Long-time 'Fence-Sitter' about to Pull the Trigger...
So I've watched all of StevePAL's excellent videos, listened to dendy and other's demos, read all the lively chat on the Audiobus forum and elsewhere...but still don't have NS2 yet.
Some of the most compelling features that I continually read about are NS2's resource optimization, low overhead, and stability. Also it's powerful and intuitive MIDI sequencer, Obsidian, effects, quick workflow, etc.
Experimenting in the 'modular world' of apeMatrix/AB3/AUM is great for generating ideas, but I'd like to go back to finishing songs in the traditional manner - in a single app, bringing all of those apeMatrix experiments as samples into a standalone app. I do it with BM3, it's been my favorite since release, and I've spent a considerable amount of time creating instrument sample banks in BM3, but now I'm thinking it was a waste of time since Obsidian should cover most of those sounds, (and I could make space by deleting ~6GB of my banks from BM3). But I wouldn't want to delete all of my BM3 drum kits. And I would miss audio tracks, disk streaming, audio transient detection, slicing, AU FX automation, mod wheel and sustain pedal support...and mostly BM3's excellent drum sampler - BM3's instrument per pad approach is excellent for drum kits and slicing long samples.
So how do I make use of both NS2 and BM3 's best features? Along with apeMatrix experiments? I'm on a 2013 Air 1 and Mini 2 - I decided not to upgrade to a newer iPad if NS2 is as efficient as I've read. I'll leave all the CPU-hungry AUs for apeMatrix for audio recording, and export the audio into BM3 or NS2's Slate. But NS2's fast workflow, efficiency, stability, and Obsidian are what I am missing. Thanks!
Comments
you can sync BM3 and NS2 together using Link.. Very soon NS2 will get Audiobus update, so you will be able record NS2 into BM3 audio tracks...
I personally like more Obsidian+sample start automation replacement for audio tracks than Slate pads - it removes need rewind to beginning of long note to hear audio + as bonus you get simple "timestretch" which works surprisingly good, especially with drum loops
Yes it is. But - think about what you can afford on latest Air/Mini :-))
Main reason for upgrading - on latest devices your needs of resampling CPU heavy AUs would be significantly lower...
It depends on complexity of your music, but there is chance you would not need resample to audio for saving CPU at all :-) With strategy, let's say for 70% of stuff use Obsidian+build in FXs (they are actually pretty good, not much people are talking about them, mostly just about Obsidian, but bundled FX's are in my opinion comparable to most of AUfxs - great overdrive/distortion, reverb, delay, chorus,phaser, filter,..) and just rest AUs for things not covered by NS, this can be reality..
All Reslice issues will be solved in upcoming update, Matt did his part of job and VirSin did their, so this is the way. If you like using pads, you can control Reslice with Slate pads ;-)
Back to your question about what possible workflow choose - there are various workflows you can use. I think @richardyot mentioned that in last 2 tracks he did most of song in NS2 and then he exported it to auria and added vocal recording and applied some auria's FX for final mix/master .. As i mentioned above - with upcoming AB support, possibilities for using more apps together with NS will be lot bigger than now.
Thanks @dendy You know I was going to ask a million questions, right?
I'll keep it at 2 questions for now.
Okay, so AB support will come for the few IAAs I still use, LINK is currently supported, Obsidian using sample-start automation - I've read your posts on how to do it.
I missed your social media post about the Mini 5's performance. That's the iPad I planned on getting initially, but it was nice to test the other 3 models over the past 2 weeks. Still waiting for my store to get the Mini 5 back in stock.
1st question: Any complaints or regrets so far on your Mini 5 purchase?
2nd question: With the same test, in your estimate, how would the Air 1 perform? (I noticed you had multiple filters and FX, including 100x long reverbs. That's impressive!)
With the Air 1, I could probably cover 90% of my synth-needs with Obsidian, and just use the heavy AUs in apeMatrix using the built-in LFOs, and print that to audio for export. I do automate synths and effects like a mad man - so NS2's internal FX can be automated, but not AU FX, until a later update.
VirSyn Reslice: never got it, have Egoist (sluggish UI), and the latest BM3 update with audio transient detection and better slicing makes them a bit unnecessary for my needs.
I think between NS2, BM3, and apeMatrix, I should be covered, even on the old Air 1, but the weakest part of the Air 1/Mini 2 is the graphics/Metal performance. It benchmarks at ~500, whereas even the 2018 base iPad is ~13000. That's where the sluggish UIs become much more apparent. I can get around the 3x-4x lower CPU performance, but the sluggish UIs are a buzz-kill.
Thanks again dendy!
As a small clarification, the mod wheel on an external controller works fine in NS2, you can assign it to the performance control of your choice (NS2 performance knobs or the XY pad). The only thing it currently won’t do is directly affect the mod wheel of an AUv3 instrument. You have to be able to assign it to a performance control.
Ah, I see, thanks for the clarification @Stiksi !
Ask how much you wan't, probably it's obvious i'm enjoying answering questions :-)) I'm "teacher" type of guy
I'm very much excited :-) Now when i grab my old iPad Air i feel like i'm holding some stone age artifact
20 instances smooth, no drops, 63% cpu load .. interesting is when i get to 71% of cpu load by adding two more instances - it is still smooth (even when i'm scrolling in mixer), but at beginning of song loop it aways makes one drop .. but just on beginning, rest of loop is playing still smooth ..
It's lot better on Mini5
Not totally smooth (you still feel, when moving sliders for example, they are a tiny bit sluggish - obviously UI of egoist is coded in let's say it politically correctly, not in best way)
Although for slicing i found Reslice LOT easier, i like it's simplicity .. and it can export all slices as zipped package - so if you use "open in nanostudio", NS automatically unzips it and place into "Inbox" folder of your library - so you can use Reslice standalone for just fast slicing and then save it to NS and load slices to Slate.
Thanks, once again, for all of your help. 🖖
I mainly just use BM3 for building kits for export to NS2. I look forward to NS2 audio tracks so I can just jam along and export larger chunks from BM3 and hopefully be able to edit them well enough in NS2.
Thanks @Audiogus Are you exporting individual kits or stereo audio tracks? I was planning on exporting Obsidian audio and apeMatrix audio into BM3, but if your method is better, then I'm all ears.
Hopefully NS2 audio tracks aren't as buggy as BM3's-recording to audio tracks in BM3 usually doesn't work for me, but sampling instead to a pad works every time. Importing audio into an audio track works fine.
@ocelot just in case you didn't purchased yet, there is 50% holiday sale
I didn't expect to see a Friday sale today, but I suppose it is a Good day for a sale.
As someone with very little income, my opinion is still that none of the apps on iOS have a monetary barrier unless someone is in a situation where they struggle to have food to eat. Some low end apps maybe time wasters for some, depending upon workflow and desired activity, but NS2 is not a time waster for anyone who wants to create music and actually get projects completed. If the goal is just to improvise with wild soundscapes with no intent to create a finished piece, then a set up like AUM with lots of synths and FX will suffice. It all depends upon what you are trying to do. And there is no need to choose one workflow exclusively. That would be like me having only one tool in my workshop. I may not use the Drill everyday, but when I need it - there it is! Happiness ensues. Mostly....
It's fun to use several tools to get the job done. As far as soundscape and ambient music, I used to do it a lot with hardware synths and a PC sequencer, but the truth is I never go back and listen to those tracks.
I became much more productive once I got a hardware sequencer, and these days maybe slightly less so with iOS since there's just so many apps on iOS for little money (thanks for the heads up @dendy).
Got the iPad Mini 5. For some reason I can't set most apps' sample rate to 44.1kHz when using the headphone jack. It's locked at 48kHz like the 2018 Pro 11.
Real-World Usage Tests Updated 4/29/2019:
apeMatrix@512 (Mini 5 @48kHz vs iPad 6 #2 @44.1kHz vs Mini 2 @44.1kHz):
Watching the CPU for 1 minute (CPU Min/Max):
Preset StepBudDixieRozRuis: Mini 5: 18-22%; iPad 6: 10-14%; Mini 2: 11-16%.
Preset StepBudDixieiDensityKosmoRozRuis: Mini 5: 24-36%; iPad 6: 17-24%; Mini 2: 17-27%.
Preset FugueCalityAparilloParticlesKronecker: Mini 5: 13-41%; iPad 6: 17-21%; Mini 2: 29-35%.
Preset CyclopRuisNoirAxonDedalus Jam 01: Mini 5: 11-15%; iPad 6: 12-18%; Mini 2: 41-46%.
Preset ThesysDixieStepBudElsaFuguePrimer: Mini 5: 15-46%; iPad 6: 17-23%; Mini 2: 26-40%.
Preset StepBudDixieAparilloAxonNoir: Mini 5: 13-16%; iPad 6: 17-21%; Mini 2: 44-60%.
Preset StepBudDixieNoirRozTrouble: Mini 5: 13-40%; iPad 6: 16-21%; Mini 2: 22-30%.
Preset RozX2AparilloX2SampleriPulsaretNoir: Mini 5: 25-44%; iPad 6: 25-29%; Mini 2: 76-85%.
Preset Modular MIDI Mayhem: Mini 5: 11-17%; iPad 6: 9-21%; Mini 2: 27-53%.
AUM@1024 (Mini 5 @48kHz vs iPad 6 #2 @44.1kHz vs Mini 2 @44.1kHz):
Watching the CPU for 1 minute (CPU Min/Max):
Preset Fugue_iceWorks+Dixie+Ruis: Mini 5: 29-44%; iPad 6: 25-33%; Mini 2: 36-44%.
Preset Roz_Heaviest Synths 3: Mini 5: 19-30%; iPad 6: Deleted Preset/Couldn't Test; Mini 2: 57-109%.
Preset Roz_Noise+Heaviest Synths: Mini 5: 27-40%; iPad 6: 35-55%; Mini 2: 88-118%.
Preset Roz_Sugar Bytes 4: Mini 5: Deleted Preset/Couldn't Test; iPad 6: 25-31%; Mini 2: 49-69%.
BM3@1024 (Mini 5 @48kHz vs iPad 6 #2 @44.1kHz vs Mini 2 @44.1kHz):
Watching the CPU for 1 minute (CPU Min/Max):
48kHz Samples Only+Bark Filter AU on Main Out: Mini 5: 18-23%; iPad 6: 10-25%; Mini 2: 14-22%. Without Bark Filter: Mini 5: 2-5%; iPad 6: 1-6%; Mini 2: 2-3%.
44.1kHz Samples Only: Mini 5: 2-4%; iPad 6: 1-5%; Mini 2: 1-3%.
Stress Tests Updated 4/29/2019:
Audiobus 3 stress tests while watching its CPU meter for 4 minutes, noting only peaks, not averages (except for CPU Min/Max for the last iPad 6 #2 tests). The Pro 11's CPU averages lower than the rest, with the Mini 5 slightly higher. Pro 11 and Mini 5 have more frequent CPU spikes. Opening synth GUIs doesn't noticeably impact the Mini 5's CPU usage, except for Quanta, where the CPU peak usage rose from 36% to 76% @256@48kHz. This is strange because when testing the 2018 iPad 6 at 64@44.1kHz, opening Quanta was fine (only opening Cyclop or Aparillo resulted in crackling), so Quanta's GUI appears to have issues with the A12 iPads.
Audiobus 3 Preset: 'CPU 110': 5 Rozetas driving 4 of my heaviest AUs (Cyclop+Aparillo+Quanta+Synthmaster One) + Ruismaker.
2013 Mini 2: 64@44.1kHz, CPU peaks @231%. Non-stop crackling; GUIs slow + unusable.
2013 Mini 2: 128@44.1kHz, CPU peaks @188%. Non-stop crackling; GUIs slow + unusable.
2013 Mini 2: 256@44.1kHz, CPU peaks @135%. Non-stop crackling; GUIs slow + unusable.
2013 Mini 2: 512@44.1kHz, CPU peaks @116%. Frequent crackling; GUIs slow + unusable.
2013 Mini 2: 1024@44.1kHz, CPU peaks @107%. Frequent crackling; GUIs slow + unusable.
2018 Pro 11: 64@48kHz, CPU peaks @56%. No crackling; GUIs smooth + useable.
2018 Pro 11: 1024@48kHz, CPU peaks @32%. No crackling; GUIs smooth + useable.
2018 iPad 6: 64@44.1kHz, CPU peaks @122%. Occasional crackling, especially with Cyclop+Aparillo GUI open, other synths OK; GUIs smooth + useable.
2018 iPad 6: 256@44.1kHz, CPU peaks @55%. No crackling; GUIs smooth + useable.
2018 iPad 6: 1024@44.1kHz, CPU peaks @38%. No crackling; GUIs smooth + useable.
2019 Mini 5: 64@48kHz, CPU peaks @52%. Occasional crackling, but only with Quanta GUI open, other synths OK; GUIs smooth + useable.
2019 Mini 5: 128@48kHz, CPU peaks @38%. No crackling; GUIs smooth + useable.
2019 Mini 5: 256@48kHz, CPU peaks @36% (76% with Quanta GUI open). No crackling; GUIs smooth + useable.
2019 Mini 5: 512@48kHz, CPU peaks @29%. No crackling; GUIs smooth + useable.
2019 Mini 5: 1024@48kHz, CPU peaks @26%. No crackling; GUIs smooth + useable.
2018 iPad 6 #2 64@44.1kHz, CPU Min/Max @43-114%. Occasional crackling, especially with Cyclop+Aparillo GUI open, but no CPU spikes, other synths OK; GUIs smooth + useable.
2018 iPad 6 #2 128@44.1kHz, CPU Min/Max @29-79%. No crackling; GUIs smooth + useable.
2018 iPad 6 #2 256@44.1kHz, CPU Min/Max @28-52%. No crackling; GUIs smooth + useable.
2018 iPad 6 #2 512@44.1kHz, CPU Min/Max @32-46%. No crackling; GUIs smooth + useable.
2018 iPad 6 #2 1024@44.1kHz, CPU Min/Max @21-39%. No crackling; GUIs smooth + useable.
Stress Tests Updated 4/30/2019:
2019 Mini 5:
apeMatrix 128@48kHz, 10x iPulsarets all using the default patch 'Factory-Sky Instr (Def)', all playing a 7-note chord; apeMatrix CPU meter 49-56% over 2 minutes. No crackling; UIs smooth. That particular preset uses the Trainlet as its grain source.
Using a Sample as the grain source, with all 10x iPulsarets using the patch 'ape_periodic-1', all playing a 7-note chord; apeMatrix CPU meter 41-58% over 2 minutes. No crackling; UIs smooth.
2018 iPad 6:
apeMatrix 128@44.1kHz, 10x iPulsarets all using the default patch 'Factory-Sky Instr (Def)', all playing a 7-note chord; apeMatrix CPU meter 63-84% over 2 minutes. No crackling; UIs smooth. That particular preset uses the Trainlet as its grain source.
Using a Sample as the grain source, with all 10x iPulsarets using the patch 'ape_periodic-1', all playing a 7-note chord; apeMatrix CPU meter 49-53% over 2 minutes. No crackling; UIs smooth.
2013 Mini 2:
apeMatrix 128@44.1kHz, 10x iPulsarets all using the default patch 'Factory-Sky Instr (Def)', all playing a 7-note chord; apeMatrix CPU meter 81-187% over 2 minutes. Non-stop crackling; UIs slow. That particular preset uses the Trainlet as its grain source.
Using a Sample as the grain source, with all 10x iPulsarets using the patch 'ape_periodic-1', all playing a 7-note chord; apeMatrix CPU meter 89-166% over 2 minutes. Non-stop crackling; UIs slow.
Dendy's Test: NS2 Obsidian playing 4 notes poly, 3x Supersaw Oscillators, 3x Filters (LP, HP, Formant), 3x FX (phase, delay, long plate reverb). 512 sample buffer. Sample Rate 44.1kHz.
2019 Mini 5: 100 Obsidian instances=70% CPU. Smooth UI.
2013 Air 1: 20 Obsidian instances=63% CPU. Smooth UI. Adding 2 more Obsidians=71% CPU, Smooth UI, but 1 dropout only at song loop start.
@ocelot i replied to your issue with 48/44 khz in AB thread, check it there.. it's really strange.. maybe try that hard reset..
oh additional question - any chance tomtrybit with different headphones ??
@Cinebient that's very cool idea, will try it... eventually combined with battery endurance test.. just give me some time, will deliver results
From BM3 I am just exporting banks of 32 samples for use as NS2 Slate kits. BM3 is essentially my audio editor for making samples that require pitching, stretching, adding fades, misc processing etc. Who knows, maybe NS2 will have a ‘copy/paste to slate pad’ right on the timeline but I am not really expecting that sort of thing and imagine BM3 will still be used even when audio tracks show up.
Yah I totally have the same issues.
Thanks Audiogus. Would be nice to have a full-fledged multi-track audio editor on iOS, without having to use a DAW.
Got NS2 a day before the big sale, no worries. Still getting familiar with its workflow. Obsidian is amazing, just wish we could expand the sample mapping window. Tempo and time sig tracks are significant.
dendy, now just half of my music apps on the Mini 5 can't be set to 44.1kHz with headphones attached. An owner of the Air 3 has the exact same 'issue'. I believe it's an app issue, not a HW issue. Also the fact that we can change the sample rate in AUM w/ or w/o headphones attached leads me to believe that it's not worth stressing about. The Mini 5 is a little beast. Funny thing is Apple recently fixed their typo, it does have 2nd Gen TouchID, not 1st Gen.
I did all those Audiobus 3 stress tests while just watching its CPU meter for 4 minutes and noting only peaks, not averages. The Pro 11 does average lower than the rest. Pro 1 1 and Mini 5 have more frequent CPU spikes. The CPU boost over the Air 1/Mini 2 is significant, but the biggest jump in performance is UI responsiveness when opening and editing AUs, mixing, etc.
Thanks everyone.
this is crazy.. i cannot understand how same app can work at 44 for me but not for you on same device ... most weird thing i ever experienced at iOS
I'm pretty curious what result you get with NS.. because there will be significant change in realtime sample rate handling - current public version works internally always on 44khz and then just at the end it upsamples result stereo audio stream to 48khz if device needs that..
New version will natively work on device sample rate - i think this change was needed because of Audiobus support and also because a lot of AU plugins implements incorrectly samplerate handling (they ignore rate set by host and use always device sample rate)
New version will also show current used sample rate in settings, so you will be able quickly check what is going on when you disconnect/reconnect headphones
you are able to set 44 in AUM with headphones disconnected ? Me not, it stays set at 48.
It's definitely strange. Mine is a USA model, no cellular. It's not unusual during electronics manufacturing for different parts to be used depending on the global region where the product is sold. Some countries have bans on certain components due to the microwave frequencies they emit, military, etc. So it could be that, or...
It certainly was the case with an old Alienware laptop, where the Nvidia chipset and motherboard components emitted certain radio frequencies that disrupted WiFi, LCD, touch bar, and many other functions. And the type of WiFi card installed also was a factor, making it even worse with certain cards. The laptop was discontinued within 6 months, replaced within a new model using an Intel chipset and mobo, but it took lawyers, RF engineers, and the FCC for Alienware to do anything about it. Of course, this anomaly with certain Mini 5 and Air 3 iPads is not as extreme as the Alienware case.
I may go to the Apple Store to test their display models, if they'll allow me to download a free music app where I can set the sample rate. But it could be that the apps are wrong. I'm burned out testing this stuff, but a final confirmation would be nice.
With headphones attached and nothing else running:
Apps that are locked at 48kHz: All apeSoft+Amazing Noises, BM3, AudioShare, Cubasis,
Apps that can be set to 44.1kHz: Groove Rider, Poison-202, SunVox, Virtual ANS, zMors Modular, KQ MiniSynth, Quanta,
With AUM, occasionally I can it to 44.1kHz w/o headphones attached, but it's rare.
I believe tackling this sample rate anomoly is going to be a lot tougher than expected, though most users don't notice it. To be honest, for users it's not worth stressing about, except for the out-of-tune synths and exporting issues. Thanks goodness some developers are working on it.