Let's talk about reverb, will you?

[Rekel]

I doubt it was pointed at anyone.

 

 

Well, it was pointed at people who Beck thought only knew reverb software plug-ins and never worked with actual hardware, wasn't it?

 

 

We all need to keep in mind that the ear is easily fooled and these things are releative. There is no proper perspective if you dont compare for yourself. Software sales people count on you not comparing along with this huge mass of folks desperate to convince themselves the software is as good so they dont have to face the cost of real gear. But its just a silly form of self deception if you want the best quality. If the software is good enough for ones needs then there is no argument. But it isnt as good, thats just the truth.

 

 

With all due respect but it really doesn't matter to John and Jane end consumer whether hard- or software reverbs were used in any given production. That's just the truth.

[roomjello]

 

Well, it was pointed at people who Beck thought only knew reverb software plug-ins and never worked with actual hardware, wasn't it? With all due respect but it really doesn't matter to John and Jane end consumer whether hard- or software reverbs were used in any given production. That's just the truth.

 

I hope you don't truly believe that. If you do you should study perception a bit.

Music and sonics are percieved on pretty deep levels.

Just because a consumer cannot put their finger on a lack of depth, doesn't mean their experience isn't comprimised.

As a recordist it should be a paramount concern that you address the perception of the end user.

To say what you just did shows a lack of desire to be the best you can at your craft.

But many are for financial and convienience reasons working hard at dismissing this truth.

[Rekel]

I hope you don't truly believe that. If you do you should study perception a bit. Music and sonics are percieved on pretty deep levels. Just because a consumer cannot put their finger on a lack of depth, doesn't mean their experience isn't comprimised. As a recordist it should be a paramount concern that you address the perception of the end user. To say what you just did shows a lack of desire to be the best you can at your craft. But many are for financial and convienience reasons working hard at dismissing this truth.

 

 

I really do. Believe me, the consumer doesn't give a single hoot whether hard- or software reverbs were used in a production. Sorry, that's just the plain truth. Like I said, I've been in this industry for quite some time.

 

And no, I always strived to be the best I can do in my craft. Without going into the gory details, my musical career is testimony to that.

 

Also, a lack of finances doesn't play any role in my personal opinion on this matter.

[roomjello]

Not all music is dense and cluttered, perhaps in your world it doesn't matter but perhaps there are other worlds where it does. Music with space and detailed ambience will be strongly affected by a crappy verb. It's perceivable and if it is perceivable then it matters to the ear and therefore the listeners experience.

And honestly it sounds to me like you don't want it to matter, if it's perceivable it matters.

[Rekel]

Not all music is dense and cluttered, perhaps in your world it doesn't matter but perhaps there are other worlds where it does. Music with space and detailed ambience will be strongly affected by a crappy verb. It's perceivable and if it is perceivable then it matters to the ear and therefore the listeners experience. And honestly it sounds to me like you don't want it to matter, if it's perceivable it matters.

I'm afraid we'll just have to agree to disagree here, roomjello. I don't think a software package like the one made by Lexicon can be called "crappy" by any stretch of the imagination.

Heck, even EpicVerb, which I already mentioned in this thread, does a great job, imo. And this is freeware for crying out loud.

The quality differences between these reverbs aren't perceivable by the end consumer. It's how you use them that counts. All else is getting into cork sniffing territory, imho

[roomjello]

I'm afraid we'll just have to agree to disagree here, roomjello. I don't think a software package like the one made by Lexicon can be called "crappy" by any stretch of the imagination. Heck, even EpicVerb, which I already mentioned in this thread, does a great job, imo. And this is freeware for crying out loud. The quality differences between these reverbs aren't perceivable by the end consumer. It's how you use them that counts. All else is getting into cork sniffing territory, imho

Turn the volume way down when you are applying them and slowly add it in. With soft verbs your substance is sort of hollow and the amount of verb required goes beyond what you want to use to get it audible at low volumes. The Lexi pack suffers bad from this due to its 2 dimensional nature. Yes the sounds are nice, but the music ends up with flat texture on it.

[Philter]

 

Turn the volume way down when you are applying them and slowly add it in. With soft verbs your substance is sort of hollow and the amount of verb required goes beyond what you want to use to get it audible at low volumes. The Lexi pack suffers bad from this due to its 2 dimensional nature. Yes the sounds are nice, but the music ends up with flat texture on it.

 

 

The Lexicon native plugin uses the same algorithms as their PCM96 hardware unit. You think different processors have their own sound? So like, the processors doing the digital math in the PCM96 sounds warmer than the processor chips in a computer? Which is warmer sounding, AMD or Intel?

[roomjello]

 

The Lexicon native plugin uses the same algorithms as their PCM96. You think different processors have their own sound?

 

 

I have not compared the soft to the hard in this case. But i have compared those soft versions to other hardware and they don't cut it.

I don't know why technically, don't really care, and i would also prefer that the plugs sounded as good, obviously, but they don't and i'm not about to close my ears to it no matter what anyone else's argument is, i know what i hear, we all do, we just have to believe it, seek the truth.

[Folder]

The Lexicon native plugin uses the same algorithms as their PCM96 hardware unit. You think different processors have their own sound? So like, the processors doing the digital math in the PCM96 sounds warmer than the processor chips in a computer? Which is warmer sounding, AMD or Intel?

 

 

The processor in most hardware reverb units are optimised for digital signal processing.

Your computer's CPU is a general purpose processor.

[UstadKhanAli]

 

I really do. Believe me, the consumer doesn't give a single hoot whether hard- or software reverbs were used in a production. Sorry, that's just the plain truth.

 

 

WE'RE the ones who are supposed to give a {censored} about sound quality.

 

I will try everything I can to get my recordings, my audio, my mixes the best I can.

 

I don't care whether a consumer cares or not what I use. What I care is that the recording supports the artistic and emotional statement. And to do that, I will make my recording sound as fantastic as possible because, again, I care about the job I do, and I care about how it sounds. I'm passionate about this.

 

I don't care whether consumers know what I am using or what I am not (and really, some of them do, but that's besides the point). What I care about is not cutting corners and not settling to the best of my ability and to the best of my current equipment.

[UstadKhanAli]

 

The quality differences between these reverbs aren't perceivable by the end consumer.

 

 

Why all this talk about the consumer? WE'RE the ones who need to care about audio. If we don't, who does?

 

Can YOU tell the difference? Do YOU care? That's what matters. Why? Because you're the one recording, not the consumer.

[Folder]

I don't like reverb.

Too much reverb can make things sound muddy. Especially in the lower frequencies.

When I listen to music, I like to be able to hear each instrument separately.

When the big arena sized reverbs of the 80's became popular I remember thinking that those records sounded like you were sitting in the very last row of the very top section of the Omni.(An Atlanta Arena) I always hated sitting in the very last row of the very top section of the Omni. The sound sucked up there.

You had to bring your binoculars and all you could hear was just a big ol' wash of noise. Why would you want to make your record sound like that? That's why we camped out for concert tickets so we could be in front of the stage and hear the sound of the band "without" reverb.

Most of the music I like has very little reverb. A lot of classic rock was recorded dry and reverb was used mainly as an effect.

I bought a Digitech and a Lexicon rackmount. I thought I needed them because they were all the rage. Neither one sounded as good as the plug-ins I use now.

My favorite reverb that I own is my Alesis Wedge. It's kind of dark and thick sounding. I can get it to sit unobtrusively in my mixes. It has the best sounding delay I've ever used and it pretty much stays in my guitar amp's effects loop. I also use "Sonar's Perfect Space" and "IK Multimedia Classik Studio Reverb" but my number one go to reverb is "Overloud's Breverb". I usually filter out all low frequencies and depending on the material most midrange as well. I like just a touch on the top end to add a sense of space.

If you notice the reverb then you've used too much in my opinion.

[roomjello]

 

Why all this talk about the consumer? WE'RE the ones who need to care about audio. If we don't, who does? Can YOU tell the difference? Do YOU care? That's what matters. Why? Because you're the one recording, not the consumer.

 

Right on Ken.

[Anderton]

you realize you're comparing software with software.. the dedicated reverb box being a computer with reverb software running in it, the plugin being a computer with reverb software running in it.

 

That is a very good point. Many companies simply port their digital algorithms they used in hardware over to software. IIRC the Korg plug-in version of the Wavestation was pretty much identical to the hardware. As to native vs. DSP, older chips like the 56000 are weaklings compared to even a fraction of the power of a modern-day multicore processor.

 

When I reviewed the Waves version of the Aphex Aural Exciter, I thought it was silly to have the option to be able to switch in the noise and hum components of the original, and asked Waves about why they did this. Waves claimed that when they tested the plug-in, those who were experienced with the sound of the Aural Exciter thought it didn't sound "right" without those elements of the original hardware included.

 

But think about this for a second: Suppose the original hardware exciter did NOT have these noise and distortion components, and Waves decided to make a plug-in that added arbitrary amounts of noise and distortion. Would people say "Finally!! Thank heavens my Aural Exciter now has some crappiness to it! It was just to damn perfect before! I'm so happy we have these plug-ins that add problems not in the original!!"

 

I think IK Multimedia got it right with their Mellotron emulation. You could emulate anything from a crappy Mellotron whose heads hadn't been demagnetized since the French revolution all the way to a "perfect" Melloton. You can probably figure out which one I choose

[Anderton]

As to Beck's point about manufacturer claims about hardware software blah blah...when Lexicon introduced the PCM-80, of course people gushed about its reverb, because it said "Lexicon" on the front panel. But the PCM-80 was never, ever designed to be a reverb. It was designed to be a flexibly multi-voice chorus, but marketing was horrified. "It has to have a reverb!" So engineering added a reverb, almost as an afterthought...and if you talk to any of the engineers who were around during that time, they'll tell you the PCM-80's reverb was not one of their finest moments. But of courrse, people think it's wonderful because it's a Lexicon piece of hardware...

 

That reverb, and the one in the Alesis MIDIverb, were the two reverbs I used on my Forward Motion CD of many years ago. Musicians would listen to it and automatically assume the MIDIverb was the PCM-80 because it sounded "better." Nope. Preset #23 on the MIDIverb was hard to beat

 

As to Rekel's comment, I think what he's saying is that if the music is compelling, the reverb you put on it is relatively inconsequential to most people. While of course the pursuit of audio excellence is a worthy goal, I've seen sessions where the engineers spent so much time obsessing over the sound they forgot to obsess about the music, and the performances suffered.

 

It's all about balance. I did a cover version of Saul T. Nads' "Black Market Daydreams" in about 10 hours. I'm sure I could have chosen some better sounds for some of the parts, but then the song would still be sitting on my hard drive, not on YouTube. I think at some point, sometimes you really do have to say "it's good enough." Although of course, it's best if that's said by someone with very high standards And if the parts themselves are good!

[Philter]

 

The processor in most hardware reverb units are optimised for digital signal processing. Your computer's CPU is a general purpose processor.

 

 

I don't think that has anything to do with how the reverb algorithms sound. I can run dozens of Lexicons on my computer in spite of it having a general purpose processor, simply because that relatively old Q6600 is light years beyond the number crunching utility of anything Lexicon ever put in a hardware unit.

[WRGKMC]

One important thing about hardware units is they have preamps.

Some like the Alesis have input, and output gains. A big part of their

sound quality comes from its analog gains staging.

 

Analog Preamps have sweet spots. Without getting technical, most preamps

will have a sweet spot when pushed 70%. Along with that if an actual potentiometer

is used to adjust gain, that pot will have an EQing effect on the signal to some extent.

Turning the gain down may darken the sound slightly, and cranking it up above the sweet spot

may add some edginess to the drive.

 

My biggest obstacle switching from analog to digital was the loss of these simple gain staging effects.

Everything is digital had a linear change, and there were no notable sweet spots mixing in the box so

all my expertise at pushing a gear chain for its optimal sweet spots disappeared. It

[Rekel]

As to Rekel's comment, I think what he's saying is that if the music is compelling, the reverb you put on it is relatively inconsequential to most people. While of course the pursuit of audio excellence is a worthy goal, I've seen sessions where the engineers spent so much time obsessing over the sound they forgot to obsess about the music, and the performances suffered.

You hit the nail on the head. That is indeed exactly what I was trying to get across in my own clumsy way. (remember, English isn't my native language).

I can't count the times where I watched an engineer wasting so much valuable time and energy (and in the process, losing the momentum of a great session vibe) while trying to edit a few bum notes with a mouse in a MIDI/audio take, whereas it would have been so much easier and faster to have the performer do a retake of this 'less than perfect part' with a simple punch in/out.

I think at some point, sometimes you really do have to say "it's good enough." Although of course, it's best if that's said by someone with very high standards And if the parts themselves are good!

^^^ This. I couldn't agree more

[UstadKhanAli]

I will take your word for it, Rekel. I just wanted to get across that WE should be the ones concerned about the sound quality, not the consumer.

 

I know that many consumers don't care about the process, only the song. But that idea doesn't change how I go about recording and mixing, and that's what I want to get across.

[UstadKhanAli]

I don't like reverb. Too much reverb can make things sound muddy. Especially in the lower frequencies.

It's for this reason that I use delay(s) to give a sense of space rather than reverb quite often, although obviously it really depends on the song. But if I use delays, I can either not use reverb or seriously back off on it.

I also roll off highs and lows on my reverb quite often to stop the mud and so forth. I like to have an EQ (and sometimes a compressor) before the reverb to give me more control over what gets, uh, "reverbed".

If you notice the reverb then you've used too much in my opinion.

So often, that's true, although that depends on the genre. Some ambient stuff, Goth, or Cocteau Twins and dreampop stuff just screams for reverb. But even then, you don't typically want it clouding up the works, so it takes some careful listening to apply lots of reverb but make it sound really musical and clear. And again, often with delays instead of so much reverb.

Good post.

[Rekel]

I also roll off highs and lows on my reverb quite often to stop the mud and so forth. I like to have an EQ (and sometimes a compressor) before the reverb to give me more control over what gets, uh, "reverbed".

Indeed. Most of the reverbs already have a low cut/ high pass filter. Also, clever use of predelay before the reverb kicks in and even gated reverb (both set in time with the meter of the song) will help much with cleaning up the mud.

[chevybusa]

Well, I'm 100% ITB (see sig), and use softies and VSTis exclusively, as well as 9 times out of 10 using loops for the drums which usually are alerady processed really well, so usuaslly when I need reverb I use the built in FX on whatever softy or VSTi I'm using...if I need it and it doesn't have it, I'll use one of Logic's built in reverbs or Lexicons suite, depending on the application.

[JeffLearman]

 

I don't think that has anything to do with how the reverb algorithms sound . I can run dozens of Lexicons on my computer in spite of it having a general purpose processor, simply because that relatively old Q6600 is light years beyond the number crunching utility of anything Lexicon ever put in a hardware unit.

well, there's a huge difference between running an algorithm on a 24-bit fixed width processor (assume 48-bit intermediates) and running "the same" algorithm on a general-purpose computer, with 32-bit float and 63-bit float.

 

I say "the same" because when you code an algorithm for hardware, one of the first things you do is find where you might run into near infinites or near infinitesimals, and rework the algorithm to minimize them, with a good idea of your value range for intermediates when you do it -- and those value ranges are hugely different for fixed vs. float!

 

However, reverbs generally shouldn't run into near-infinites, or near-infinitesimals in ways that matter. (Especially convolution reverbs, where the problem simply doesn't happen for sane filter kernels -- "impulses").

 

 

My biggest obstacle switching from analog to digital was the loss of these simple gain staging effects. Everything is digital had a linear change, and there were no notable sweet spots mixing in the box so all my expertise at pushing a gear chain for its optimal sweet spots disappeared. It

[JeffLearman]

 

My "secret" for reverb is adding a whole bunch of individual, short prime number delays to give good early reflections. That's where many reverbs fall down.

I like using early reflections with minimal tails to help give different locations to different instruments, putting them in different places in a virtual room. That is, the track that's to the left will have faster early reflections on the right than the left, and vice versa. The thing I like about this is it gives a spacious, non "mono-pan" sound, but without all the mud that using lots of reverb can add up to. Then a master channel reverb to add a little more life/warmth to the space, again to avoid mud. And channel-specific reverbs where appropriate, like the spring reverb in a guitar amp.

 

SIR is way cool in concept, though I haven't actually used it much. I don't know whether they "fixed" this, but I didn't like its huge latency. The latency is necessary for a true convolution, but reverbs only need half the convolution kernal. (The first half of a convolution kernel is essentially a time machine, where the signal can affect the results ahead in time -- like "pre-echoes". Reverbs don't need pre-echoes.) Anyway, I'd like a convolution reverb like SIR that was zero latency, adding no more latency than your DAW's normal audio output buffering latency.

 

Philbo, do you use a starter pistol, or a shotgun? I always wondered how much different the results would be. Is there a way to normalize the impulse based on the input? I guess you'd have to fire your pistol in an anechoic chamber, and then figure out how to process it. Regardless, I think it's be cool to collect a lot of good impulses! I suspect the main reason I gave up on SIR was the amout of time and effort required to collect the good impulses.

[Mark L]

Here's another subject I know nothing about