Rupert Neve designs Portico summing mixer

[warhead]

Originally Posted by jjdpro

Well, i agree with the previous poster.. Some mag just did asumming shoot-out usunf the 5k summing boxe, mackie onyx mixer, and an apogee.. the conclusion was startiling. The difference btw the high-end and low end was not wide at all. Some testers even prefered the mackie onyx mixer..

Little known fact about the Mackie Onyx, there is a pad on the output of the mixer that brings it down to microphone level. So you could sum through the Onyx (which has tons of headroom, very nice 4 band eq's, 6 auxes etc) and then plug it into any preamp of your choice for makeup gain.

Kinda like somebody knew it might be used this way, although when talking to their guys they mentioned it was so if you were sub-grouping drums etc live you could plug it into another mixer's preamps if needed.

Either way it's functional if you want to choose makeup gain pres.

The preamps on board are just a little step away from designs Amek and others have used in the past (9098 etc), and sound very nice top to bottom. Talkback capability and the ability to have full monitoring while tracking as well...all add up to a serious little mixer.

No, you can't push it for "color", but you can't push some of the other devices on the market sold solely as summing devices. The only real con to some would be the name Mackie on the face of it, not high end enough.

War

[murp]

Originally Posted by warhead

Little known fact about the Mackie Onyx, there is a pad on the output of the mixer that brings it down to microphone level. So you could sum through the Onyx (which has tons of headroom, very nice 4 band eq's, 6 auxes etc) and then plug it into any preamp of your choice for makeup gain.

The way you worded that makes it sound to me like it goes into the summing amp, then it's padded down by resistors, which I'd imagine wouldn't be of any use to anyone. What would be useful is if the switch bypassed the summing amp altogether - or is that what you meant?

[NathanEldred]

Originally Posted by Brad McGowan

Mixing OTB is all about the things that Nathan and a couple others have cited: phase shift, intermodulation distortion, harmonic distortion, etc. Once you figure out that you can add those things without using a mixer or "summing box" then you'll be well on your way to a more "analog" sounding mix.

Sure, you're right in many ways. I've heard the difference for myself of the same techniques/skill set, etc of guys using ITB and OTB (who know their stuff on both platforms adhering to proper analog or digital gain staging), both on high quality gear...using outboard also ITB (by sending out, processing, then printing). By using the analog outboard, it sound FAR better than using plug ins IMO. But when hearing the difference between the console mixes and the ITB mixes were noticeable. The ITB, on their own merits, were as good or better than most people's than I've heard. The console mixes were ear candy, a definite step up (again, same engineer, same techniques, even much of the same outboard).

So just to me, in my own experience, this verifies that it's something beyond just noise, phase shift, distortion & digital headroom issues. Maybe it's just because PCM conversion makes the original analog signal smaller sounding (tonally speaking...and it does, even with the best conversion, it's like a cold day at the beach in speedos for the analog when transferring from the 2" 16 track...serious shrinkage). So one conversion out via D/A, one back in to the DAW via an A/D.

But that still doesn't explain it, because one conversion in via an A/D, then mixing via multiple D/A's into a console and then converting the stereo master one more time is the requirement for mixing on a hybrid DAW/console system. That's one more conversion than the ITB/outboard gear method. Not trying to be argumentative, just kicking some ideas and opinions around. It's just that every time the two are compared, even in the best circumstances, the OTB wins for me in an obvious way. It also seems a lot more convenient with the DAW/analog console system. The integration with the outboard gear is a lot easier, not to mention you get console EQ in most cases, auxes, etc, etc. Is ITB generally prefferred by some engineers because of convenience, cost, or physically space issues, or is it truly that they feel it's a sonic improvement over the alternative?

[Kiwiburger]

Originally Posted by Duardo

You've actually got it backwards...you don't start at the bottom and count up, you start at the top and count down. Whether you're mixing on a 24-, 16-, or 8 bit system, 0 dBFS is the loudest you can get and if you take a 16-bit file that peaks at 0 dBFS and import it into a 24-bit system it will still peak at 0 dBFS and be just as loud as it was in the 16-bit system...no less and no more. More bits don't let you get louder, they let you get quieter.

In addition, most DAW users don't even come close to using the full 144 dB dynamic range they have abailable theoretically, or even 114 dB, or 80, or 60, at least not as far as popular music is concerned...put in most commercial CD's and you'll see that the meters probably don't go more than 10 dB down. It is true that there would be artifacts if you tried to push a signal that truly utilized 144 dB of dynamic range through a 114 dB "hole" but they'd all be so quiet you wouldn't hear them.

-Duardo

Thanks Duardo, but I still stand by what I said.

Consider this torture/stupidity test: create a sinewave that peaks at 0dBFS, and then gate it on and off with a plugin. See how your speakers like that.

In that case, the gated silence is digital black - 000000000000000000000000. The peaks of the sinewave would be full 24 bits - 111111111111111111111111.
24 bits represenst 144dB dynamic range.

Now you said that "most DAW users don't even come close to using the full 144dB dynamic range" - but I disagree. This torture test is using the full 24 bits - the 144dB range. And this is not very different to what a lot of DAW users do with soft synths, samples or gated effects.

I know what you are saying about over-compessed mixes - that is a different issue. You are talking average peak levels - i'm talking about wave cycles.

What I am suggesting is that many DAW users are using the full 144dB mathemically possible range (by peaking at 0dBFS), and that this is very extreme and unnatural.

Your converters have published dynamic range specs of say 114dB. That to me suggests that there is some 30dB range that is masked by noise and distortion - and very probably makes the sound more acceptable to our ears.

I agree with the people who suggest that our DAW meters are calibrated wrong. The are mathematically correct, but they pschologically causing DAW users to mix too hot. Especially since the top 6dB usually occupies half of the wave graph, or half of the fader range. It makes it seem 'wrong' not to get your audio peaking into that top 6dB.

I'm not disputing that OTB can sound better (or a lot worse, depending on gear and skill). I'm just suggesting it is all about noise and distortion. And in theory that could be done digitally.

Incidentally - I also don't agree with the theory that modern music has no dynamics, so therefore poor converters are acceptable. My thinking, is that if you are going to smash everything with hyper compression, the noise floor of your A/D converters should be ultra low, otherwise you will hear it. Therefore, you should choose the widest dynamic range if you plan on making undynamic music. That's my theory anyway.

[John Suitcase]

For me, ITB is my standard operating procedure, but OTB is something I'll do if time permits and I feel the project will benefit. I'm confident that I get very good ITB mixes, but sometimes, particularly on complex rock mixes, going OTB does add some clarity. Again, this may be a headroom issue (it's hard to imagine how much headroom is needed when you're summing 40 or 50 tracks), or it may be something with pleasant sounding distortion, or even just the color of the board I'm using.

I guess I don't believe one is "head and shoulders" over the other, but in some situations, the extra effort is worth the results.

Generally, I don't use outboard compression or eq after I track, simply because I want my mixes to be recallable, should the client want "a smidge more guitars" or whatever...

[mixerguy]

Originally Posted by Brad McGowan

... (snip)..... Mixing OTB is all about the things that Nathan and a couple others have cited: phase shift, intermodulation distortion, harmonic distortion, etc. Once you figure out that you can add those things without using a mixer or "summing box" then you'll be well on your way to a more "analog" sounding mix.

Brad

Have you ever done a mix on a really good analog console, then repeated the mix ITB?

Just curious.

I have; the analog console mix smoked the ITB mix.

[warhead]

Originally Posted by murp

The way you worded that makes it sound to me like it goes into the summing amp, then it's padded down by resistors, which I'd imagine wouldn't be of any use to anyone. What would be useful is if the switch bypassed the summing amp altogether - or is that what you meant?

It's padded down by resistors before the main L/R output.

War

[Brad McGowan]

Originally Posted by NathanEldred

But that still doesn't explain it, because one conversion in via an A/D, then mixing via multiple D/A's into a console and then converting the stereo master one more time is the requirement for mixing on a hybrid DAW/console system. That's one more conversion than the ITB/outboard gear method. Not trying to be argumentative, just kicking some ideas and opinions around. It's just that every time the two are compared, even in the best circumstances, the OTB wins for me in an obvious way. It also seems a lot more convenient with the DAW/analog console system. The integration with the outboard gear is a lot easier, not to mention you get console EQ in most cases, auxes, etc, etc. Is ITB generally prefferred by some engineers because of convenience, cost, or physically space issues, or is it truly that they feel it's a sonic improvement over the alternative?

You make some interesting observations, Nathan. Speaking for myself--I have a mixerless studio because of physical space limitations. And the convenience of instant recall is quite nice. I would never make the claim that mixing ITB is sonically better than mixing through a console. It's just different. I personally have never had the opportunity or luxury to do a direct comparison with any material I've been working on. I just try to make the best music with the tools I have at hand. I'm sure many of us here are in the same boat. The goal is to find a set of tools that helps you create the results you envision in your head. Some people like Snap-On while others actually prefer Craftsman.

Brad

[Brad McGowan]

Originally Posted by mixerguy

Have you ever done a mix on a really good analog console, then repeated the mix ITB?

Just curious.

I have; the analog console mix smoked the ITB mix.

Oh I wasn't suggesting at all that ITB was better or worse than OTB. I was just making the point that it's possible to create a good sounding mix ITB if that's what you have to work with. The challenge is figure out how to add the distortions the mix is missing. I don't claim to know that trick.

And just to play devil's advocate: "smoked" is matter of preference. I bet that many classical and jazz guys might make the claim that the ITB "smokes" the OTB mix for a given tune.

I personally have never had the chance to perform the experiment you are describing. I'm 98% positive that if I did I probably might have the same feeling as you. But I'll likely never have that opportunity so my time is best spent trying to make the most of the medium that I have to deal with on a daily basis.

Brad

[not_so_new]

Originally Posted by NathanEldred

Not trying to be argumentative, just kicking some ideas and opinions around. It's just that every time the two are compared, even in the best circumstances, the OTB wins for me in an obvious way.

So Nathan my question about that would be.....

Do you think this is the analog summing buss or the difference between the physical vs. virtual faders causing the improvement in your eyes (ears)?

That is to say, discounting ITB and external EQ / compression are the benefits of something like a Folcrom causing the good vibrations or do you think it is the faders on a full blown console making the difference?

I have my opinion on this already but I am wondering what other people think, open to having my mind changed.

For the record I do think that you can get a good mix ITB and if that is how you work cool, but I do believe that hardware tools like compression and EQ are just slightly ahead of (or in some cases miles ahead of) software tools, the gap is closing to be sure.

[NathanEldred]

Originally Posted by not_so_new

So Nathan my question about that would be.....

Do you think this is the analog summing buss or the difference between the physical vs. virtual faders causing the improvement in your eyes (ears)?

IME and IMO it's the physical process of the mix being moved OTB. It has nothing to do with physical faders. You can do all your automation moves in the computer, as long as it's going out to the console or summing box, you should hear a difference (and to be preemptive, this is under the assumption that the console or outboard gear being used is of high quality).

[Duardo]

Consider this torture/stupidity test: create a sinewave that peaks at 0dBFS, and then gate it on and off with a plugin. See how your speakers like that.

In that case, the gated silence is digital black - 000000000000000000000000. The peaks of the sinewave would be full 24 bits - 111111111111111111111111.
24 bits represenst 144dB dynamic range.

Actually, that's not the case...digital black doesn't equal -144 dBFS, it's theoretical number that's essentially minus infinity. Digital black is the same whether you're talking about a 16-, 24- or 8-bit system. You'll never "hear" it because your converters have noise at a level that's much higher than that.

Now you said that "most DAW users don't even come close to using the full 144dB dynamic range" - but I disagree. This torture test is using the full 24 bits - the 144dB range. And this is not very different to what a lot of DAW users do with soft synths, samples or gated effects.

This test isn't using the full 24 bits and isn't practical in any case. When we talk about dynamic range we're talking about the difference between the highest and lowest levels of the signals we're recording. Nobody's recording digital black. Even if you use gated effects there are still other tracks present in the mix so the level of the mix never drops down to digital black. And if you want to put forth a purely theoretical example of a mix where the entire mix is gated to digital black again, digital black isn't -144 dBFS, or any real number...you'll only hear the "silence" be as quiet as the converters you're listening through.

What I took issue with was your statement that "Most DAW users mix using the full 144dB range. In other words, they mix right up to 0dBFS" which is simply not true...mixing up to 0 dBFS doesn't say anything about how much dynamic range you're using. How quiet the quietest signal in your mix is does. More bits lets you record quieter, not louder.

What I am suggesting is that many DAW users are using the full 144dB mathemically possible range (by peaking at 0dBFS), and that this is very extreme and unnatural.

They're only using the full mathematically possible range if their music extends way down in level. Classical music probably comes closest as the dynamics vary so widely.

Your converters have published dynamic range specs of say 114dB. That to me suggests that there is some 30dB range that is masked by noise and distortion - and very probably makes the sound more acceptable to our ears.

It's not likely that you'll ever hear the noise in your converters. Your ears can only handle a dynamic range of, say, 120 dB. So if you're listening to something that peaks at 0 dBFS, and your system's set up so that equals, say, 120 dB SPL (which is approximately what the human threshold of pain is, although it varies with frequency) then even if you were to have a theoretically perfect converter you wouldn't be able to hear anything below 0 dB SPL, or -120 dBFS. And that's just a theoretical case, as temporary threshold shift would even further the dynamic range of signals you'd be able to hear. Believe me, people are not even coming close to using that dynamic range.

Incidentally - I also don't agree with the theory that modern music has no dynamics, so therefore poor converters are acceptable. My thinking, is that if you are going to smash everything with hyper compression, the noise floor of your A/D converters should be ultra low, otherwise you will hear it.

In most cases the noise floor of the converters will be a non-issue as the level of the source you're recording will be much higher. Chances are if you're recording a signal that even comes close to the noise floor of your converters you're either doing something wrong or are recording a signal with enough dynamic range that you're not going to hypercompress everything. But it certainly is something to be aware of.

it's hard to imagine how much headroom is needed when you're summing 40 or 50 tracks

Not really...you need 6 dB every time you double the number of tracks...

-Duardo

[John Suitcase]

Originally Posted by Duardo

Not really...you need 6 dB every time you double the number of tracks...

-Duardo

Ok, so now you have my ear. Let's say I have one track. I don't need any headroom, correct?

But if I add a track, I now should have 6db of headroom, 2 tracks would be 12, and so on:

1 - odb
2 - 6db
4 - 12db
8 - 18 db
16 - 24 db
32 - 30 db
64 - 36 db


I assume that you're saying that my master fader meter (ITB) should be peaking at the indicated level?

[Brad McGowan]

Originally Posted by John Suitcase

Ok, so now you have my ear. Let's say I have one track. I don't need any headroom, correct?

But if I add a track, I now should have 6db of headroom, 2 tracks would be 12, and so on:

1 - odb
2 - 6db
4 - 12db
8 - 18 db
16 - 24 db
32 - 30 db
64 - 36 db


I assume that you're saying that my master fader meter (ITB) should be peaking at the indicated level?

I think what he's saying is that if you have one track and it peaks at 0 dBFS and then you add a second track that peaks at 0 dBFS then you need to pull both tracks down 6 dB to keep your master faster from clipping 0 dBFS.

Someone correct me if I have this wrong.

Brad

[Brad McGowan]

FWIW...I just emailed Steinberg and recommended to them they should add a metering scale to their GUI that correlates to all the outboard analog gear we use. I got a note back saying he would pass the recommendation along.

Everyone should contact their favorite DAW manufacturer and make this same recommendation.

Brad

[Duardo]

I think what he's saying is that if you have one track and it peaks at 0 dBFS and then you add a second track that peaks at 0 dBFS then you need to pull both tracks down 6 dB to keep your master faster from clipping 0 dBFS.

Right...and then if you add two more tracks it's another 6 dB...four more, another 6 dB...etc. So if you have a 12-track mix that's peaking at 0 dBFS and you bumped it up to 24 you'd need 6 dB more headroom, and another 6 dB if you bumped it up to 48.

-Duardo

[John Suitcase]

OK, that makes more sense.

It still leaves the quandary of whether your mix should peak at 0 db, or well below (per the advice of others on this board.)

Also, let's say I have a mix and it's hitting 0 or above and I turn down the master fader. Does that solve anything? Is the master fader reducing all of the channels before summing them, or simply lowering the output level after the fact (thereby still clipping but being "quieter")?

From what I understand from reading other posts, you're best served by keeping the individual tracks from hitting 0db (having them peak at -10db is ideal), and making sure that the master channel has several db of headroom.

What about using group channels?

[BorrelNote]

Originally Posted by AnalogBob

. but we have yet to make a musical sounding mixing "engine" for our DAWs....The modern daw engine is clean clinical bt not musical, and I think it will be quite some time before this gets adressed. Besides...I shudder when I think abou how old the mix engine is in protools or logic. Sure I got all these new hot plugins that sound great on there own....put em together and well its just not right. Where is the weakest link in audio right now? We have the sum of the parts but not the whole......

I think there already is such a musical sounding mixing "engine" and it's called Sawstudio.... Look past the initial appearance of the DAW (there are good skins for it) but this software has the musical sounding engine you may be looking for...I know I am...Just saving up for it since I invested in another DAW before I found out about Sawstudio...
Oh, and the 'onboard' effects sound GREAT as well...there are some videos you can download on the sawstudio site... fun to watch.. And I am not affiliated with Sawstudio as a disclaimer.

[Kiwiburger]

OK - just found the flaw in my logic. 24 bit audio does not mean that the waveform is divided into 16777215 discrete steps, as basic digital audio theory would seem to suggest. It seems that there is a bit of overhead, and only about 20 bits are used to describe the waveform.

20 bits x 6dB per bit = 120dB - which explains to me why the best converters only have a dynamic range of around 114dB. The missing 6 dB is easier to explain as noise than 30dB.

Of course consumer D/A has a much more restricted dynamic range than the best studio converters. So I still don't think it's wise to slam everything in the top 6dB, even if that's what everybody does. But i'm now happier about peaking at 6dB.

[Brad McGowan]

Originally Posted by John Suitcase

OK, that makes more sense.

It still leaves the quandary of whether your mix should peak at 0 db, or well below (per the advice of others on this board.)

Also, let's say I have a mix and it's hitting 0 or above and I turn down the master fader. Does that solve anything? Is the master fader reducing all of the channels before summing them, or simply lowering the output level after the fact (thereby still clipping but being "quieter")?

From what I understand from reading other posts, you're best served by keeping the individual tracks from hitting 0db (having them peak at -10db is ideal), and making sure that the master channel has several db of headroom.

What about using group channels?

It depends on the math of your DAW. If you use a floating 32-bit app like Nuendo or Cubase SX then you can theoretically run the individual channels in the red and simply reduce the master fader to scale the 2-mix back so that you will not clip your D/A when the waveform is reconstructed. However, in reality you still must be mindful of intersample peak distortions (do a search to learn more). It is because of this phenomenon that we are best served by keeping our gain staging between plugins conservative as well as giving ample headroom on the master fader. Otherwise you run the risk of introducing distortions due to intersample peaks not picked up the common peak meters of most DAW's.

Brad

[John Suitcase]

Thanks for the info, Brad!

It's a confusing thing, trying to keep everything in line. I usually try to avoid overs at each stage, but with the whole "intersample peak distortion" thing, it gets a bit fuzzier, so to speak.

I work in Cubase SX, and generally have pretty good luck, although I occasionally do have a mix that gets that "behind a screen" kind of sound going on, which I usually attribute to some type of clipping. It's not that you can hear distortion, it's just that everything sounds like it's a generation away from the original, if that makes sense.

Does anyone else ever hear this? That feeling like you're hearing a recording of a recording, rather than a recording?

Am I nuts?

[Duardo]

It still leaves the quandary of whether your mix should peak at 0 db, or well below (per the advice of others on this board.)

I'd advise to have it peak well below, at least as far as tiyr final mix is concerned...in mastering they'll typically push it up there anyhow. Not that it's necessary...

Also, let's say I have a mix and it's hitting 0 or above and I turn down the master fader. Does that solve anything? Is the master fader reducing all of the channels before summing them, or simply lowering the output level after the fact (thereby still clipping but being "quieter")?

No reason that shouldn't work...it does depend on how the digital mixer is set up, but any decent digital system should provide plenty of headroom for signals "above" zero, and as long as you bring your final level down so that you're not peaking above zero (and, again, preferably a ways below) you should be fine. I know that in Digidesign's HD systems, which have a fixed 48-bit internal resolution, they leave 9 bits for signals "above" 0 dBFS (giving you an extra 54 dB of headroom) and the rest below -144 dBFS, so you can maintain full "resolution" over an extremely wide dynamic range. That doesn't mean that you don't need to pay attention to your levels, of course, as there are still plenty of places you can clip internally and it makes sense to have your faders in "normal" places, but you shoudln't have to worry about clipping the internal mixer itself. I don't know exactly how all of the other systems on the market do it but most of them have 32-bit floating internal resolution and should be able to handle the same kinds of issues with no problems. And in any case these mixers will have way more dynamic range than any analog mixer/summer. Not that that means they'll sound better...I'm certainly not saying that they do...just that that's not one of their limitations.

OK - just found the flaw in my logic. 24 bit audio does not mean that the waveform is divided into 16777215 discrete steps, as basic digital audio theory would seem to suggest. It seems that there is a bit of overhead, and only about 20 bits are used to describe the waveform.

Actually, 24-bit audio does divide the waveform into 16,777,215 discrete steps (which is two to the power of twenty-four). It's just that due to the limitations of the analog circuitry involved we can only get about 120 dB of dynamic range, so the "last" four bits or so are just "describing" noise.

Of course consumer D/A has a much more restricted dynamic range than the best studio converters. So I still don't think it's wise to slam everything in the top 6dB, even if that's what everybody does. But i'm now happier about peaking at 6dB.

Again, slamming everything into the top 6 dB has nothing to do with the restricted dynamic range of cheaper converters (don't get me wrong...there are plenty of reasons not to slam everything into the top 6 dB...that's just not one of them). Even with a consumer-grade CD player if you have everything slammed into the top 6 dB it's highly unlikely that you'd be listening at a high enough level to, say, hear the noise of the converters in the silence between tracks. What you will lose with those cheaper consumer converters is detail at lower levels.

-Duardo