even cared. Same with Baxendall type EQs.

Yup, FabFilter Pro-Q 2 has tilting options. Completely different price range, but it was nice to see it included!

[itneveris]

I really enjoy Sonimus Softube

[Ichad.c]

A good quote from Melda MDynamicEQ manual:

Upsampling also has several disadvantages you should be aware of before you start using it. Firstly, upsampled processing induces latency (at least in high-quality mode, although you can select low-quality mode in the plugin settings), which is not very usable in real time applications. Secondly, upsampling also takes much more CPU power, due to both the processing being performed at a higher sampling rate (for 16x upsampling at 44kHz, this equates to 705kHz!), and the complex filtering. Finally and most importantly, upsampling creates some artifacts of its own and for some algorithms processing at higher sampling rates can actually lower the audio quality, or at least change the sound character. Your ears should always be the final judge.

So if a developer says "use your ears", it's better to provide optional oversampling support. If a developer says "I know what I'm doing, my plugin sounds crap without OS but sounds great with 8x" he uses mandatory OS then.

Good point. Most people forget that developers stare at different plots alot(maybe too much), all/most artifacts are considered before a single line of code is written, very tough decisions need to be made. Otherwise we would have just all used 32x oversampling with 10000tap latency inducing linear phase filters, minimized for less pre-ringing, and made the cpu load the end-users problem. Sometimes we just need a bit of trust...

Btw - from an end-user perspective I never use any "high quality offline options" because when I spent 8 hours mixing a track and it gets mixed down - it should sound 100% the same as when I mixed. Imagine some plugins create a tad of aliasing that bunches up around 10kHz - you hear the tops are too harsh so you dip 3dB out, now you "HQ offline render" - the aliasing is removed - so no harshness but your prior EQing decision is still there, so now your tops are theoretically 6dB less and too dark... I prefer real-time rendering - call me old-school if you must.

Not that I'm against having different quality setting on plugs - does come in handy if you need it for a live situation; and for non-linearities elements - the Dev can only guess how hot the signal is that your sending into the plug-in.

Regards
Andrew

[Fritze]

Recently I read about tilt eqs and wanted one. Then I found out I already have one: preFIX has one. And it sounds great.

[vladg]

Good point. Most people forget that developers stare at different plots alot(maybe too much), all/most artifacts are considered before a single line of code is written, very tough decisions need to be made. Otherwise we would have just all used 32x oversampling with 10000tap latency inducing linear phase filters, minimized for less pre-ringing, and made the cpu load the end-users problem. Sometimes we just need a bit of trust...

The developer (or team of developers) usually acts as 3 different persons:
#1. Scientist. The guy all about numbers, perfect curves and models.
#2. Audiophile. The guy that tells you that the sound is not to "fat" or not "warm" or too boring.
#3. User. This guy bothers about usability (and CPU and latency are about usability too). For example guy #3 can say to guy #1 than 100 parameters he able to control is not cool it he want to control not more than 10 of them and so on

So the final product is a kind of compromise between these 3 guys.

Btw - from an end-user perspective I never use any "high quality offline options" because when I spent 8 hours mixing a track and it gets mixed down - it should sound 100% the same as when I mixed. Imagine some plugins create a tad of aliasing that bunches up around 10kHz - you hear the tops are too harsh so you dip 3dB out, now you "HQ offline render" - the aliasing is removed - so no harshness but your prior EQing decision is still there, so now your tops are theoretically 6dB less and too dark... I prefer real-time rendering - call me old-school if you must.

I dislike quality settings too! Indeed, when mixing the song for several hours the offline render is not the final step at all. You're just sitting in the silence for 5 minutes waiting for mixdown to complete then listening to the result and then you're working on this mix again. So it's just a time for a short rest. Now imagine you're mixing with musicians behind your back. You tell them, guys the mix sounds like sh*t but that's all just because I use low quality processing mode in plugins... Then they sit in silence for 10 minutes waiting for high quality render to complete and then... you're mixing again. That's why I prefer online render mode because you get absolutely what do you hear.

Not that I'm against having different quality setting on plugs - does come in handy if you need it for a live situation; and for non-linearities elements - the Dev can only guess how hot the signal is that your sending into the plug-in.

Speaking of SlickEQ, the new version is on the way and it features alternative processing mode with almost zero latency and reduced CPU usage. Acting as those guy #3 I see this mode intended to use on tracks while the original processing mode for buses but really the final decision is for user.