Summing with Scope?

[medway]

Hi,
imho, floatingpoint summing is nice as long as singal scales are almost identical.
(For 32 bit floatingpint signals, you have a audiosignal -24 bit and a scale - 8bit)
If you are summing floating-signals with different scales (loud and quiet signals), fixed point seems to be better.



Open a project with 16 tracks or less. Drive the hell out of the mix, push it way into the red and make sure the submix clip lights come on occasionally. Don't use any paris eqs, directx or eds effects on this mix. Keep it dry and confined to faders and panning only. Make it SLAMMING. Now add another submix to the project. Take the submix with your slamming hot mix and switch it to a NATIVE submix. I'm sure you will be able to hear the difference in seconds. All kinds of gnarly nasty **** going on. “

I'm finding lots of sonic similarities between Paris and CW. The mixes sound very organic and punchy rather than bright and shiny like Pro Tools mixes, but that's good, IMHO. It's easy enough to dial in an overly bright top end with a filter if I want that.

The reason the native saturation in paris sounds nasty is becuase it uses the old method of wrapping around waveforms during overs. Some older DAWs used to do this, hence the reputation for bad sound. All daws handle overs like Paris now so this is nothing special.

[areptiledysfunction]

Hi,
imho, floatingpoint summing is nice as long as singal scales are almost identical.
(For 32 bit floatingpint signals, you have a audiosignal -24 bit and a scale - 8bit)
If you are summing floating-signals with different scales (loud and quiet signals), fixed point seems to be better.



Open a project with 16 tracks or less. Drive the hell out of the mix, push it way into the red and make sure the submix clip lights come on occasionally. Don't use any paris eqs, directx or eds effects on this mix. Keep it dry and confined to faders and panning only. Make it SLAMMING. Now add another submix to the project. Take the submix with your slamming hot mix and switch it to a NATIVE submix. I'm sure you will be able to hear the difference in seconds. All kinds of gnarly nasty **** going on. “

I'm finding lots of sonic similarities between Paris and CW. The mixes sound very organic and punchy rather than bright and shiny like Pro Tools mixes, but that's good, IMHO. It's easy enough to dial in an overly bright top end with a filter if I want that.

The reason the native saturation in paris sounds nasty is becuase it uses the old method of wrapping around waveforms during overs. Some older DAWs used to do this, hence the reputation for bad sound. All daws handle overs like Paris now so this is nothing special.

Have you actually mixed on a Paris system? There are many gain staging options in Paris that can be used simultaneiouly which are DSP based..individual channel faders, submix faders, EQ makeup gain pots, DSP based compression that can be set up as pure gain (threshold totally cranked out) and global faders, not to mention the ability to use VST plugins on the individual channels. Each of the DSP based Paris gain stages can interact with the others to produce varying sonic footprints.

There is a difference between what it can do and what a native system can do if mixed in the same manner. The difference is that while 32 bit floating point math will certainly allow this kind of mixing without distortion, it will not allow the final rendering/bounce without distortion. You either have to lower the channel/bus levels or apply compression to them to keep everything below 0dBfs. In Paris, it is the DSP itself doing this and doing it in an extremely transparent way which is, IMHO, vastly superior to the way I have to mix on a native system. It is possible for me to crank the hell out of tracks in cubase at 32 bit float and then lightpipe them directly to corresponding Paris channels and the "phatness" that can be achieved in a mix is amazing because not only am I eliminating the cipping that would happen when summing a native mix above 0 dBfs, I also have 24dB of additional headroom and massive additional options for processing and gain staging. A Paris system can sum and bounce up to 128 tracks (if using 8 EDS cards, which is also possible to do without PCI bus overloading) without the stereo image collapsing and the detail disappearing in the mix bus. The only kind of summing I have ever seen that will allow high track counts and this kind of transparency in the mix bus is an extremely expensive (and huge) analog console and it certainly didn't have 128 channels available for summing. I have received ITB Paris mixes here for mastering that were 60+ tracks already rendered at RMS levels of -12 dBfs. This is with no bus compression and it was pretty darned transparent, even at that extreme. I did send it back, BTW because there was no real sonic headroom available at that point that I could use if I needed to boost so it is possible to abuse this system, as with any other. thing is, there was no audible digital clipping.

Anyway, I appreciate your post and I'm not meaning this to sound like I disagree with your statement "in theory" but unless you have mixed on this system, it's hard to describe how it behaves other than to compare it with an analog desk due to the ability to "push/pull" the gain staging in a manner that can produce a very transparent or colored result. It really should act as you describe but it doesn't act that way at all.


Very best regards and I hope you have a happy new year,

[medway]

Hey DJ, yes I have mixed on a Paris system, since 2001. You might remember me as Jesse Skeens from the NG.

I purchased a small Paris system a year ago after having sold all my stuff previously. This time I did many tests to try and figure out what was special about it. And contrary to what many on the NG said as well as what Brian shows in the DVD I did not find anything special about how Paris saturates. The coding further supports this as it shows all Paris is doing is flat topping signals.

I was basically laughed at for being crazy to say that you can acheive the same saturation that Paris produces just by raising the master fader in Wavelab and clipping a signal. Yet a few people finally tried it and agreed. I think the whole Paris sounds better thing at this point is mental. As when I got my system last year and cranked some trims and faders I too thought "hey this sounds analog" But when I tried the same thing in other DAWs I realized they sounded just the same.

Now I won't say Paris doesnt have a sound. But for me its more due to the effects and converters as well as the ability to be able to interfae hardware devices easily. And to a point the gain structure does lend itself to some creative clipping. But this is nothing that can't be acheived with another DAW set up similarly.

Let me add that I really wanted Paris to show me something unqiue in its saturation/summing so if anything my tests were done as exhasutively as I could to find something "special".

Paris is/was a great system but in the meantime I have matched or bettered all my mixes in it with current DAWs such as SX3 and Live 6.

I'm interested to see some examples of SCOPE summing. I've always wondered if it might make a difference yet at the same time I have not heard a concinving comparison that shows summing really amounts to anything.

Jesse

[bosone]

since there are units which cost 2.000 $ and are used to analog sum up to four stereo signals... why not to build an emulation for scope that you can use twith the 8 mono bus of the stm2448? no setting, no levels... just 8 in, 2 out with some "analog/warm/fat" sound...

[astroman]

you cannot 'emulate' analog summing in the digital world, except for a few effects that aren't strictly related to summing (like saturation or softclipping).

Unfortunately the audio files of the 'analog summing comparision' (I mentioned on the previous page) aren't online.
On the other hand that (supposed to be...) objective material has so much flaws that it's rather to be considered a kind of end-of-year-paid-advertisement-review.
Even a 300 Euro box 'outperformed' a ProTools HD for a full class difference.

Math theory seems simple, but Wolf and Ch. Steinberg have contradicting viewpoints. DJ says Paris does magic (in a certain context), Medway says he could set the same performance parameters on (almost) arbitrary systems once he got his mind free from any prejudice...
it's a bit confusing

if I'd buy (for example) that cheapo 4 x 2 channels analog wonder, and feed it 4 busses from my A16, will the extra conversion stage screw the final result to a degree that I might have been better off with a proper digital summing in Scope ?
or will the little box just do it's magic thing ?
what if you open it and find nothing but 4 NE5532-Opamps and a few resistors and caps inside...

cheers, Tom

[voidar]

since there are units which cost 2.000 $ and are used to analog sum up to four stereo signals... why not to build an emulation for scope that you can use twith the 8 mono bus of the stm2448? no setting, no levels... just 8 in, 2 out with some "analog/warm/fat" sound...

The Switch-it device actually does this, without the "analog/warm/fat" part. It's just Creamware clean.

[medway]

you cannot 'emulate' analog summing in the digital world, except for a few effects that aren't strictly related to summing (like saturation or softclipping).

Unfortunately the audio files of the 'analog summing comparision' (I mentioned on the previous page) aren't online.
On the other hand that (supposed to be...) objective material has so much flaws that it's rather to be considered a kind of end-of-year-paid-advertisement-review.
Even a 300 Euro box 'outperformed' a ProTools HD for a full class difference.

Math theory seems simple, but Wolf and Ch. Steinberg have contradicting viewpoints. DJ says Paris does magic (in a certain context), Medway says he could set the same performance parameters on (almost) arbitrary systems once he got his mind free from any prejudice...
it's a bit confusing

if I'd buy (for example) that cheapo 4 x 2 channels analog wonder, and feed it 4 busses from my A16, will the extra conversion stage screw the final result to a degree that I might have been better off with a proper digital summing in Scope ?
or will the little box just do it's magic thing ?
what if you open it and find nothing but 4 NE5532-Opamps and a few resistors and caps inside...

cheers, Tom

My conclusion is that "summing" isn't an issue. I've done mixes in Logic that sounded bigger and wider than a Neve 9098i using the same exact files.

The only difference I hear in analog mixing solutions compared to digital is saturation/distortion. There is probably also subtle phase differences due to the electronics.

In the end though I worry about getting good source material and my mixing skills and not summing.

[medway]

Ok so I've now heard the file Neil made. There is deffinatly a difference and its not too subtle but I'd like to see the process to figure out whats going on.

[wolf]

Math theory seems simple, but Wolf and Ch. Steinberg have contradicting viewpoints.

We have no different viewpoints beside the nonsense, that pan law is the cause of difference between different mixing engines (i.e. - 3dB panlaw just removes the need to readjust the volume while panning left/right). Afair Charlies answer was mainly to "defend" the cubase/nuendo mix engine against Logic (which is nonsense anyway because both are 32bit float, but hey .. it's all about marketing), so he doesn't really talk about float vs. integer.

DJ says Paris does magic (in a certain context), Medway says he could set the same performance parameters on (almost) arbitrary systems once he got his mind free from any prejudice...
it's a bit confusing.

well .. not really. If not mixing at the limit, all summing devices should sound the same. The difference comes with driving the levels near the edge of the summing devices headroom.

[areptiledysfunction]

Hey DJ, yes I have mixed on a Paris system, since 2001. You might remember me as Jesse Skeens from the NG.

I purchased a small Paris system a year ago after having sold all my stuff previously. This time I did many tests to try and figure out what was special about it. And contrary to what many on the NG said as well as what Brian shows in the DVD I did not find anything special about how Paris saturates. The coding further supports this as it shows all Paris is doing is flat topping signals.

I was basically laughed at for being crazy to say that you can acheive the same saturation that Paris produces just by raising the master fader in Wavelab and clipping a signal. Yet a few people finally tried it and agreed. I think the whole Paris sounds better thing at this point is mental. As when I got my system last year and cranked some trims and faders I too thought "hey this sounds analog" But when I tried the same thing in other DAWs I realized they sounded just the same.

Now I won't say Paris doesnt have a sound. But for me its more due to the effects and converters as well as the ability to be able to interfae hardware devices easily. And to a point the gain structure does lend itself to some creative clipping. But this is nothing that can't be acheived with another DAW set up similarly.

Let me add that I really wanted Paris to show me something unqiue in its saturation/summing so if anything my tests were done as exhasutively as I could to find something "special".

Paris is/was a great system but in the meantime I have matched or bettered all my mixes in it with current DAWs such as SX3 and Live 6.

I'm interested to see some examples of SCOPE summing. I've always wondered if it might make a difference yet at the same time I have not heard a concinving comparison that shows summing really amounts to anything.

Jesse

Hi Jesse,

Sure I remember you. I'm glad to hear you are getting good results with your summing. I think there may be some more to the Paris summing bus than simple level attenuation and it *may* be modelling in the DSP. There has been a bit of speculation that the summing bus of his custom built Helios console may have been modelled in the same way that other digital devices *model*, like as the POD, for instance. I don't know this for sure. I think the way summing is accomplished in Paris is easer and more transparent than doing the same thing in a native system, but also don't doubt that the same kind of gain staging can be accomplished manually. The summing in Paris is 52 bit "fixed" rather than floating. This same kind of thing will surely be possible in native systems once newer audio software and hardware drivers are coded for 64bit operation. I'm not a code jockey, but I know a few and I have been told that those extra bits may come in handy for a lot of things. That's why I'm not selling my RME cards until I actually *use* 64 bit CW drivers on a 64 bit OS and know that they work. One thing is for sure, RME will be all over any new developments. I am hoping CW is thinking along the same lines.

Best regards and Happy New Year. good to hear from you again.

[areptiledysfunction]

Math theory seems simple, but Wolf and Ch. Steinberg have contradicting viewpoints.

We have no different viewpoints beside the nonsense, that pan law is the cause of difference between different mixing engines (i.e. - 3dB panlaw just removes the need to readjust the volume while panning left/right). Afair Charlies answer was mainly to "defend" the cubase/nuendo mix engine against Logic (which is nonsense anyway because both are 32bit float, but hey .. it's all about marketing), so he doesn't really talk about float vs. integer.

DJ says Paris does magic (in a certain context), Medway says he could set the same performance parameters on (almost) arbitrary systems once he got his mind free from any prejudice...
it's a bit confusing.

well .. not really. If not mixing at the limit, all summing devices should sound the same. The difference comes with driving the levels near the edge of the summing devices headroom.

Agreed 100%. This stuff I'm talking about is only evident when mixing "hot"

Regards,

[medway]

Ok so after speaking with Neil I see there are different limiters being applied to both the mixes so perhaps its not completely on even ground. Neil did say they limiting was very subtle but still I'd like to hear something removing that variable.

In fact I'll just do a test of a current track I'm working on and see if there is a difference.

Jesse

[Neil]

Ok so after speaking with Neil I see there are different limiters being applied to both the mixes so perhaps its not completely on even ground. Neil did say they limiting was very subtle but still I'd like to hear something removing that variable.

In fact I'll just do a test of a current track I'm working on and see if there is a difference.

Jesse

You mean after speaking with Neil, he explained that the limiters were barely kissing the peaks on a couple of the submixes so there wouldn't be any difgital overs going through lightpipe, and that if they were doing more than that, they would serve to color the sound in a negative manner, rather than open it up & make it more detailed, and he further explained that if he could have plugged the same limiter across the Pulsar 2-buss as the one in SX, he would have, but unfortunately Pulsar can't accept Izotope Ozone VST plugin across its two-buss, so he set the Optimaster limiter for the same settings.

Let's at least get it clear, OK?

It's the closest thing to an a/b comparison I can achieve considering the two separate platforms, and I have no favored position on either one. If i had a wimpy li'l mix that didn't feature optimally-managed gain structures in SX (this means "not too hot, not too low, but rather "just right""), I would gladly try that & have no need to kick down a handful of overs on the submixes OR two-buss - however, I have no such mix of that sort.

Like I said on the other forum, do your own & see for yourself.

Neil

[medway]

Ok I just did a test on my own. Took a track and divided it amongst 4 stereo busses. Drums in 1/2, Bass in 3/4, Vocals in 5/6, and synths in 7/8.

I lowered these 4 busses by an equal amount so they did not clip. I then sent them out to SCOPE and summed there and recorded the mix via STS5000.

I then bounced 4 pairs of audio files within Cubase from the same busses at the same level they were at for SCOPE. I imported these into a new project and summed them in Cubase.


I then took the two files and compared them in Cubase. I looped small sections over and over on 2 separate tracks. Didn't hear anything different. Even turned down the moniter so I couldnt see which was which, let it loop and tried to spot one take from the other, no luck.

I did try inverting the phase of one in relation to the other and they didn't cancel, there was high end info present. But due to capturing the mix using the STS the bounces werent exactly aligned. And I didn't hear a change anyhow so this delta doesn't seem important anyways.

I have to say I would love SCOPE to provide better summing and clearer mixes just by running busses to it yet I don't see that happening here. I truly don't belive summing is an issue in mixing today, its always down to other factors.

Jesse

[hubird]

thanks for nice testing

[medway]

I will say that strapping 4 Sim-Jims across the busses and using the saturation slider yielded a nicer bigger mix. There might be some native plugs that can compete though. Shame I can't use my Oxford Dynamics with warmth control in Live 6 or SX3 without the hassle of rewire.

[garyb]

my mixes can tell the difference. possibly it's the better toolset i have in scope, but i notice a definite difference. others might not. maybe it's a thing like guitar string brands, a lot of companies, only a few manufacturers and one or two makers of wire. it's hard to prove that there are many differences in strings, so ultimately, you have to use what you feel good and comfortable with...

[medway]

my mixes can tell the difference. possibly it's the better toolset i have in scope, but i notice a definite difference. others might not. maybe it's a thing like guitar string brands, a lot of companies, only a few manufacturers and one or two makers of wire. it's hard to prove that there are many differences in strings, so ultimately, you have to use what you feel good and comfortable with...

Theres a big difference between saying you get better mixes out of SCOPE versus SCOPE having superior summing. I don't doubt people get a better mix out of a particualr DAW or set of tools.

But I have not even heard hardware consoles or summing boxes make a difference in summing (they sound different perhaps but in some cases the DAW actually sounded more open and punchy). So I really doubt SCOPE is going to provide better summing when an SSL or multi thousand dollar summing unit doesn't.

Jesse

[wolf]

That said it is really up to the guy driving the unit.
In my case I usually use 32bit float for dance & club tracks, while I prefer to use 32bit integer for rock & pop songs and for shure for classic stuff.
Integer is a bit heavier to work on though, but it's worth it.

Nice we scoper's have an option, isn't it ?

[LHong]

So far I'm very happy with the SCOPE Mixing engine! Not best but the result (mixing down) seems be pretty quality!

For someone might be interested:
http://www.purgatorycreek.com/documents/44_193.html

LongStudio