The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Hollywood Strings in Logic: Does it actually work? Virtual Instrument Plugins
Old 11th August 2014
  #61
Lives for gear
 

A good drum or cymbal sample takes a LONG time to decay. Take a well sampled drum kit. A snare hit for example has a mono sample for the top and bottom head, a stereo sample for the overheads and ambient mics, then spill samples from the rest of the kit mics. Suddenly that ONE hit is using 12 voices of polyphony. Factor in the long decay and a 16th note fill is chewing up polyphony like a bastard. Yes you could hire a real drummer but then you have to mop up all the drool afterwards. (I kid 'cause I love.)

Do an up tempo Hollywood style trailer with a wall of epic ambient percussion with lots of round robin samples, (you guys know what RR samples are right?) and in the space of a bar of music.... say goodbye to hundreds of voices of polyphony.

If you make a living out of this sort of thing then high polyphony counts count which neatly takes me back to Play...... It sucks at this sort of thing.
Old 11th August 2014
  #62
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by Marsdy View Post
A good drum or cymbal sample takes a LONG time to decay. Take a well sampled drum kit. A snare hit for example has a mono sample for the top and bottom head, a stereo sample for the overheads and ambient mics, then spill samples from the rest of the kit mics. Suddenly that ONE hit is using 12 voices of polyphony. Factor in the long decay and a 16th note fill is chewing up polyphony like a bastard. Yes you could hire a real drummer but then you have to mop up all the drool afterwards. (I kid 'cause I love.)

Do an up tempo Hollywood style trailer with a wall of epic ambient percussion with lots of round robin samples, (you guys know what RR samples are right?) and in the space of a bar of music.... say goodbye to hundreds of voices of polyphony.

If you make a living out of this sort of thing then high polyphony counts count which neatly takes me back to Play...... It sucks at this sort of thing.
Woah, wait. The 16th note, on a snare? You realize whenever you 'hit' the snare, the sound is effectively restarted (I am talking real life, not your whacky samples), even with a round robin - the original note should not be still playing (reverb aside).

The maximum polyphony for orchestral scores should never be more than the number of instruments, times the number of possible notes a person can play ( say a piano) or number of total notes that can be struck/strummed like a harp (if you pluck the same string twice, the original 'pluck' should not continue), multiplied by the number of mics (4 at most). If you use more than that, it won't sound like a real orchestra at all and your 'score' won't sound like anything of the sort.

Even emulating the loudest and most complex crescendo of a full symphonic orchestra should use little more than 1,000 (100 instruments, 4 mics per instrument (actually an overkill on most instruments) and perhaps allowing for 50 - 60 note polyphony on the polyphonic instruments.

I have some pretty complex orchestral scores, and I try to keep it as realistic as possible (18 violins can only play one note at a time each) - and my max polyphony is in the 300 notes range.
Old 11th August 2014
  #63
Lives for gear
 

Quote:
Originally Posted by Zoolook View Post
Woah, wait. The 16th note, on a snare? You realize whenever you 'hit' the snare, the sound is effectively restarted (I am talking real life, not your whacky samples), even with a round robin - the original note should not be still playing (reverb aside).

The maximum polyphony for orchestral scores should never be more than the number of instruments, times the number of possible notes a person can play ( say a piano) or number of total notes that can be struck/strummed like a harp (if you pluck the same string twice, the original 'pluck' should not continue), multiplied by the number of mics (4 at most). If you use more than that, it won't sound like a real orchestra at all and your 'score' won't sound like anything of the sort.

Even emulating the loudest and most complex crescendo of a full symphonic orchestra should use little more than 1,000 (100 instruments, 4 mics per instrument (actually an overkill on most instruments) and perhaps allowing for 50 - 60 note polyphony on the polyphonic instruments.

I have some pretty complex orchestral scores, and I try to keep it as realistic as possible (18 violins can only play one note at a time each) - and my max polyphony is in the 300 notes range.
Wow..... What libraries are you using? Garritan?

Sorry dude but you have this COMPLETELY wrong although nice of you to edit out the "dick" comment. The idea that one note = one voice of polyphony is nonsense in the sample world.

No the snare "hit" is NOT simply "restarted" as you suggest. It's overtones and decay outlive the next hit before you even factor in ambience, air and spill. With a good library, hit a floor tom and it triggers a snare rattle that is picked up by the top and bottom snare mic samples, the O/H and ambient mic samples and the spill in all the other drum mics and so on. If the samples are "restarted" and kill off the previous hit you are taking instant "machine gun" effect. If you don't believe me just try it.... Set your drum sounds to one voice and see how realistic it sounds. Even a close mic'd drum has some room, air and spill captured in the sample.... In other words you need a LOT of polyphony... period... Take BFD for example.... I think I'm right in saying one hit triggers 12 voices. A ride cymbal can take a minute to decay at higher velocities. Eighth notes over the course of 1 bars at 120 bpm = 96 voices of polyphony. Over 8 bars 768 voices JUST for the ride. And as I say, if the next hit cuts off the previous one, it sounds terrible. You do NOT want the machine gun thing.

So let's talk about orchestral instruments. You mention harp.... ONE harp pluck can take many tens of seconds to decay in the lower registers if not more. Play a 24th note glissando and you can be using dozens if not a 100 voices in the space of a few bars. That's a third of the 300 notes you claim to be your maximum polyphony count.

These new fangled orchestral libraries huh? Don't you just love them. You play ONE note of a legato violin patch and five stereo samples are triggered to allow for the mod wheel dynamic crossfading, another overlapping batch of samples kick in for the legato transition which overlap with the next note, then there are those release triggers... It soon adds up. We are talking just ONE note here and 20, 30.... even more voices of polyphony are eaten alive. I'm amazed you've never come across this well known gotcha of the sample world. THIS is why sampler performance is important. THIS is why Play sucks.

So let's move on to those massive walls of epic cinematic percussion that are all the rage, the 8Dio stuff for example ....A single drum instrument with a four second tail built in. Play 16th notes at 120 bpm, that's 32 voices of polyphony after two bars. Multiply THAT by a whole cinematic percussion section, trailer stylee....OK... it may not be the art of music at its zenith but it's a LOT of polyphony. And did I mention those staccato strings playing along with the percussion... jeez... again with those greedy round robins...

300 voices.... Meh..
Old 11th August 2014
  #64
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by Marsdy View Post
Wow..... What libraries are you using? Garritan?

Sorry dude but you have this COMPLETELY wrong although nice of you to edit out the "dick" comment. The idea that one note = one voice of polyphony is nonsense in the sample world.

No the snare "hit" is NOT simply "restarted" as you suggest. It's overtones and decay outlive the next hit before you even factor in ambience, air and spill. With a good library, hit a floor tom and it triggers a snare rattle that is picked up by the top and bottom snare mic samples, the O/H and ambient mic samples and the spill in all the other drum mics and so on. If the samples are "restarted" and kill off the previous hit you are taking instant "machine gun" effect. If you don't believe me just try it.... Set your drum sounds to one voice and see how realistic it sounds. Even a close mic'd drum has some room, air and spill captured in the sample.... In other words you need a LOT of polyphony... period... Take BFD for example.... I think I'm right in saying one hit triggers 12 voices. A ride cymbal can take a minute to decay at higher velocities. Eighth notes over the course of 1 bars at 120 bpm = 96 voices of polyphony. Over 8 bars 768 voices JUST for the ride. And as I say, if the next hit cuts off the previous one, it sounds terrible. You do NOT want the machine gun thing.

So let's talk about orchestral instruments. You mention harp.... ONE harp pluck can take many tens of seconds to decay in the lower registers if not more. Play a 24th note glissando and you can be using dozens if not a 100 voices in the space of a few bars. That's a third of the 300 notes you claim to be your maximum polyphony count.

These new fangled orchestral libraries huh? Don't you just love them. You play ONE note of a legato violin patch and five stereo samples are triggered to allow for the mod wheel dynamic crossfading, another overlapping batch of samples kick in for the legato transition which overlap with the next note, then there are those release triggers... It soon adds up. We are talking just ONE note here and 20, 30.... even more voices of polyphony are eaten alive. I'm amazed you've never come across this well known gotcha of the sample world. THIS is why sampler performance is important. THIS is why Play sucks.

So let's move on to those massive walls of epic cinematic percussion that are all the rage, the 8Dio stuff for example ....A single drum instrument with a four second tail built in. Play 16th notes at 120 bpm, that's 32 voices of polyphony after two bars. Multiply THAT by a whole cinematic percussion section, trailer stylee....OK... it may not be the art of music at its zenith but it's a LOT of polyphony. And did I mention those staccato strings playing along with the percussion... jeez... again with those greedy round robins...

300 voices.... Meh..
It's funny... I read this and still didn't believe you, until I checked Kontakt myself - you're absolutely right - this is something I really didn't know. I plucked the 300 number out of thin air too... what a dick!

This is a project, one of 4 Kontakt instances... before your post, I would not have thought this would be any more than 20 or 30 notes of polyphony. No wonder I am always running out of resources.



Sorry for being a total tool.
Old 11th August 2014
  #65
Lives for gear
 

Quote:
Originally Posted by Zoolook View Post
It's funny... I read this and still didn't believe you, until I checked Kontakt myself - you're absolutely right - this is something I really didn't know. I plucked the 300 number out of thin air too... what a dick!

This is a project, one of 4 Kontakt instances... before your post, I would not have thought this would be any more than 20 or 30 notes of polyphony. No wonder I am always running out of resources.



Sorry for being a total tool.
No worries!!! A SSD is huge help here!

Coincidently I was using EW Hollywood Strings this morning. One of the BIG legato "powerful system" violin patches. On a simple monophonic part with slightly overlapping notes to trigger the legato transition, polyphony passed 50 voices!!!

Imperfect Samples 1908 Walnut Concert Grand sounds gorgeous but uses up to SIX voices for one note!! Play a fast arpeggio with the sustain pedal down and its not difficult to trigger 200 voices!
Old 11th August 2014
  #66
Lives for gear
 
EastWest Lurker's Avatar
 

For those who would like some insight into what makes HS tick, from EW: '

" Understand folks that HS is a unique approach and not comparable to other libraries. Some instruments in this collection require a ram buffer to handle the stub file loading of 20,000+ samples, not 2,000+ samples you would find in other collections. What is a stub file? It’s the amount of samples immediately loaded into a ram buffer to overcome latency (the time difference between hitting a key and hearing the sound), the rest of the samples in the instruments are streamed into and out of ram as needed".

AND:

"You couldn’t do this prior to 1998 when EW and Nemesys introduced GigaSampler to the world of sampling. That invention took out the market leaders in hardware samplers that were all ram based including Akai, Emu, Kurzweil and Roland.

We offer Gold and some cases Silver versions of our collections for those that don’t require that level of detail or multiple mic positions included in the Platinum/Diamond versions. Who else has orchestral strings with true legato, all these articulations, and 5 mic positions? It also sounds incredible, as evidenced by the amazing demos on the product page and number of awards it has received."
Old 11th August 2014
  #67
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by Marsdy View Post
No worries!!! A SSD is huge help here!

Coincidently I was using EW Hollywood Strings this morning. One of the BIG legato "powerful system" violin patches. On a simple monophonic part with slightly overlapping notes to trigger the legato transition, polyphony passed 50 voices!!!

Imperfect Samples 1908 Walnut Concert Grand sounds gorgeous but uses up to SIX voices for one note!! Play a fast arpeggio with the sustain pedal down and its not difficult to trigger 200 voices!
We're going to need a bigger boat...
Old 11th August 2014
  #68
Lives for gear
 

Quote:
Originally Posted by Zoolook View Post
We're going to need a bigger boat...
With a lot of RAM...
Old 13th August 2014
  #69
TGV
Gear Addict
 

I can imagine it surpasses 50 voices, although it does indicate problems with resource management. You've got to imagine that every aspect of the string sound that is being modeled requires a voice. Things are in stereo, so everything doubles. Then you need one voice per layer that can switch while holding a note. That would be a voice (or rather 2, since it's stereo) per dynamic layer, times 2, possibly more, for vibrato. Then there is a comparable amount of voices for the legato transition. So if there are 4 dynamic layers and 2 vibrato levels, you've got 2 x 4 x 2 x 2 = 32 voices. Then there are trailing release voices, so the count can exceed 50. It could be trimmed considerably by flexible voice scheduling while the note is constant, but when there's a lot of CC movement, voice count will necessarily be high.

I have my misgivings about Play too, in particular in combination with Symphonic Choir on the mac. It was so frustrating that I bought Voices of Prague and removed SC. Much happier now. I did buy HB Silver in the recent sale though, and it seems to work fine. Let's hope they keep improving Play, and be more open about its limitations.
Old 18th August 2014
  #70
Gear Maniac
 

i have Hollywood strings gold edition if anyone is interested in buying it from me
Old 18th August 2014
  #71
Lives for gear
 
EastWest Lurker's Avatar
 

Quote:
Originally Posted by On The Swarm View Post
i have Hollywood strings gold edition if anyone is interested in buying it from me
EastWest does not allow the resale of its products. This is clearly stated in the licensing agreement.
Old 18th August 2014
  #72
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by EastWest Lurker View Post
EastWest does not allow the resale of its products. This is clearly stated in the licensing agreement.
You've really got the PR down haven't you?
Old 18th August 2014
  #73
Lives for gear
 
EastWest Lurker's Avatar
 

Quote:
Originally Posted by Zoolook View Post
You've really got the PR down haven't you?
It is not my job to do PR. It is my job to give people factual information and that is simply a fact.
Old 18th August 2014
  #74
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by EastWest Lurker View Post
It is not my job to do PR. It is my job to give people factual information and that is simply a fact.
"Clearly" is a subjective value statement, so you went well beyond stating a fact and seemed to make an effort to antagonize the poster. You could have sent a PM to them instead.
Old 18th August 2014
  #75
TGV
Gear Addict
 

Quote:
Originally Posted by Zoolook View Post
You've really got the PR down haven't you?
I'm not fond of EW, but it's only honest to mention it.
Old 18th August 2014
  #76
Lives for gear
 
EastWest Lurker's Avatar
 

If I send a PM I am only giving that information to one person. If I post it, I am giving it to many.

I am certainly not trying to antagonize anyone intentionally but sometimes that happens unfortunately when you write things that you either know to be true or believe to be true that people don't always want to hear.
Old 18th August 2014
  #77
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by EastWest Lurker View Post
If I send a PM I am only giving that information to one person. If I post it, I am giving it to many.
Perhaps you could credit the rest of us with already knowing?

Maybe I over reacted, but your comments seem a little self-righteous occasionally.
Old 18th August 2014
  #78
Lives for gear
 
EastWest Lurker's Avatar
 

Quote:
Originally Posted by Zoolook View Post
Perhaps you could credit the rest of us with already knowing?

Maybe I over reacted, but your comments seem a little self-righteous occasionally.

You would be surprised how many people do not know that because they do not read the terms of their agreement.

I accept that criticism. It is an unfortunate aspect of my personality that I will try to be more mindful of.
Old 15th September 2014
  #79
Lives for gear
 

Quote:
Originally Posted by EastWest Lurker View Post
It will work but since Logic does not spread patches within a multi-timbral instance throughout the cores, and since the patches are big and take a long time to load, I strongly recommend VE Pro.

Also, you are not going to be able to use that much with only 16 GB.
As the last paragraph at of this support article mentions
Logic Pro/Express: Tips for balancing multi-core performance

The following two videos show you can load multitimbral instruments in Logic Pro X and achieve core distribution following the above article, last paragraph.

Shows core distribution with Kontakt 5 loaded as a multitimbral. I show the steps to get core distribution.
https://www.youtube.com/watch?v=59YmTxZHKfM

41 instruments loaded across 4 instances of PLAY (v 4.2.0) and 2 instances of Kontakt 5 (v 5.3.1.37).
https://www.youtube.com/watch?v=Vdg8po01nwM

BTW - The last video shows running on a MacMini Server with 16GB RAM and 2 500GB hard drives at 7200rpm.

The last video is playing a song for around 11 mins with no issues and that's with all the tracks I have loaded playing all the time.
Some projects probably have 41 tracks but those tracks play now and then, this is showing 41 tracks playing all the time.
So once you get core distribution working, Logic Pro X has no issue on multitimbral instruments.
Old 16th September 2014
  #80
Lives for gear
 
EastWest Lurker's Avatar
 

That all works well for playback, ValliSoftware. The problem comes when you arm a track that has a multitimbral instance with a lot of demanding patches and/or auxes. At that point as all go into Live mode, you may get crackling or system overload messages, depending on the power of your computer and what else is going on in the project. However if you load that same multi in Vienna Ensemble Pro and connect to it in Logic Pro, it distributes it through the cores before it gets to Logic so that when you connect to it, all works better.
Old 16th September 2014
  #81
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by ValliSoftware View Post

The last video is playing a song for around 11 mins with no issues and that's with all the tracks I have loaded playing all the time.
Some projects probably have 41 tracks but those tracks play now and then, this is showing 41 tracks playing all the time.
So once you get core distribution working, Logic Pro X has no issue on multitimbral instruments.
Nice technical demo, but that music was... well, difficult to listen to for 11 minutes!
Old 16th September 2014
  #82
Lives for gear
 

Quote:
Originally Posted by Zoolook View Post
Nice technical demo, but that music was... well, difficult to listen to for 11 minutes!
Yeah, sorry about that.
I just needed to show constant playing to show that the CPU Core Distribution works in Logic Pro X.
If there was an issue, it would have shown up heh

Thanks for watching.
Old 16th September 2014
  #83
Lives for gear
 

#1
Quote:
Originally Posted by EastWest Lurker View Post
That all works well for playback, ValliSoftware.
Logic Pro X does handle CPU Core Distribution when you follow the Support Article.
Logic Pro/Express: Tips for balancing multi-core performance

#2
Quote:
Originally Posted by EastWest Lurker View Post
The problem comes when you arm a track that has a multitimbral instance with a lot of demanding patches and/or auxes.
Not sure why you would select that same multitimbral for recording since you just applied #1 from the Support Article.
By selecting the multimbral again, you defeat what you just accomplished from #1 which was core distribution, now your right back where you started from which is no core distribution because you selected the multimbral again.

If you want to have core distribution even while recording, you need to create a new track with the instrument you want to record.
Once you're finished, move that MIDI region to the MIDI track of the muiltimbral and reselect your empty track that you create from #1 to gain full core distribution again.



This shows that Logic Pro X handles core distribution just fine, when done correctly but also notice in the video while I show creating from three different instruments, the results on the CPU go from great to bad depending on the instrument. The video clearly shows which instrument handles core distribution while recording.

Instead of just saying buy this other $$$oftware, show in Logic Pro X that core distribution can be done either in playback or recording.
Old 17th September 2014
  #84
Lives for gear
 
EastWest Lurker's Avatar
 

I understand and have used that workflow. I find it at best inelegant and I am still telling you that you will find that with Vienna Ensemble Pro, you needn't jump through those hoops and yet core distribution will be better than Logic's. (Also, when switching projects, you will not need to wait for samples to reload. )

I have tried this every which way from Sunday, but people are free to either take my advice or not.
Old 17th September 2014
  #85
Lives for gear
 
Zoolook's Avatar
Quote:
Originally Posted by ValliSoftware View Post

If you want to have core distribution even while recording, you need to create a new track with the instrument you want to record.
Once you're finished, move that MIDI region to the MIDI track of the muiltimbral and reselect your empty track that you create from #1 to gain full core distribution again.
That's the way I have always done it anyway, because using a new track makes it easier to do over-dub recording, edits and roll-back to a prior version if you don't like the 3rd or 4th take. It's just cleaner, if a little old-fashioned. Also, back in the day when you had to freeze all the time, this would be the only way to layer MIDI on frozen tracks.

I think people who have been doing this for a while (for at least 15 years) never complain about not having enough CPU, because in 1997, your workflow was always being crafted around technical limitations. People who started figuring out their workflows in an age of apparently limitless RAM and CPU never seem to have enough. It's one of life's ironies.
Old 18th September 2014
  #86
Lives for gear
 

Quote:
Originally Posted by EastWest Lurker View Post
I understand and have used that workflow. I find it at best inelegant and I am still telling you that you will find that with Vienna Ensemble Pro, you needn't jump through those hoops and yet core distribution will be better than Logic's. (Also, when switching projects, you will not need to wait for samples to reload. )
It was never mentioned to me about workflow. I was showing a fact that Logic Pro X handles core distribution on playback on multitimbral instruments.

Quote:
Originally Posted by EastWest Lurker View Post
I have tried this every which way from Sunday, but people are free to either take my advice or not.
I merely replied to show others core distribution with Logic Pro X is possible without any additional $$$oftware.
Unfortunely you replied again regarding core distribution while recording so I replied again to show that it's possible for core distribution on recording as well.

And no, I'm not seeking your advice, that would imply that I'm having trouble with something and I need help.
The videos clearly show how to get Logic Pro X working on core distribution on playback and recording, this has nothing to do with workflow.
Old 18th September 2014
  #87
Lives for gear
 

Quote:
Originally Posted by Zoolook View Post
That's the way I have always done it anyway, because using a new track makes it easier to do over-dub recording, edits and roll-back to a prior version if you don't like the 3rd or 4th take. It's just cleaner, if a little old-fashioned. Also, back in the day when you had to freeze all the time, this would be the only way to layer MIDI on frozen tracks.

I think people who have been doing this for a while (for at least 15 years) never complain about not having enough CPU, because in 1997, your workflow was always being crafted around technical limitations. People who started figuring out their workflows in an age of apparently limitless RAM and CPU never seem to have enough. It's one of life's ironies.
Thanks for getting it. heh
Old 18th September 2014
  #88
Lives for gear
 
EastWest Lurker's Avatar
 

Quote:
Originally Posted by ValliSoftware View Post
It was never mentioned to me about workflow. I was showing a fact that Logic Pro X handles core distribution on playback on multitimbral instruments.


I merely replied to show others core distribution with Logic Pro X is possible without any additional $$$oftware.
Unfortunely you replied again regarding core distribution while recording so I replied again to show that it's possible for core distribution on recording as well.

And no, I'm not seeking your advice, that would imply that I'm having trouble with something and I need help.
The videos clearly show how to get Logic Pro X working on core distribution on playback and recording, this has nothing to do with workflow.
Got it, you're fine.
Topic:
Post Reply

Welcome to the Gearslutz 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…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump