OrangeTree Evolution Mandolin - very disappointed

[fret_man]

Firstly, I apologize for this long message, but I need to share my frustration recording mandolin.

I have a very difficult time tracking myself due to having no soundproofed rooms available. And now that it's nice outside it is even worse with dogs, kids, street noises, rain, wind, etc outside. I've tried RX8 and have had some success but it takes a long, long time to get a result that's clean and not all phasey/watery sounding.

So I have switched to just go ahead and play the piece using Midi-Guitar or Melodyne to convert it to MIDI. This captures my timing and dynamics, but I have to edit it to restore hammer-ons, pull-offs, and set where on the fretboard I'm playing. Sometimes the MIDI conversion misses some of the faster trills so I have to restore those as well. Lots of work but I can do it MUCH faster than waiting for quiet to record clean takes and/or using RX8.

So now it comes down to finding a very good mandolin VSTi. My vision of a good clean mandolin sound can be heard here: https://www.youtube.com/watch?v=gHMjGQ3iNAM. Yes, he plays it too fast - showing off, I guess, cuz he can. But he has a great sound. I favor the sound of archtop "bluegrass-style" mandolins as opposed to the classical "potato bug" bowl-back models. Those seem much too shrill and strident to me and is what I expect to hear when the Sheriff walks out of the saloon onto a dusty main street to stare down his ugly nemesis in a cheap Western movie. Flat-top Celtic models are better. But archtops have that magic woodie quality to them while the other models are too steely. So that's the desired sound I'm looking for.

Also, the only fret buzz you hear in that video is on the low G string. Us players try very hard to not get that. The same is true with fret noise as the fingers slide up/down and note release noises (which is really another form of fret buzz). It is important to me to not have those if I don't want them.

My musical style swings from Classical, Bluegrass, Folk, New Americana, Classic Rock, World Music, ... I guess quite a range there. My current projects are mostly duet & ensemble instrumentals so realism is important since the mix is so sparse.

So, armed with those goals, I've listed all the mandolin VSTi's I've found below:

1.  https://8dio.com/instrument/advanced-guitar-series-mandolin-for-kontakt/ ($118). This IS a potato bug and is very thin sounding. It may be a great classical model, but not for me. ?
2.  https://www.vir2.com/instruments/acou6tics/ ($400 for 6string, 12string, classical, guitalele, Uke, Mandolin). May be quite the deal but it doesn't sound very good to me. Again, a very thin sound. ?
3. https://www.indiginus.com/the-mandolin ($59). Sounds like another potato bug (even though their picture shows an archtop) for cheap western movie music. ?
4. https://www.boldersounds.com/index.php?main_page=product_music_info&products_id=122 ($60). Sounds passable. I've heard nice things about his stuff so, lucky for me, I picked this up on sale a few days ago. ?
5. https://www.orangetreesamples.com/products/evolution-acoustic-bundle ($790 for Songwriter, Flatpick, Classical,  Steel Strings, Jumbo 12, BG Banjo, Mandolin). Again, sounds promising. The banjo does, too. ?

My conclusions after playing with these last two are:

BolderSounds Pure Mandolin: This is not as tinny as the bowl-backs in 1, 2, 3 above, but not as full/woodie sounding as OrangeTree's mandolin. The killer for me, though,  is the A string pairs are not in tune with each other! They beat against each other no matter which fret you play and reminds me of a new, amateur musician that hasn't developed his ear yet. Sad, as this could have easily been caught and fixed before the sampling session. Also, I am convinced this is an archtop based on the documentation that came with the software, but it sounds like a cheap laminated model. My apologies to the owner but this just doesn't capture the richness these mandolins can have. ?

Update 2021-4-6: This VSTi contains 3 different instruments: Sustain (with a scripted tremolo), Tremolo (sampled), and Chop Chords. The Tremolo instrument has variable speed, which I like, and, strangely, the mistuned A string pairs don't sound so bad - probably due to the masking effects of all the other stuff happening during  the tremolo. So ? for their Tremolo instrument! It sounds very natural and is the best I've found so far. Unfortunately, it doesn't seem to handle double-stops (2 note chords that are tremolo'd together) very well. It sounds like two mandolins playing and results in lots of busyness. This seems to be true for ALL mandolin VSTi's so far.

So that leaves me with OrangeTree's Evolution Mandolin. It is the best of what I've found so far but there are still many problems:

1. It sounds like the action was too low on this mandolin because EVERY NOTE on EVERY STRING has fret buzz. I can minimize this by adjusting the velocity curve in Setup to favor soft and keep all velocities below 50 (which severely limits the dynamics I can achieve), but they're still not gone. Playing up the neck sometimes helps. Since I have to play at such low velocities, I have to turn up the output to restore the volume. This makes other items too loud, so those have to be turned down as well. Lots tweaking is required, but it is still impossible to get a clean sound. Edit: not sure what's happening but the more I use this the less this is an issue for me as long as I adjust the velocity curve to "soft" and play at low velocities. Maybe it was the new files I just re-downloaded? I seem to be making this work.
2. Some of the Round Robin samples still have sympathetic body vibrations even though they are turned off in Setup. Pick noise is a similar story. So is note release noise. Sure wish I could get rid of these but I'd like to keep the Round Robin sample cycles. This seems to be a post-processing editing error. Aarggh... Edit: not all notes have this problem, but some do, and only on some of the Round Robin samples.
3. I wish I could select fretboard location with keyswitches. That's how players play - moving around the fretboard. All I can do is a) force a single fretboard location for the entire song, and b) force which string each note is played on, which gets cumbersome when playing up the neck. That's a lot of manual keyswitch placement but it works. This seems like an oversight to me since much cheaper VSTi have this keyswitched fretboard location feature (like BolderSounds). Edit: Fret Position CAN be selected via MIDI CC command #079. I haven't gotten this to work yet but I'll assume this will do what I want.  So I retract this statement.
4. Since I have to play at such reduced velocity I noticed that pick noise does NOT track velocity! This doesn't follow reality. Hit the string hard -> loud pick noise -> loud note volume. I might even consider this a bug rather than a feature. Edit: I'm told pick noise DOES track velocity. Yet I have to reduce Pick Noise even more to maintain balance for me.
5. The fret noise engine is broken. Fret noise occurs when the fingers slide along the lower wound string pairs. It seems there needs to be enough space between the notes before fret noise will be inserted. Ok, I can work around that. Well, when I play a note, then a rest, then play the same note again on the same fret, fret noise is placed in the rest! Guys, there's no finger sliding when you're playing the same note back-back! I think I even heard it when I was playing on the high unwound strings! I just turn this off since it doesn't work the way real life does. Sheesh...
6. Bugs - the worst involves pick modeling, which is a cool concept as it models how close/far the right hand plays to the bridge. Unfortunately, if you enable it, then change your mind and disable it, volume goes away and tonality is very light/brittle. If you go back to enable it, the bug stays. The natural middle position has the same quiet/brittle sound. You have to delete/reload the whole instance of the instrument to restore it. Lesson: don't bother with pick modeling until you're ready to leave it set. Update 2021-4-6: I re-downloaded the library per Greg's suggestion and this problem has gone away! Not sure what happened, but Yay!
7. Hammer-ons and Pull-offs sound artificial and awful. I usually find a way to do without them even though it's not what I want. How could they accept this as OK?
8. Resonance is normally thought of as sympathetic string vibrations. If you play a high D, the low open D string starts vibrating. That's part of the sound of acoustic instruments, but that's not what this is. For OrangeTree, resonance is sounds like the vibrations of the strings behind the bridge and/or behind the nut. Anyone who has done enough recording knows to wind some leather, cotton, or rubberbands thru those strings to dampen their vibrations because mics will definitely pick those up. Why would anyone want to hear those is beyond me. All the players I know take steps to keep them from vibrating. I wish they had modelled the desired sympathetic vibrations instead.

So, to summarize, I am somewhat disappointed in OrangeTree. I'll have to wait for Ample Sounds to do their mandolin. They seem to do things mostly right. In the meantime, I'm making due with OrangeTree, working around its short-comings.

But I do appreciate an value Greg's responsiveness and activity on his forum. He seems to be one of the good guys.

Please let me know if I'm missing something. I would LOVE to get OrangeTree working better.

Edited April 7, 2021 by fret_man

[Byron Dickens]

I have had good results recording acoustic guitar and Bouzouki with the iRig acoustic mike. Does a pretty good job of isolating the instrument.

 

You can also make a booth out of moving blankets which, while not perfect, helps considerably.  I have done that in the past. I made a frame out of PVC pipe and hung the blankets with tarp clips and shower curtain rings. Blackout curtains on the windows help some, too.

[Vernon Barnes]

Greg at Orange Tree is very open to suggestions and it may well be worth communicating with him about improvments. 

[Tezza]

Can you get decent pickups for the mandolin? Is it possible to use a not so sensitive microphone? Can you record only at a quiet time?

The only way for you to get want you want is to record live in a professional studio with great mics. Everything else is going to be a compromise, I've had the same battles with Acoustic guitar. I have a quiet time around 7.30 - 10pm so if I want to do vocals or acoustic guitar it has to be then to stop outside noises, but that does not correct for a poor room reverb. For Vocals, I've fixed this by getting closer to the mic and using hanging dunas and that works reasonably well. For acoustic guitar, I generally record it separately at the right time with sound deadening material but still not ideal. I get a better but different sound with a Shure 57 than I do with my AKG C214 condenser because the dynamic mic just doesn't pickup room reverb or extraneous sounds as much, again, a compromise.

I can transcribe what I play to midi and run a VST but it won't sound the same as what I play and never will, so it's a compromise. With the VST solution, there is always something wrong or different. For example, I like the reduced fret noise and finger squeak in VST midi recordings, I don't experience that problem. There's some that allow you to set the fret (playing area) so that's good. But then some of them have uncontrollable round robin up/down strokes and the way I play, sliding up/down, playing in triplets and having hammer ons/pull offs etc conform to that signature doesn't work. My natural playing is not recorded. It might be ok for some things but it takes some fiddling around.

I recently got a Jazz Archtop guitar that has 2 humbucker pickups and have been experimenting with that. Different experience, I can get up at 6.30 and record stuff in a bad room with roosters, angle grinders, dogs barking, parrots, etc going off and when I play it back, it sounds like a pristine studio grade recording, the humbuckers don't pick up the acoustic cacophony going on outside of the guitar string . It's a different sound to a mic recorded acoustic guitar but with the style I play, I only need to make some small changes to my playing style for it to work...again, just a bit of a compromise. I only just got the guitar so still experimenting but it is such a relief to be able to record when I want, where I want, without having to fiddle with midi and end up with a really high quality professional sound straight off the bat.

I don't know what your playing style is like, whether it's suited to pickup style recording or whether there exists decent pickup systems for mandolins but I would highly recommend trying it out to see if it suits, it fixes a lot of problems. There will no doubt be a compromise of some sort, that might be acceptable to you or not. I am finding that I am starting to prefer the sound of the Jazz guitar for my style of playing anyway, so for me it may end up not being a compromise at all, just an improvement. There are still some issues with pickup height and string balance etc but I'm sorting those out at the moment.

 

 

Edited April 2, 2021 by Tezza

[Byron Dickens]

Along those lines, I would also add: don't make perfect be the enemy of good enough.

[fret_man]

Thanks, guys.  I forgot my mandolin has a built-in Schertler pickup. I should have remembered because I paid a pretty penny for that and it sounds fantastic live! Will give that a shot. Also, I noticed that the longer I sit there mucking with things the more I hear problems and get more frustrated. Gotta take a step away more often.

[jerrydf]

If you record with your pickup,  you might want to try a mandolin impulse response. There's some good ones around and they tame the piezo sound of the pickup. 

[fret_man]

Thanks, Jerry. The Schertler is a sound body / microphone pickup that sounds very realistic live. No piezo quack here! Still, I'll look into the IR idea. I didn't know those existed for mandolin modelling.

[Jim Fogle]

@fret_man, Thank you for sharing your experience seeking out the "perfect" mandolin VSTi.  I very much appreciated that you included links to the VSTis you mentioned.  While not in the market for a Mandolin VSTi at this time, the feedback you provided is helpful just the same.

[fret_man]

FYI, I found out it IS possible to specify the fret  position by using MIDI CC. Unfortunately, I have yet to find out which CC. I requested a list of the various CC options but haven't seen any yet. It is not in the user guide.

[Matthew Sorrels]

14 hours ago, fret_man said:

FYI, I found out it IS possible to specify the fret  position by using MIDI CC. Unfortunately, I have yet to find out which CC. I requested a list of the various CC options but haven't seen any yet. It is not in the user guide.

Can you right click on the Fretting Position number in the setup tab and then assign it any CC# you want with the learn?  I don't think it comes assigned by default.  There are some Host automation parameters but using those instead of MIDI can be tricky.  #079 maybe?

Edited April 5, 2021 by Matthew Sorrels

[fret_man]

Thanks, Matthew Sorrels. Yes, I think #079 is correct. According to their Acoustic Slide manual there are two ways to do this: 

This is achieved by two methods. Firstly, you can right-click on any automatable knob and use the MIDI learn option to assign the MIDI CC #. The second method is to manually drag a MIDI CC # from the listing of CCs in KONTAKT's automation section on KONTAKT's left sidebar onto one of the knobs on Steel Strings' interface. This is accessed in the "Auto", then "Midi Automation" tab.

I wish this was in the Mandolin manual as well.

[Greg Schlaepfer]

 I'm sorry to hear you're disappointed with our mandolin sample library!

Hopefully I can give you a little insight on the library that might be helpful. And if the library doesn't seem like it's going to work out for you, please get in touch with me through our support form (there's a direct email address on that page, too).

On 3/31/2021 at 8:23 AM, fret_man said:

It sounds like the action was too low on this mandolin because EVERY NOTE on EVERY STRING has fret buzz. I can minimize this by adjusting the velocity curve to favor soft and keep all velocities below 50 (which severely limits the dynamics I can achieve), but they're still not gone. Playing up the neck sometimes helps. Since I have to play at such low velocities, I have to turn up the output to restore the volume. This makes other items too loud, so those have to be turned down as well. Lots tweaking is required, but it is still impossible to get a clean sound.

As you mentioned, there's fret buzz, particularly on the loudest dynamic and on the G string. We try to sample a wide dynamic range, meaning that the loudest dynamic is going to naturally give you that fret buzz from when you really dig into the strings. Any fret buzz on the soft or medium dynamic is not intentional, so if you're running into any fret buzz elsewhere, that's something I'll take a look at for an update to the library.

For more graceful parts, I'd recommend lowering the dynamic curve setting so that it's biased towards the softer dynamics. Admittedly the default velocity scaling should probably put the loudest velocity a little higher so that it doesn't happen so readily.

Overall, it's a delicate balance, making sure to record imperfections like fret noises, buzz, and other extraneous noises are included in order for the library to sound more human, while still allowing you to get purer, more pristine playing when needed. It's my experience that if you sample an instrument too cleanly, it tends to sound robotic and lifeless--which is ironic, because as a player, you strive to eliminate all those little extraneous noises (as you mentioned). Noises like that are what help sample libraries sound more believable, which is one of the reasons we included sampling a misfretted sustain articulation that you can mix into faster passages to help with the realism.

On 3/31/2021 at 8:23 AM, fret_man said:

Some of the Round Robin samples still have sympathetic body vibrations even though they are turned off. Pick noise is a similar story. So is note release noise. Sure wish I could get rid of these but I'd like to keep the Round Robin sample cycles. This seems to be a post-processing editing error. Aarggh...

You can change the volume/amount of added pick noise in the SETUP section of the interface, and if you need the releases to sound cleaner and more subtle, you can reduce their volume in the SETUP section as well.

In the future I'd like to add more control over the release samples to allow you to adjust the smoothness of the transition between the sustain and release samples, and possibly a way to limit the length of the release samples, too.

On 3/31/2021 at 8:23 AM, fret_man said:

Since I have to play at such reduced velocity I noticed that pick noise does NOT track velocity! This doesn't follow reality. Hit the string hard -> loud pick noise -> loud note volume. I might even consider this a bug rather than a feature.

The pick noise actually does track velocity, I just double-checked to be sure. Though as I mentioned before, if you need to reduce the amount of added pick noise, you can do that from the SETUP section of the library's interface. At 100%, it's still audible at lower dynamics, so you might try something like 25-50% instead (or just leave it off, of course).

On 3/31/2021 at 8:23 AM, fret_man said:

The fret noise engine is broken. Fret noise occurs when the fingers slide along the lower wound string pairs. It seems there needs to be enough space between the notes before fret noise will be inserted. Ok, I can work around that. Well, when I play a note, then a rest, then play the same note again, fret noise is placed in rest! Guys, there's no finger sliding when you're playing the same note back-back! I think I even heard it when I was playing on the high unwound strings! I just turn this off since it doesn't work the way real life does. Sheesh...

It all depends on where the fretting position is set to. If you play a note outside of the current fretting position, half a second later (or so), it triggers a position change noise to simulate the player returning to the desired fretting position. If you're playing a lot of notes outside of that fretting position, try moving the fretting position control to that range instead. The fretting position can also be automated to a MIDI CC if you want to control it in real-time. I usually just leave it in one general area, though. And as you noted, you can always disable the automatic fret noises and manually trigger them using the performance effect key for the fret noise, which is mapped above the main playing range.

On 3/31/2021 at 8:23 AM, fret_man said:

Bugs - the worst involves pick modeling, which is a cool concept as it models how close/far the right hand plays to the bridge. Unfortunately, if you enable it, then change your mind and disable it, volume goes away and tonality is very light/brittle. If you go back to enable it, the bug stays. The natural middle position has the same quiet/brittle sound. You have to delete/reload the whole instance of the instrument to restore it. Lesson: don't bother with pick modeling until you're ready to leave it set.

I tried it on my end and can't reproduce what you're running into. It's also strange that the middle position would have a quiet/brittle sound, since it should sound pretty much just like when the pick modeling is disabled. If you load the library in the standalone Kontakt application (just to eliminate some variables), do you run into the same issue?

On 3/31/2021 at 8:23 AM, fret_man said:

Hammer-ons and Pull-offs sound artificial and awful. I usually find a way to do without them even though it's not what I want. How could they accept this as OK?

We use real samples of hammer-ons, pull offs, and slides. However, I do have some ideas on how to improve their realism through scripting, particularly when it comes to playing trills or depending on the current duration of the note. I'm also open to any recommendations you have, and as I mentioned before, feel free to get in touch with me via email!

On 3/31/2021 at 8:23 AM, fret_man said:

Resonance is normally thought of as sympathetic string vibrations. If you play a high D, the low open D string starts vibrating. That's part of the sound of acoustic instruments, but that's not what this is. For OrangeTree, resonance is the vibrations of the strings behind the bridge and/or behind the nut. Anyone who has done enough recording knows to wind some leather, cotton, or rubberbands thru those strings to dampen their vibrations because mics will definitely pick those up. Why would anyone want to hear those is beyond me. All the players I know take steps to keep them from vibrating. I wish they had modelled the desired sympathetic vibrations instead.

There's a resonance amount setting in the SETUP section that should be helpful to dial in how much resonance you want--this feature recreates the sympathetic resonance between strings. There's still going to be body resonance in the articulations themselves, and in some cases possibly a little sympathetic resonance. The engineer we worked with is no stranger to recording mandolin, so muting the string ends and tracking down any other little unwanted vibrations before recording instruments is a routine part of the process.

Anyway, hopefully some of that information is useful to you, and I do appreciate any feedback and suggestions you have, too.

[Greg Schlaepfer]

11 hours ago, fret_man said:

This is achieved by two methods. Firstly, you can right-click on any automatable knob and use the MIDI learn option to assign the MIDI CC #. The second method is to manually drag a MIDI CC # from the listing of CCs in KONTAKT's automation section on KONTAKT's left sidebar onto one of the knobs on Steel Strings' interface. This is accessed in the "Auto", then "Midi Automation" tab.

I wish this was in the Mandolin manual as well.

I should really add that to all the Evolution guitar libraries' manuals. It's part of Kontakt's own operation, so it's documented more thoroughly in that manual, but I can see that it would be useful to include that tip in the Evolution manuals as well.

Honestly, it would probably be worth making a short video tutorial about automation. There are some cool things can be done from the automation panel that people might not realize, like adjusting the min/max ranges of the automated MIDI CC, or even assigning the same MIDI CC to multiple parameters.

[fret_man]

Thank you, Greg. I really, really appreciate your responses to my tirade above. I don't want to be that guy who goes off halfcocked with unwarranted complaints. Maybe I had unrealistic expectations but guitar VIs are so good now-a-days that I figured Mandolins would be there as well. But I'm thinking not quite yet. I made a few edits above based on what I've learned from playing around some more with the new file I recently downloaded. That helped a lot. I admit I need to learn Kontact much better.

[fret_man]

On 4/3/2021 at 3:22 AM, jerrydf said:

If you record with your pickup,  you might want to try a mandolin impulse response. There's some good ones around and they tame the piezo sound of the pickup. 

So, I recorded two channels: 1 thru a mic and 2 thru DI. I played each note from the lowest to the highest (~3.5 octaves)and measured the RMS of each channel for each note. By comparing these I generated an EQ curve to convert the DI sound to be more like the mic and it sounds GREAT! The DI was a little boxy sounding (which the EQ gloriously fixed) and the DI seems to impart some sort of compression effect because it's notes don't die down as quickly. It's probably not the transducer per se but more due to sound transmission thru the maple wood, bracing, etc. Still, it is a highly useable signal and will definitely solve my tracking issues.

I'm glad I brought up my frustrations because y'all made some very good suggestions. Even tho' I use Studio One now, the Cakewalk Forum has always been a great place to stay and share. Thanks!

[jerrydf]

7 hours ago, fret_man said:

So, I recorded two channels: <snip> ...

Yes, that's a fair approach.

If you want to go a little deeper, you could monitor the comparative frequency domains of the ranges of notes and derive the EQ from there. This could be by taking a simple audio editing program (like Audacity) and analysing the harmonics generated for each note (or at least evry few ful-tones or so) from the pick up compared to the microphone version. This would show where the pick up is adding or subtracting harmonic content throughout the range.  You may end up with something like the EQ you already have, but it would be a good scientific check. 

However that doesn't cover the other subtleties of things like phase-shift responses both within the air space of the mandolin body,  or of those conducted through the body. There's not a lot you can do about that side of things on a simple EQ correction. Maybe you could create your own IR from your own instrument.  Of course, now we're into professional lab/studio set-ups and selected microphones.

jdf

[fret_man]

And in case anyone's curious, the EQ I came up with looks like this: