The No.1 Website for Pro Audio
Digital reverbs - why outboard?
Old 3 weeks ago | Show parent
  #91
Lives for gear
 
Tom H's Avatar
 
🎧 15 years
Quote:
Originally Posted by gearstudent ➡️
another reason could be tracking. perhaps the hardware unit has less latency/less hassle then setting up an unprinted reverb in the daw. so the engineer puts some outboard non-printing reverb into the performer's monitor mix.
Latency on reverb is usually not really a problem I think, just some extra pre-delay. But when you track on an analog console with enough sends a hardware unit could indeed be convenient when tracking. In most hybrid setups its much easier to pull up a reverb from the daw. Whatever works.
Old 3 weeks ago | Show parent
  #92
Lives for gear
 
nednerd's Avatar
 
🎧 5 years
Quote:
Originally Posted by Sharp11 ➡️
Earlier in this thread, there’s a copy of posts made by the designer of the revered lexicon hardware and software verbs, he states unequivocally the port from the hardware to the software is 1:1.
This argument is moot.
It is my understanding that the PCM 96 you are referring to was designed specifically so that it could be ported to software.
Therefore it does indeed sound pretty much like the Lexicon Native Reverb Plugin.

Both the PCM96 and the Lexicon Native Plugin were designed form a separate algorithm heritage namely the PCM90/91 long after Dr. David Griesinger, the father of the famous Lexicon reverb algorithms left the company.

Hence the PCM96 could in itself be considered a modern take on and/or inspired by the classic Dr. Griesinger line of Lexicon reverbs like the 224,224X,224XL,480L,960L and their budget derivatives.

The Lexicon 960L was the last Lexicon reverb Dr. Griesinger worked on.

To my knowledge none of the classic Lexicons have been ported to software.
It is not possible to move the algorithms over from their dedicated hardware processing platforms to common CPUs without rewriting them (I am not a CPU expert, so I write this with caution).
Also it has been written that most of the classic source code got "lost" when Lexicon was acquired by Harman.

Corrections are welcome in case I got the Lexicon history wrong here.
Old 3 weeks ago
  #93
Gear Head
 
🎧 5 years
I've used some of the best reverb plugs and still the verb sound of a relatively inexpensive DSP verb FX within a sampler like the Roland SP-404SX sounds way better. I don't understand it, but I can see that comparing ITB verb at 24/44.1 will never get close to HW. You will need to at least render your project verb at 24/96 or better. Even though the HW verb is 16/44.1.
Old 3 weeks ago | Show parent
  #94
Gear Nut
Quote:
Originally Posted by Sharp11 ➡️
I think the plugin gets used more, a lot more. Why? Because it’s ubiquitous and people who need to get work done use it and don’t moan on gs about.

As for reverbs, what this thread is about in case you misunderstood - aside from plates, most software reverb is exactly the same as the hardware equivalent - they’re digital reverbs - the only difference being the hardware unit is dependent upon the computer built into its own box, the software unit uses your computers host dsp.

Earlier in this thread, there’s a copy of posts made by the designer of the revered lexicon hardware and software verbs, he states unequivocally the port from the hardware to the software is 1:1.
As someone who's designed quite a bit of "audio adjacent" hardware utilizing AD DSP devices (ADAU, Blackfin, Sharc), there simply ISN'T a 1:1 software conversion from the DSP architecture to an x86 or ARM based architecture. Just because they're utilizing an object-oriented design approach that abstracts those differences away behind high level languages doesn't make them 1:1 in anything other than pre-compiled/high-level-software, which doesn't exist at run-time.

You have to understand that it's in Lexicon's best interest to suggest they're identical. They're trying to sell a new product after all.
Old 3 weeks ago | Show parent
  #95
Gear Maniac
 
franktree's Avatar
 
Quote:
Originally Posted by LLSentelle ➡️
Plugins emulate famous analog gear for a very good reason, the analog gear sounds amazing, if, this is not so, why emulate? If digital is better than analog, stop emulating analog gear, and create something which sounds better. Seems to me this, would end the argument.
But we're not really talking about analog here, right? We're talking about digital hardware, versus digital software. I, at least, am happy to accept the idea that true analog hardware can (at least sometimes) sound "better" than digital emulations. But here, we're comparing digital to digital (except that one version--the hardware version--involves some amount of physical (analog) hardware). But the core is really a digital vs. digital comparison.
Old 3 weeks ago | Show parent
  #96
Lives for gear
 
nednerd's Avatar
 
🎧 5 years
Quote:
Originally Posted by franktree ➡️
But we're not really talking about analog here, right? We're talking about digital hardware, versus digital software. I, at least, am happy to accept the idea that true analog hardware can (at least sometimes) sound "better" than digital emulations. But here, we're comparing digital to digital (except that one version--the hardware version--involves some amount of physical (analog) hardware). But the core is really a digital vs. digital comparison.
You are grossly oversimplifying digital processing if you regard it as merely calculator in a blackbox with an input and an output.
Old 3 weeks ago | Show parent
  #97
Lives for gear
 
🎧 10 years
the main reasons are:

1.: trusted workflow. If you used a device for years you know its strengths and weaknesses. Humans are creatures of habit. So why change? (example would be guys like Andy Wallace)

2.: It's a box. Maybe with a shiny remote. Humans are haptic creatures. We like 'stuff'. The more height it occupies in a rack the better.

3.: 'Pro'-ness. People who were drooling over the now classic magazine cover studio archetype (SSL 4/9k desk, 480/960 + TC6k remote on top, lavalamps, NS10) still associate these boxes as the absolute pinnacle of never-being-able-to-afford-this-ness.

4.: Many people on GS are people who can afford to play around endlessly, print back signals, listen for tiny (apparent) differences in sound and never need to get stuff done. That's what a hobby is like.

5.: Many people on GS don't use these devices as tools to get stuff done. They use music to be able to play around with gear. Different hobby. Similar to the difference between normal people who use stereos to listen to music, and audiophiles, who use music to listen to their gear.


For me personally, I still have a fully loaded TC6k in my home studio. Don't have it connected for at least a year now.
I really love the box and will probably install it at some point again.
I don't expect it to sound any different to the just recently published plugins with the same algorithms though.

In my dayjob (audio post) we got rid of the last signal processors that are not plugins 7 years ago in all studios (surround reverbs), and will never go back. Speed and total recall are absolutely mandatory, and the sound of let's say Stratus3D is as good - if not better - as any hardware reverb. Of which none can handle immersive audio.

In the old days I loved the 224XL. Do I miss it? Not at all.

Last edited by kosmokrator; 3 weeks ago at 07:27 PM..
Old 3 weeks ago | Show parent
  #98
Lives for gear
 
syra's Avatar
 
1 Review written
🎧 15 years
Quote:
Originally Posted by Sharp11 ➡️
Earlier in this thread, there’s a copy of posts made by the designer of the revered lexicon hardware and software verbs, he states unequivocally the port from the hardware to the software is 1:1.
Correct me if I’m wrong but doesn’t UA claim 1:1 port of 480L and AMS RMX? Neither of those sound as good as my hardware. Especially the AMS.
Old 3 weeks ago | Show parent
  #99
Gear Maniac
 
franktree's Avatar
 
Quote:
Originally Posted by kosmokrator ➡️
Similar to the difference between normal people who use stereos to listen to music, and audiophiles, who use music to listen to their gear.
I love this formulation.
Old 3 weeks ago | Show parent
  #100
Gear Guru
 
thethrillfactor's Avatar
 
4 Reviews written
🎧 15 years
Quote:
Originally Posted by syra ➡️
Correct me if I’m wrong but doesn’t UA claim 1:1 port of 480L and AMS RMX? Neither of those sound as good as my hardware. Especially the AMS.
The 480L wasn't a 1:1 port as the original algorithms used in the 480L are now gone( or that's what I remember reading somewhere).

The original RMX16 did not run at a 44.1khz sampling rate I believe.I think it was made to run at 40khz. So the software has to be recoded so the delays and filters are scaled back to sample rate convert back down to 16 bit/40khz. This will affect the sound somewhat. It's also why even hardware reverb reissues like the AMS RMX16 500/Eventide 2016 / Dynatron 255 don't sound the same either.

None of the digital reverbs that have been ported over to plug ins( 224, 480L, Eventide SP 2016, EMT 250) sound exactly like the boxes. The boxes in all these cases sound better, but you have to live with the "unobtanium" of parts and possible support.
Old 3 weeks ago
  #101
Lives for gear
 
toledo3's Avatar
 
🎧 15 years
I think it all boils down to the judgement of the teams that create the hardware, or the plugins, and whether their idea what sounds good jives with yours.

The other thought that comes to mind, is that there may be something about the plugin format that causes some developers to focus on whatever they perceive to be algorithmic purity, and elegance in the code. That often leads to needing a fair amount of pre and post processing with some popular plugin reverbs to get them to sit right in a mix...though this does depend on sonic goals for sure.

If you can get the sound you want quicker and with less deliberation with hardware, why not just do that? If you have a little twinge in your gut that you need to “do something” to the plugin reverb because it doesn’t sit right...it can turn into chasing something and eating away at time and effort.

At the same time, I suppose the opposite could be true in other scenarios, if there is some plugin that you just LOVE the sound of. No point in questioning it.
Old 3 weeks ago | Show parent
  #102
Lives for gear
 
🎧 5 years
Quote:
Originally Posted by Sharp11 ➡️
Earlier in this thread, there’s a copy of posts made by the designer of the revered lexicon hardware and software verbs, he states unequivocally the port from the hardware to the software is 1:1.
Never let facts spoil a good argument.
Old 3 weeks ago | Show parent
  #103
Lives for gear
 
Sharp11's Avatar
 
🎧 15 years
Quote:
Originally Posted by Seamoss ➡️

You have to understand that it's in Lexicon's best interest to suggest they're identical. They're trying to sell a new product after all.
Conspiracy theories are often the last resort in a losing argument.

They wouldn’t have to lie if the difference were night and day though, would they?
Old 3 weeks ago | Show parent
  #104
Lives for gear
 
Sharp11's Avatar
 
🎧 15 years
Quote:
Originally Posted by thethrillfactor ➡️

None of the digital reverbs that have been ported over to plug ins( 224, 480L, Eventide SP 2016, EMT 250) sound exactly like the boxes. The boxes in all these cases sound better, but you have to live with the "unobtanium" of parts and possible support.
“Sounds better” to whom, you?
Old 3 weeks ago | Show parent
  #105
Lives for gear
 
Sharp11's Avatar
 
🎧 15 years
Quote:
Originally Posted by syra ➡️
Correct me if I’m wrong but doesn’t UA claim 1:1 port of 480L and AMS RMX? Neither of those sound as good as my hardware. Especially the AMS.
I can’t tell you what you think “sounds better” - that’s a subjective assessment.
Old 3 weeks ago | Show parent
  #106
Gear Nut
Quote:
Originally Posted by Sharp11 ➡️
Conspiracy theories are often the last resort in a losing argument.

They wouldn’t have to lie if the difference were night and day though, would they?
Please.. in all your inherent wisdom and winning arguments, define where and how they're 1:1. What does this mean to you?

They're different things so they're clearly not identical.

What level of technological differences are allowed to be considered 1:1 in your almighty opinion?
Old 3 weeks ago | Show parent
  #107
Lives for gear
 
Sharp11's Avatar
 
🎧 15 years
Quote:
Originally Posted by Seamoss ➡️
Please.. in all your inherent wisdom and winning arguments, define where and how they're 1:1. What does this mean to you?

They're different things so they're clearly not identical.

What level of technological differences are allowed to be considered 1:1 in your almighty opinion?
I didn’t bring it up - go back to post #62 , the thread isn’t that long, and if you need further elucidation, search GS for Michael Carnes input on the code he wrote for the pcm 96.
Old 3 weeks ago | Show parent
  #108
Gear Nut
Quote:
Originally Posted by Sharp11 ➡️
I didn’t bring it up - go back to post #62 , the thread isn’t that long, and if you need further elucidation, search GS for Michael Carnes input on the code he wrote for the pcm 96.

"Code" is an algorithm/recipe. I can give 15 people the exact same recipe for boiled water and I wouldn't end up with the same product on the other end. Thus is the same for compilers/assemblers/linkers and "ideal" analog circuits, not to mention different processor architectures. Run-time differences is a sophomore level course in any accredited EE/embedded-systems program. At least in the US where I have experience teaching such courses. We didn't even get to the differences between FPGA/CPLD/DSP devices and a more standard "computer" architecture.

This is simple ****. I understand your confusion, but you're missing the forest for the trees.
Old 3 weeks ago | Show parent
  #109
Gear Guru
 
thethrillfactor's Avatar
 
4 Reviews written
🎧 15 years
Quote:
Originally Posted by Sharp11 ➡️
“Sounds better” to whom, you?
Yes! As in my studio, my opinion is the only that really matters.
Old 3 weeks ago | Show parent
  #110
Gear Maniac
 
Quote:
Originally Posted by Seamoss ➡️
"Code" is an algorithm/recipe. I can give 15 people the exact same recipe for boiled water and I wouldn't end up with the same product on the other end. Thus is the same for compilers/assemblers/linkers and "ideal" analog circuits, not to mention different processor architectures. Run-time differences is a sophomore level course in any accredited EE/embedded-systems program. At least in the US where I have experience teaching such courses. We didn't even get to the differences between FPGA/CPLD/DSP devices and a more standard "computer" architecture.

This is simple ****. I understand your confusion, but you're missing the forest for the trees.
Maybe you're the guy who can answer my question from a ways back in the thread: What is the difference between how the Intel chip in my laptop handles a reverb algorithm that is "inferior" to the way a SHARC or FPGA chip handles it in a dedicated box? And is the difference audible, given the same algorithm?
Old 3 weeks ago | Show parent
  #111
Gear Nut
Quote:
Originally Posted by Rob 28 ➡️
Maybe you're the guy who can answer my question from a ways back in the thread: What is the difference between how the Intel chip in my laptop handles a reverb algorithm that is "inferior" to the way a SHARC or FPGA chip handles it in a dedicated box? And is the difference audible, given the same algorithm?
I think your question is the real crux of this entire debate. There is absolutely nothing "inferior" about how an Intel/AMD chip would handle a reverb algorithm. In fact, there's no reason to suggest that the CPU based approach, maybe aided by GPUs, can't provide an implementation that is far "superior" in a large list of metrics. My argument is that they're not "identical" due to the vast amount of variables associated.

Which leads to the ultimate issue that this is all incredibly subjective and test equipment/blind-tests can't settle the debate in-my-opinion.
Old 3 weeks ago
  #112
Lives for gear
 
toledo3's Avatar
 
🎧 15 years
It is possible to port an algorithm without getting into emulating the real world levels of precision that various parameters had in the hardware units as a byproduct of the actual parts implementation. It can become a matter of philosophy whether or not to code in various bandwidth or dynamic range limits, etc, that aren’t really part of the effects algorithms themselves in some cases.

Whether or not that has a cumulative difference is going to be on a case by case basis.

Then there is the effect of any analog parts, their impedance and capacitance values, the way they interact with the rest of the signal chain...for some units this is not really an issue because they really are all digital except conversion, or you can use digital i/o.

A rough parallel analogy is video game emulation. You can emulate an early video game without necessarily emulating the CRT screen it originally rendered to, but the result may actually be different than it was originally perceived to be.
Old 3 weeks ago | Show parent
  #113
Gear Nut
Quote:
Originally Posted by toledo3 ➡️
It is possible to port an algorithm without getting into emulating the real world levels of precision that various parameters had in the hardware units as a byproduct of the actual parts implementation.
As a synth nerd too, I've never loved any of the "virtual analog" stuff, but I highly respect the design philosophy. Most of those dude(ette)s are the last people to try to claim their stuff is "identical" or "1:1."
Old 3 weeks ago
  #114
Lives for gear
 
toledo3's Avatar
 
🎧 15 years
Yep, totally.
Old 3 weeks ago | Show parent
  #115
Gear Maniac
 
Quote:
Originally Posted by Seamoss ➡️
I think your question is the real crux of this entire debate. There is absolutely nothing "inferior" about how an Intel/AMD chip would handle a reverb algorithm. In fact, there's no reason to suggest that the CPU based approach, maybe aided by GPUs, can't provide an implementation that is far "superior" in a large list of metrics. My argument is that they're not "identical" due to the vast amount of variables associated.

Which leads to the ultimate issue that this is all incredibly subjective and test equipment/blind-tests can't settle the debate in-my-opinion.
Thank.
You.

My, it took a long time to get a reasonable explanation of this. I have no quibble whatsoever with those who like hardware boxes over software. I just get so tired of the highly subjective hubris that "even my cheapo rack unit sounds night and day better than any plugin". It leads me to think that many of this opinion simply have not tried the latest (sometimes expensive!) plugins, or that they haven't done due diligence to route them in various ways, with EQ, compression, etc.

In the end, I don't care that my LX 480 plug is slightly off the 480L - it still gives me that wonderful lush tail and complex modulation. For about $12,000 less.
Old 3 weeks ago | Show parent
  #116
Gear Nut
Quote:
Originally Posted by Rob 28 ➡️
Thank.
You.

My, it took a long time to get a reasonable explanation of this. I have no quibble whatsoever with those who like hardware boxes over software. I just get so tired of the highly subjective hubris that "even my cheapo rack unit sounds night and day better than any plugin". It leads me to think that many of this opinion simply have not tried the latest (sometimes expensive!) plugins, or that they haven't done due diligence to route them in various ways, with EQ, compression, etc.

In the end, I don't care that my LX 480 plug is slightly off the 480L - it still gives me that wonderful lush tail and complex modulation. For about $12,000 less.
Definitely. Use all the tools you have available to you, and I can't imagine there are very many circumstances where an amazing song was held back by "inferior" equipment when there was a good engineer familiar with that equipment around.

I use a PCM60/42 and ValhallaRoom on almost everything I do.

Only the sith deal in absolutes.
Old 3 weeks ago | Show parent
  #117
Lives for gear
 
Sharp11's Avatar
 
🎧 15 years
Quote:
Originally Posted by Seamoss ➡️
I think your question is the real crux of this entire debate. There is absolutely nothing "inferior" about how an Intel/AMD chip would handle a reverb algorithm. In fact, there's no reason to suggest that the CPU based approach, maybe aided by GPUs, can't provide an implementation that is far "superior" in a large list of metrics. My argument is that they're not "identical" due to the vast amount of variables associated.

Which leads to the ultimate issue that this is all incredibly subjective and test equipment/blind-tests can't settle the debate in-my-opinion.
Which is why the spitfire audio blind test is often posted in threads of this type - there are so many now I’ve stopped checking to see if it’s here.
Old 3 weeks ago | Show parent
  #118
Lives for gear
 
Sharp11's Avatar
 
🎧 15 years
Quote:
Originally Posted by Seamoss ➡️
"Code" is an algorithm/recipe. I can give 15 people the exact same recipe for boiled water and I wouldn't end up with the same product on the other end. Thus is the same for compilers/assemblers/linkers and "ideal" analog circuits, not to mention different processor architectures. Run-time differences is a sophomore level course in any accredited EE/embedded-systems program. At least in the US where I have experience teaching such courses. We didn't even get to the differences between FPGA/CPLD/DSP devices and a more standard "computer" architecture.

This is simple ****. I understand your confusion, but you're missing the forest for the trees.
I’m not missing anything because I’m not looking - this issue was settled for me long ago
Old 3 weeks ago | Show parent
  #119
Lives for gear
 
DougS's Avatar
 
🎧 5 years
Quote:
Originally Posted by Rob 28 ➡️
Maybe you're the guy who can answer my question from a ways back in the thread: What is the difference between how the Intel chip in my laptop handles a reverb algorithm that is "inferior" to the way a SHARC or FPGA chip handles it in a dedicated box? And is the difference audible, given the same algorithm?
Hi Rob - did some research on SHARC Chips - here's what I found:

SHARC chips are made primarily by Analog Devices (ticker: ADI) and are designed for high precision floating point calculations. Essentially real-time math applications where normal rounding error is too much error - like a radar system used to guide an air-to-air anti-aircraft missile, where you're trying to pinpoint an object's exact position in real time based on the time it takes radio waves to bounce back - the math must be done fast and at high precision. They are architected specifically for these kinds of applications and their major advantage is they are architected to only use one large word size allowing for fast storage and retrieval of high precision floating point numbers (many decimal places).

Analog Devices makes a line of SHARC chips specifically designed for digital audio. If your car has surround sound option that converts regular two channel sources - its probably on a SHARC chip. If your car pipes extra engine notes into the cabin to enhance the driving experience - its probably a SHARC Chip. ADI sells an entire mother board based around its SHARC chip you can use to quickly build a sound bar, smart speaker and other audio/DSP products.

As to your question - if the reverb algo was directly ported to a laptop, would the difference be audible? I think the question is - if the algo was processed to the same level of precision would the laptop be able to keep up and process it in real time? I have not direct info on that.

Of course, in a plugin you can create many instances and I have to presume that part of plugin design is A) finding mathematical shortcuts to preserve cpu cycles and B) tapering precision when and where its least audible. Constraints in a dedicated SHARC environment will be different. So who knows?

On the other hand, if my life depended on shooting down the enemy jet, I think I'd rather have the SHARC chips doing the math then my laptop. ( Funny how the context makes the answer easy.)
Old 3 weeks ago | Show parent
  #120
Lives for gear
 
syra's Avatar
 
1 Review written
🎧 15 years
Quote:
Originally Posted by thethrillfactor ➡️
The 480L wasn't a 1:1 port as the original algorithms used in the 480L are now gone( or that's what I remember reading somewhere).

The original RMX16 did not run at a 44.1khz sampling rate I believe.I think it was made to run at 40khz. So the software has to be recoded so the delays and filters are scaled back to sample rate convert back down to 16 bit/40khz. This will affect the sound somewhat. It's also why even hardware reverb reissues like the AMS RMX16 500/Eventide 2016 / Dynatron 255 don't sound the same either.

None of the digital reverbs that have been ported over to plug ins( 224, 480L, Eventide SP 2016, EMT 250) sound exactly like the boxes. The boxes in all these cases sound better, but you have to live with the "unobtanium" of parts and possible support.
Thanks Thrill that may very well explain it.

Quote:
Originally Posted by Sharp11 ➡️
I can’t tell you what you think “sounds better” - that’s a subjective assessment.
Obviously. I’m a little confused though on what your argument is. Do they sound the same to you?
📝 Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 547 views: 63316
Avatar for psycho_monkey
psycho_monkey 1st October 2015
replies: 150 views: 42486
Avatar for 10pints
10pints 3 weeks ago
replies: 609 views: 14148
Avatar for s wave
s wave 18 minutes ago
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump