2CAudio B2: Full Body. Maximum Attitude.

[Andrew Souter]

BTW, someone on GS asked how to create your own hall vs chamber vs plate presets with B2...


Here is what I said:

In the new/imminent preset expansions I have made several folders that are distinctly classified as Hall vs Chamber vs Plate etc presets. Once our final official update is out and these are available, I will write some explanations on how to achieve these various "classes" of presets using the B2 framework...

I would say though, IMHO, often these classes are somewhat arbitrary in algorithmic reverb, and in real acoustic space they are all just semantic boxes that we humans put things into for general classification purposes to make sense of the complexity found in the real world. In other words, a particular acoustic space does not know if we human beings call it a Hall or Chamber.

Furthermore, one developer's or even one product's take on these classes can vary widely from another...

In fact, we should put it to the test: Who here can explain the difference between a "chamber alg" vs a "hall alg"? (other reverb developers are banned from answering.) Quoting someone else's marketing lit is also cheating.

For extra credit, describe why you would want to use a chamber vs a hall etc?

I feel slightly different about Plates vs 3D spaces, but even in these cases the difference between some product's Chamber's vs their plates is quite small...

In general I think these differentiations came from the marketing department in the early "classic device" manufacturers... boing

IMHO it would be more wise to talk about various perceptual and psychoacoustic attributes of the classes of things, and what the goals of each are, and why one choice might be better than another, and under what circumstances. I think that would be a very helpful and enlightening dialog to have. We should have it here in the near future... kfhkh

First, I am sincerely interested in hearing what people think about the above questions...

what do you guys think?

[Breeze]

IMHO it would be more wise to talk about various perceptual and psychoacoustic attributes of the classes of things, and what the goals of each are, and why one choice might be better than another, and under what circumstances. I think that would be a very helpful and enlightening dialog to have. We should have it here in the near future... kfhkh

You're assuming that it's necessarily a logical process, that it can be quantified and analyzed; while that may be true in a situation where you're applying a realistic model of a reverb to accurately simulate an acoustic space, most of our application of reverberation is quite the opposite: often the goal is simply making things sound better, more highlighted, more pleasant, more cohesive. I think there are too many parameters involved with that (not the least of which is the inscrutable instinct of the artist or producer) to have a meaningful discussion on the subject.

If you're talking about the formulas (algos, settings, eq) that are in common usage, and why these formulas work from the perspective of their psycho-acoustic effect on the ear including an analysis of the way a reverb reacts to a sound based on the characteristics of the sound and the reverb algo... that could be very interesting. But I could also see that as thesis material a well.

Let's not forget that a lot of effects that become common in our productions are there because of the technology itsef. Some are contrived but some are just happy accidents; both require the technology to exist. There's no logic in some things: if it sounds good then why not do it again?

So while I'm here (if relevant at all!):

In general, a Chamber is typically smaller, so bright and fast ER content, short pre-delay and short bright reverberation; a Hall is typically longer and duller ER's, longer pre-delay on the reverberation, duller but much longer fading reverberation. For extra credit: I'd use a Chamber for making a dry drumkit sound "live" and a big Hall to flatter a relatively slow moving resonant vocal line. But other factors like surfaces, materials and shape will affect the character of the reverberation as well. (This last was for extra-extra credit.)

[Andrew Souter]

You're assuming that it's necessarily a logical process, that it can be quantified and analyzed; while that may be true in a situation where you're applying a realistic model of a reverb to accurately simulate an acoustic space, most of our application of reverberation is quite the opposite: often the goal is simply making things sound better, more highlighted, more pleasant, more cohesive.

Yup. Preaching to the choir. This is why I generally avoided these classifications in the initial B2 presets in favor of grouping them rather by the structural families of the various themes of how B2 could be set up.

This is my general feeling. But I am curious what others think...

I think there are too many parameters involved with that (not the least of which is the inscrutable instinct of the artist or producer) to have a meaningful discussion on the subject.

If you're talking about the formulas (algos, settings, eq) that are in common usage, and why these formulas work from the perspective of their psycho-acoustic effect on the ear including an analysis of the way a reverb reacts to a sound based on the characteristics of the sound and the reverb algo... that could be very interesting. But I could also see that as thesis material a well.

Let's not forget that a lot of effects that become common in our productions are there because of the technology itsef. Some are contrived but some are just happy accidents; both require the technology to exist. There's no logic in some things: if it sounds good then why not do it again?

Sure. What I am interested in, is trying to get a general feeling for what the level of understanding is among musicians/producers/artists when they choose a reverb alg, plug, preset. Why do they reach for Hall vs Chamber, for example, and what do they expect the differences will be.

I know what these differences are "supposed" to be of course in the context of classic devices. But I also know that the classification is quite nebulous and blurry. Is a Room with a huge size still a room? Is a small, dense Hall, still a Hall? Etc etc. Is a Plate with a size of 30 meters still a plate? Is a hall really supposed to be sparse in its initial attack? That may be true of Lx algs, but it is certainly not true of M7 algs for example, and it is not really true of impulse reposes of real halls either, yet it DOES work well on some sources... etc. etc. etc.

Hence I am curious what people's experiences and expectations are... I hypothesize the general level of understanding of these topics is fairly low and this could be an interesting conversation...

So while I'm here (if relevant at all!):

In general, a Chamber is typically smaller, so bright and fast ER content, short pre-delay and short bright reverberation; a Hall is typically longer and duller ER's, longer pre-delay on the reverberation, duller but much longer fading reverberation. For extra credit: I'd use a Chamber for making a dry drumkit sound "live" and a big Hall to flatter a relatively slow moving resonant vocal line. But other factors like surfaces, materials and shape will affect the character of the reverberation as well. (This last was for extra-extra credit.)

cool.

Anyone else? Jump in. Let's hear what you think...

[Echoes in the Attic]

As far as why we'd use different algorithms, to be honest, I don't know what the difference is supposed to be between a big room and a chamber or a big chamber and a hall. Or any of that. Generally I just know that a room algorithm is what to choose if I want a smaller reverb with a more confined sense of space and I'll pick a hall for something bigger and more distant. One thing that is difficult to get is the real sounds of instruments being recorded in a nice hall, such as an orchestra in the AIR studios (like Spitfire Audio's recordings), or the Sony stage (like Cinesamples recordings). It's especially noticeable with percussion samples. I haven't heard a reverb give the BOOM sound to percussion as you get when they are recorded in a nice hall or room like those.

[Andrew Souter]

Thanks for your thoughts.

As far as why we'd use different algorithms, to be honest, I don't know what the difference is supposed to be between a big room and a chamber or a big chamber and a hall. Or any of that. Generally I just know that a room algorithm is what to choose if I want a smaller reverb with a more confined sense of space and I'll pick a hall for something bigger and more distant.

My guess is that is most likely the way it is for most users, more or less...

Anyone else care to share their thoughts?

One thing that is difficult to get is the real sounds of instruments being recorded in a nice hall, such as an orchestra in the AIR studios (like Spitfire Audio's recordings), or the Sony stage (like Cinesamples recordings). It's especially noticeable with percussion samples. I haven't heard a reverb give the BOOM sound to percussion as you get when they are recorded in a nice hall or room like those.

Interesting. I wonder how you will like the new presets.

[Andrew Souter]

Ok, we seem to have fixed crashing issues on AVX-capable CPUs. (It seems we used an AVX2 instruction in the AVX code, so we were trying to use an instruction that does not exist on some CPUs--hence the crash.)

If you had crashes, please re-download and re-run the installer. Please confirm all is well, if it is...

http://www.2caudio.com/support/updates#_b2

[elxsound]

Just got it installed and so far, all seems great.

I'm on OSX 10.8.4 (AVX - no AVX2) running Live 9.0.5

[Echoes in the Attic]

Just out of curiosity, how do we know if we are AVX etc. (I'm not installing the beta as I don' thave it but just curious)

[Andrew Souter]

Just out of curiosity, how do we know if we are AVX etc. (I'm not installing the beta as I don' thave it but just curious)

Sandy Bridge (2011) and later Intel CPUs have AVX.

See here:

http://en.wikipedia.org/wiki/Advanced_Vector_Extensions


Haswell (2013) and later Intel CPUs have AVX2/FMA.


Our plugs also now show what you are using as a suffix appended to the version number on the info page...

[aMUSEd]

Ok, we seem to have fixed crashing issues on AVX-capable CPUs. (It seems we used an AVX2 instruction in the AVX code, so we were trying to use an instruction that does not exist on some CPUs--hence the crash.)

If you had crashes, please re-download and re-run the installer. Please confirm all is well, if it is...

http://www.2caudio.com/support/updates#_b2

What's the code again?

Please can you give us a proper user area instead of having to dig through threads to find the update instructions every time?

[Andrew Souter]

2CAudio_B2_Update


It's listed on the page linked above...

[Breeze]

OK! Working fine here, W7-64 and it says "[B229] AVX" but I don't see much of an improvement on my system; in fact it looks like it might be eating a little more CPU than SSE. One instance of B2 using the "4C Cross Pollenization" presets and 2X oversampling used to max out under 50% and now with AVX I'm seeing 50% or more. Is that possible? Could this be Host-dependent?

[Andrew Souter]

OK! Working fine here, W7-64 and it says "[B229] AVX" but I don't see much of an improvement on my system; in fact it looks like it might be eating a little more CPU than SSE. One instance of B2 using the "4C Cross Pollenization" presets and 2X oversampling used to max out under 50% and now with AVX I'm seeing 50% or more. Is that possible? Could this be Host-dependent?

"Should" not be possible. But, we never know until we try...

Testing a single instance and looking at the host CPU meter is not very precise in measurement. It would be better to make a new project with just one audio track and one B2 on it. Then choose a preset. Then duplicate this audio track as many times as you can before you get dropouts. Note the max tracks/instances you can get. Then try the other version and see what the result is.

One limitation is that when using XTRM mode, and or OS, we can saturate the cache, and then we have to exchange data with the much slower main memory. Thus in the most intensive presets cache size and memory speed are the biggest limitations on performance.

The "4C" presets are typically using both engines set to XTRM plus OS. This is about as good/bad (depending on your perspective) as it gets.

Might be worthwhile to do the AVX vs SSE test comparison using the default preset which can be found by changing the "load last" pref to "No" and simply reloading the plug.

Also note these AVX/AVX2 optimizations are only partial/preliminary thus far. 100% of the DSP code is not updated yet. More optimization is to come in future updates for AVX/AVX2... AVX is fairly new, and AVX2 is bleeding edge, and I don't think many people are using it quite yet. AFAIK, at least...

[Sampleconstruct]

I got my celtic harp today which I bought for an upcoming sample project and this is the first quick and dirty test, played by me = dilettante, I don't usually play the harp.
She's still a bit hard to tune as she just got out the UPS lorry and was probably freezing and maybe she feels homesick too. B2 reverb was added in the mix to pimp my diletantism....

http://soundcloud.com/sampleconstruct/c ... st-with-b2

[xtrax]

I got my celtic harp today which I bought for an upcoming sample project and this is the first quick and dirty test, played by me = dilettante, I don't usually play the harp.
She's still a bit hard to tune as she just got out the UPS lorry and was probably freezing and maybe she feels homesick too. B2 reverb was added in the mix to pimp my diletantism....

http://soundcloud.com/sampleconstruct/c ... st-with-b2

Must admit you are not quite dilettante on the harp, but reverb is stunning now, after improvements new 1.2.0.
Woow ... This B2 thing is amazing. Kudos to Andrew and Dennis indeed.