Gearslutz

Gearslutz (https://www.gearslutz.com/board/)
-   The Sweetwater Virtual Mic Shootout 2018 (https://www.gearslutz.com/board/the-sweetwater-virtual-mic-shootout-2018-a/)
-   -   Sweetwater's Virtual Mic Shootout (https://www.gearslutz.com/board/the-sweetwater-virtual-mic-shootout-2018-a/1213691-sweetwaters-virtual-mic-shootout.html)

Lynn Fuston 30th April 2018 06:52 PM

Sweetwater's Virtual Mic Shootout
 
Is it possible for an affordable solid-state large-diaphragm mic to sound convincingly like a world-renowned vintage tube mic worth tens of thousands of dollars? That's what we at Sweetwater wanted to find out. So we put the Slate Virtual Microphone System, the Townsend Labs Sphere L22 Microphone Modeling System, and the Antelope Audio Edge Large-diaphragm Modeling Microphone in the studio right next to the mics they modeled. Then we had two vocalists sing into each mic, in a carefully calibrated test to see what differences we could or couldn't hear. We call it the Sweetwater Virtual Mic Shootout. The differences/similarities were absolutely enlightening to us. And we'd love to hear your opinion after you listen to these samples.

Are virtual mics the future of vocal mics? Sound off on what you think.

Check out how we set up the listening test and hear the virtual and vintage mics side by side. Listen to 320K MP3s or download the high-resolution 24/96 files here:
Virtual Mic Shootout — with Sound Samples - inSync | Sweetwater

https://www.sweetwater.com/insync/me...-1-650x340.jpg

Whitecat 30th April 2018 10:02 PM

Yes!!! Big welcome to Lynn again and a big welcome to everyone participating in this latest mic lab from our pals at Sweetwater. Here we go!! :cowbell:

latweek 30th April 2018 10:15 PM

Mhhmmmm...subscribed.
https://images.talkbass.com/attachme...p-jpg.2564530/

Wolf LeProducer 30th April 2018 10:45 PM

I think it would be very difficult in a blind test for people to tell the difference between a $25,000 microphone and any of the three $1000 - $15000 mic modelers tested here.

This antelope edge is so on my list to buy - I just think its brilliant and I have brand bias. As a brand Antelope seems to be best of the three.

Another thing to consider -- once autotune or variant is implemented, telling the difference between the $1300 edge strip and a vintage mic is just not capable for a great majority of ears - maybe even close to 90%, 95% -- 99.9%??


I would love to see the results of say about, "5," very super famous mastering engineers brought into the Sweetwater studios for a blind test. Can they pick out which mic, was on which recording without knowing in advance...

nikoli 30th April 2018 10:48 PM

First pass, only difference that stuck out to me was between the male vocal U67 and Townsend LD-67. The Townsend sounded less crisp in the highs and ess sounds. The female vocal take sounded the same though.

DatMicrophoneDo 30th April 2018 11:31 PM

******** sony gpac lol. once you learn to read frequency responses you wont need to care about shootouts.

mdme_sadie 30th April 2018 11:57 PM

Hmm, the U67 on female vocals seems to be the one with the widest disparity to me, mostly in the "loudness" factor where they're all markedly different. The original is a little smoother sounding and seems to have some sort of soft limit or slight distortion compared to the digital mic's, never gets quite as sibilant or loud in the upper registers. On the rest though I really couldn't tell. Then again honestly my ears were getting fatigued. It's super subtle differences that come out more in the mix as a bit more forward or backward, a little more focused etc. Didn't have time to check out the male vox yet.

TheShadowV 1st May 2018 01:43 AM

All of these virtual mic systems have their benefits. Slate’s has the most microphone emulations and Steven isn’t one to stop where he is. I’m sure there are more on the horizon; like a Classic FET Collection. The Townsend’s stereo capabilities far outweigh their accuracy of their emulations. As a room mic or for stereo capture of an instrument, the possibilities are amazing. Personally I thought the Antelope sounded closest to the originals. Except Slate’s FG12, which was my favorite emulation on that mic and which shines here. That said, I still preferred the originals by a lot. That’s not to say you can’t make a good recording with any of these mics.

What I would have liked to have heard was a source with some weight. I have found the low end to be very revealing on these mic systems. It’s one of the things I didn’t particularly like about the Slate VMS when I owned it; a lack of low end on the emulations that didn’t match their vintage counterparts. I haven’t heard the MKII versions though.

DigitalPapaya 1st May 2018 03:43 AM

Quote:

Originally Posted by TheShadowV (Post 13288806)
Slate’s has the most microphone emulations

I'm not certain that is true. AFAIK, Antelope has 17 mic emulations vs 13 for Slate.

TheShadowV 1st May 2018 06:08 AM

Quote:

Originally Posted by DigitalPapaya (Post 13288924)
I'm not certain that is true. AFAIK, Antelope has 17 mic emulations vs 13 for Slate.

I didn’t realize that.

Peller 1st May 2018 11:40 AM

Neat test, but I think it's important to point out that this approach undersells the Townsend system a bit. Whereas the other two are simply about recreating the on-axis response of vintage mics, the Sphere goes way further than this -- its ability to manipulate proximity effect and polar patterns after the fact, and equalise the off-axis sound is really cool. Unlike the Slate and Antelope systems, the Sphere does a lot more than 'just' modelling vintage mics. For me personally, those extra capabilities would totally win the argument in its favour if I were buying a modelling mic, but they don't really come into play in a test like this.

javamad 1st May 2018 07:09 PM

I also agree that the Townsend's dual capsule feature-set is not on show but to keep it apples-to-apples I think that is fine. The purchasing decision should be taken by weighing in other factors such as if you already have an Antelope or UAD interface for example. Will you be using stereo, etc. etc.

I am really happy to have heard the plain voice comparisons and I like the detail and effort that was put into making a pretty straight comparison possible. I thought the mixes had waaay too much reverb on the voice to be useful but no matter.

I have downloaded the audio files and will run through this in the studio tomorrow but from what I am hearing on earbuds today my overall impression is that all the microphones sound "good enough for rock 'n' roll" rockout

That said, I did detect a lot of differences on the 67 - on both female and male - but I think that the variations still sounded good.

I have to admit that I listened with my own personal confirmation bias .... I am currently kitting out a small studio and it has come time to buy some good mics (47, 67, 87, 251) and after going over budget on the build I am left with some tough decisions to make. I was just making a list to compare my options as follows:

Official mics (brand names, vintage or not)
Good Clones
Cheap Clones
Virtual

and I was preparing the usual excuses/reasons to opt for each yingyang. I have a couple of UAD Apollos so I wanted the Townsend to be great. I didn't necessarily care if it beat the Slate or the Antelope I just wanted it to be great. And it is. kfhkh

Up to today, I was not really paying much attention to the Virtual option but I think once I listen to the downloaded tracks at the studio I might bring this option up in the order or things. Up to now I was pondering forgetting about the German mic dream for the moment and going for a set of Stam clones or something but now I think this warrants some further thought. A Pair of Townsends could be pretty handy to have around from what I am hearing.

This is to get things started in the studio and I do have the intention of swinging back around for some 'real' mics.

anon 1st May 2018 08:49 PM

What was the point of doing that test and then getting a guy that can't sing at all, that test was ridiculous, get an actual singer to do the test.

nikoli 1st May 2018 10:41 PM

Seems like a few people have noticed the differences in 67 and the Townsend.

Makes me wonder if that’s a result of Townsend’s modeling of the 67. Or perhaps the result has something to do the proximity features that Peller pointed out earlier.

Steven Slate 1st May 2018 11:45 PM

Quote:

Originally Posted by nikoli (Post 13290560)
Seems like a few people have noticed the differences in 67 and the Townsend.

Makes me wonder if that’s a result of Townsend’s modeling of the 67. Or perhaps the result has something to do the proximity features that Peller pointed out earlier.

Thanks Sweetwater and Gearslutz for this great comparison. What I think is immediately clear is that all three products can provide a great result, and that it’s a fantastic time to be making music!

Now, regarding the 67, in defense of Townsend, out of all the mics we examined in the modeling process, 67s varied THE MOST. What makes it even more difficult is that many vintage 67s have been modded. Our default 67 has the famous “filter mods” which open the top end. Our 67 in the Classic Tubes 3 upgrade is stock. They don’t sound the same at all!

Thanks again guys,

Steven

MickeyMassacre 2nd May 2018 02:47 AM

Quote:

Originally Posted by Peller (Post 13289308)
Neat test, but I think it's important to point out that this approach undersells the Townsend system a bit. Whereas the other two are simply about recreating the on-axis response of vintage mics, the Sphere goes way further than this -- its ability to manipulate proximity effect and polar patterns after the fact, and equalise the off-axis sound is really cool. Unlike the Slate and Antelope systems, the Sphere does a lot more than 'just' modelling vintage mics. For me personally, those extra capabilities would totally win the argument in its favour if I were buying a modelling mic, but they don't really come into play in a test like this.

FWIW the Antelope is a dual capsule system as well.

bill5 2nd May 2018 03:44 AM

Mic tests that aren't blind have no value IMO. Confirmation bias is stronger than the Force is with Luke Skywalker. Also how about some more mics that us average joes can actually afford vs all the high-priced stuff?

tkaitkai 2nd May 2018 05:57 AM

Quote:

Originally Posted by bill5 (Post 13291050)
Mic tests that aren't blind have no value IMO. Confirmation bias is stronger than the Force is with Luke Skywalker. Also how about some more mics that us average joes can actually afford vs all the high-priced stuff?

How hard is it to simply appreciate the fact that a team of people took time out of their professional lives to provide informational material to you free of charge?

mdme_sadie 2nd May 2018 07:32 AM

Quote:

Originally Posted by nikoli (Post 13290560)
Seems like a few people have noticed the differences in 67 and the Townsend.

Makes me wonder if that’s a result of Townsend’s modeling of the 67. Or perhaps the result has something to do the proximity features that Peller pointed out earlier.

I wanted the Townsend to really shine because the tech seems amazing and I have a UAD, but am I wrong that there's a slightly processed sound quality to it on the breathy sounds? Like you really hear the back of the throat sounds as if there's a compressor working overtime. Maybe it just needs a better screen.

Pianolando 2nd May 2018 07:39 AM

Quote:

Originally Posted by bill5 (Post 13291050)
Mic tests that aren't blind have no value IMO. Confirmation bias is stronger than the Force is with Luke Skywalker. Also how about some more mics that us average joes can actually afford vs all the high-priced stuff?

I agree that all audio comparisons are much more valuable if listened to blindly, but surely you can easily take these files and put in a ABX software? This one for instance, works on both mac and PC:
ABX/Shootout audio testing tool

Whitecat 2nd May 2018 11:08 AM

Quote:

Originally Posted by bill5 (Post 13291050)
Also how about some more mics that us average joes can actually afford vs all the high-priced stuff?

I guess the point of these modellers is to do the holy-grail stuff... where you get a mic locker full of tens of thousands of dollars of unobtanium. The "average joe" mics are not really ripe for modelling (yet) because I guess they are more accessible in the first place.

BackHand 2nd May 2018 01:04 PM

I have now listened to the 24/94 clips on my AKG 550 MK3 with Sonarworks 4 on flat mode. I also cheated by using Izotope insight to compare frequencies.

I think the result is a quite even race. All of the microphones had a microphone that they sounded closest to. The differences was so small that the difference in the actual microphone being modeled is most probably the biggest reason of any differences.

All and all I am very impressed. This clearly feels like a new era where there is no good reason of buying anything else than virtual microphones.

Virtual studio is the future.:cowbell:

Lynn Fuston 2nd May 2018 01:35 PM

Quote:

Originally Posted by BackHand (Post 13291662)
I think the result is a quite even race. All of the microphones had a microphone that they sounded closest to. The differences was so small that the difference in the actual microphone being modeled is most probably the biggest reason of any differences.

Anyone who has used a wide variety of 40- or 50-year-old vintage mics knows that there are great variances from mic to mic. The variance in the sounds represented here is probably smaller than the differences you'd find between two 47s or 67s. Who knows what changes have been made to the vintage mics, unless you're buying a used mic that has been in the hands of a single owner since new and he has records of all the mods/repairs that have ever been executed on it.

Lynn Fuston 2nd May 2018 01:37 PM

Quote:

Originally Posted by bill5 (Post 13291050)
Mic tests that aren't blind have no value IMO. Confirmation bias is stronger than the Force is with Luke Skywalker. Also how about some more mics that us average joes can actually afford vs all the high-priced stuff?

Thankfully you can download all the files, then have someone else rename them and you can listen blind. That solves that problem.

bill5 2nd May 2018 06:21 PM

Quote:

Originally Posted by tkaitkai (Post 13291175)
How hard is it to simply appreciate the fact that a team of people took time out of their professional lives to provide informational material to you free of charge?

Pretty hard, since
1. It wasn't provided for or to me personally as you seem to imply
2. As I already said, it has no value to me. That's simply a statement of such shootouts in general, not this specific one, and I'm not trying to slam them in saying that, I just stated a fact. It's like if someone took time to do an in-depth study of diff brands of tractors, I wouldn't appreciate it either, because it has no value to me. If others did, good for them.
3. Not following "took time out of their professional lives" as if this was some altruistic act of charity. This is a marketing tool by Sweetwater to get people interested in their site/products. (And certainly nothing wrong w/that)

Clearer now?


Quote:

Originally Posted by Whitecat (Post 13291506)
I guess the point of these modellers is to do the holy-grail stuff... where you get a mic locker full of tens of thousands of dollars of unobtanium. The "average joe" mics are not really ripe for modelling (yet) because I guess they are more accessible in the first place.

? If they're more accessible, I'd think that would by WHY they are ripe for modelling vs the other way around. But again, to each their own.



Quote:

Originally Posted by Lynn Fuston (Post 13291717)
Thankfully you can download all the files, then have someone else rename them and you can listen blind. That solves that problem.

Good point. Could you do that? ;)

nikoli 2nd May 2018 07:02 PM

Quote:

Originally Posted by bill5 (Post 13292355)
2. As I already said, it has no value to me.

That’s not what you said. Any claim of “fact” you are making now is fallacious and intellectually dishonest.

What you have is an opinion. Everyone here thinks it sucks. Probably would be best for you to stop derailing this thread now.

nikoli 2nd May 2018 07:09 PM

Quote:

Originally Posted by mdme_sadie (Post 13291264)
I wanted the Townsend to really shine because the tech seems amazing and I have a UAD, but am I wrong that there's a slightly processed sound quality to it on the breathy sounds? Like you really hear the back of the throat sounds as if there's a compressor working overtime. Maybe it just needs a better screen.


I think it sounds pretty great. Seems to me the only discrepancy likely boils down to the specific 67 that was modeled.

Personally, I’d love to own the Townsend.

DougS 2nd May 2018 07:22 PM

Thanks for putting this together. To me this provides a lot of value. kfhkh

bambamboom 2nd May 2018 07:23 PM

Quote:

Originally Posted by Steven Slate (Post 13290705)
Now, regarding the 67, in defense of Townsend, out of all the mics we examined in the modeling process, 67s varied THE MOST. What makes it even more difficult is that many vintage 67s have been modded. Our default 67 has the famous “filter mods” which open the top end. Our 67 in the Classic Tubes 3 upgrade is stock. They don’t sound the same at all!

This is a very important point.

If there is a mic that people notice a few differences on between the emulations and the original, it is not necessarily because of any inherent weakness in the modeling technology.

It is primarily because no 2 vintage mics sound exactly the same. Blackbird for instance has about 25 Telefunken 251's and they all sound slightly different from one another.

Each of the companies doing the modeling have to decide on a reference point that they think is a good representation of a classic mic, and they model that reference point VERY accurately.

Change the mic and now you've changed the reference point. Of course it will sound slightly different.

All of the debates about "is the technology there yet" need to end. It's more than fine. If people like the sound of the Slate, Townsend or Antelope choices (of what those companies thought was a good reference) then that should be the end of it - time to make music!

Wolf LeProducer 2nd May 2018 07:27 PM

Quote:

Originally Posted by Wolf LeProducer (Post 13288511)
I would love to see the results of say about, "5," very super famous mastering engineers brought into the Sweetwater studios for a blind test. Can they pick out which mic, was on which recording without knowing in advance...

There is a very specific reason why I said this, and dropped it.

When I wrote this thought out, I even began to think, "we could make it multiple choice." Then I scratched voicing that because my next thought was, "multiple choice tests are totally inaccurate."

The thing is, I wrote this from the perspective of, "it would be interesting." But not much more accurate than what has already been provided.

You use own ears, and your own judgement. The reality that technology has made unbelievable leaps and bounds.

Forget these modelling technologies.

Any $1000 mic from a good brand purchased today, has enough sound quality to make it on a hit record.

If you spend $2000


And how many mics for $800 are being used left and right.

These modelling mics are just unreal. True examples of where recording tech is in 2018.

You need a $35000 major label mic? Good for you. It's called "$1200" now.

and bill5 you are welcome to your opinion, I also disagree