Audiobus: Use your music apps together.
What is Audiobus? — Audiobus is an award-winning music app for iPhone and iPad which lets you use your other music apps together. Chain effects on your favourite synth, run the output of apps or Audio Units into an app like GarageBand or Loopy, or select a different audio interface output for each app. Route MIDI between apps — drive a synth from a MIDI sequencer, or add an arpeggiator to your MIDI keyboard — or sync with your external MIDI gear. And control your entire setup from a MIDI controller.
Download on the App StoreAudiobus is the app that makes the rest of your setup better.
Comments
KQ Dixie, use a single operator, fudge the envelope so it ends high.
Drambo, don't use an envelope.
You can generate a file in Audacity and then play it with a file player. This way you can verify the actual output frequency with the spectrum analyzer in Audacity and know exactly what you are playing in the host you are looking at.
Audacity has the ability to generate various tone types and will let you specify frequencies high enough if you set your sample rate for the Audacity project where you want to test. You can then open the generated file from the iOS DAW in Audacity and directly compare the two files.
iVCS is the only prominent Synth I don't own 😅
I have miRack and The Oscillator, and some others from the quoted topic.
Thanks!
But I thought that I need something different, which goes higher than 22.05kHz...
I also own KQ Dixie and Drambo.
Also thanks to @SevenSystems , @uncledave and @NeonSilicon
It may be best to create such a file with Audacity on the Mac, this way I am not already unsure what to do and how to act while creating a test file and can be sure that it is not upsampled.
I like such topics 😅
Will report back!
@tja I also like such purely technical stuff and toying around... and the perfectionist aspect of it. So... interested in your findings (I'm almost sure that none of what I earlier described as "reasonable" will come true 😆)
Hehe, I think that you have a very good sense for such things 😅
I did not yet get the interface, but will create a 96k file at the Mac tomorrow
I ran the tests using Audacity and AUM. I generated a 36kHz sine into a 96k 32-bit float format file in Audacity. I ran the file with the file player input node in AUM. I recorded the output in a 48kHz session and a 96kHz session and brought these files back over to audacity.
The original file and the 96k AUM file have the same spectral signature with a peak at 36k. The 48k AUM session has no audio and no peak in the spectrum. This is what I would expect if AUM down samples the input from the File Player to avoid aliasing in a lower sample rate session when it plays the file.
At least in AUM, it looks like everything is being done as it should be.
Very very interesting!
You stole my job 😅😅😅
I can't remember which iPad you have - is it "stuck" without interface?
In this case, could you try the same within MultiTrack DAW or nanostudio 2, without the interface attached?
Both can still export 96k files!
But without the interface, this would be upsampled and should be empty.
My iPad Pro 12.9 2nd gen. can run more rates than the iPad Air 3.... I will compare those later.
I have the M1 Pro. It does appear to be fixed at 48kHz without an interface attached.
I don't have either Nanostudio 2 or MultiTrack DAW so I can't test those.
Not 100% sure what is being chattered about exactly but for me I think I can tell if something exports as genuine 44 or 48 depending on the pitch of a trackbounce that uses Egoist. If it is 44k the pitch sounds correct, if it is 48k the pitch is screwed up. In the case of NS2 hosting Egoist it seems I need my Roland Duo Capture which has a specific 44/48k switch (set to 44) to get proper bounces otherwise it seems my M1 running NS2 with headphone dongle seems locked at 48. Selecting 44k when exporting from NS2 seems to be downsampled from 48.
Anyway, have not dug super deep but it is super derp.
Auria Pro?
Some random musing:
We have plugins that can oversample internally.
Pro-L2 offers even up to 32 times oversampling, this is 48k * 32 = 1536k...
As AUv3 and also as standalone.
Why can't the host freely choose it's sample rate, as any plugin or other App seemingly can?
This does not seem to be related to the hardware at all, for such plugins.
@NeonSilicon haha, that's some proper detective work. Can't believe anyone else is nuts enough to actually try this
I'm positively surprised by AUM's technical precision 👍️ (but then again it's not that surprising)
That oversampling is internal to the app/AUv3, for its processing. It has to be downsampled to pass along. And that's the problem, that the whole universe has to agree on the sample rate, since you might add another app or plugin. I suspect that allowing change when an interface is connected is a concession made to work with hardware having specific requirements.
Yeah, I've got that one. I'll give it a go.
So, with Auria Pro and no interface connected but setting the project to 96k the resulting mixdown has the 36kHz peak and no aliasing. Looks good to me.
Would also be cool to test that with Cubasis 3. Many people (including me) are using it for mastering. I think this would be a totally cool topic for a video on YT to compare the popular iOS AU hosts in this regard. A collab between you guys and @jakoB_haQ would be awesome.
Oha.
That's a sensational finding!
Real 96k material can be handled and exported fine as 96k material again, without loosing content due to downsampling - without an attached interface!
That is, what I was hoping for, but did not think it to be possible.
I'm even confused that this is really possible. But maybe I missunderstood.
My Mac is booted, I will try something similar with MultiTrack DAW and nanostuduo 2 later.
Also, today I will receive the interface.
I feel what you wrote, but still don't technically understand.
Why does it need to be downsampled to pass along?
If some piece of software has content in a 1536k resolution, why should it not be able to write that out to a file or send it to other software?
Just as a habit to be sure that other software can work with that content?
Dropping this would allow a DAW to run at any sample rate, as long as CPU and OS can carry that.
This would be fantastic.
Just filling that into some brain regions of @SevenSystems
😅🤗
That's the problem.
Cubasis 3 does not allow to export at a different rate, when the iPad is hardware locked.
Basically, in a reasonable world, sampling rate is equivalent to image resolution -- it is completely arbitrary.
Let me give the following example: Imagine you launch Photoshop, and then you're trying to load a 3000x2000 px JPEG. The most normal thing in the world.
Now imagine Photoshop errors out with "Sorry, Photoshop can only open images with your current monitor's resolution (1920x1080)". That would be pretty ridiculous, right?
It's the same with AUv3 hosts and plug-ins. They should work with (and be able to export to) any arbitrary sampling rate.
It's a bit (😁) different with bit depth: I wouldn't expect a DAW to support a bit depth of 19 bits, because this is something that has to do with the underlying hardware architecture as well (it's very difficult to work with numbers that aren't directly supported by the hardware or compiler), so it's reasonable to expect for this to be limited to certain standard depths (16 bit, 24 bit, etc.) -- it's the same with images after all (there's no 11 bit per channel RGB PNGs).
Yes, @SevenSystems
But a DAW that could run at 44.1k, 48k, 88.2k, 96k, 192k or even higher, accepting, handling and exporting such content, without needing an attached interface, would simply be fantastic for all people creating music on iDevices that are hardware locked to 48k ... hint... hint 😅
Yeah, that's what I'm saying. To be honest I'm baffled that this isn't the case. It's not fantastic, it's normal!
XAW
Ideally a DAW should dither / resample the output to match the playback device...
...back in the days of 8-bit colors displays it was still possible to process an image with higher bit-depth we just could not see all the colors.
I think what led to this 'hardware locked sample rate apps chaos' has to do with the early apps that targeting a specific hardware sample-rate and when the hardware sample-rate got changed we ran into issues...
This was mainly done to optimize the performance of an app and retain the quality as early Core Audio libraries frankly sucked at sample-rate conversion.
So in a DAW such as Cubasis we should be able to set the sample-rate at which everything is processed and the output is then dithered/downsampled to match the playback device...
To be honest, XAW can currently only operate at whatever the hardware sampling rate currently is and can't even change it
patience!
Yes.
As it seems, Auria Pro is at least more near to that.
It seems, that it can run a project at up to 96k fine, even on locked hardware.
MultiTrack DAW could be similar.
nanostudio 2 cannot switch the sample rate for a project, it can only change it at mixdown / export.
So this will most probably be upsampled.
I will try this today
I changed Audacity to 96k, 32-bit float, with conversions set to Best Quality and no dither (not sure why it offers dither here, that would depend von wether you want to reduce the bit depth)
I needed to restart Audacity to get a 96k project when using the Generate / Tone menu and created a 10 second project with 4 mono tracks: sinus wave at 40k, 32k, 24k and 16k
I then exported the tracks as single 32-bit float WAV files - just to have them - and as a mixdown of all 4 tracks.
Not sure why it only goes down to -90dB on a 32-bit float ....
And why the 24k spike looks so different than the others.
The mixdown file attached.
Playing this file on an iPad in AUM over File Player and watching with Oscilloscope, this is capped at 20.5k and only shows the 16k line, as expected.
To make a long story (very) short:
the ratio of price, convenience and result is fairly low for any sampling rate above 48hkz.
As mentioned earlier (almost) all critical sections in digital processing today use internal oversampling where appropriate. No need to extend this to the whole processing chain.
Doubling the sample rate divides available time slots by two, increasing CPU demands by an even higher rate.
Input channels need a significant quality increase to take advantage of higher resolution.
In other words: a 50cent interface preamp would only reveal it‘s weaknesses more clearly in a proper 96khz digital chain.
For plain synthesized content this doesn‘t matter much, but play back in realtime would increase processing demands (see above).
That‘s why some devices are locked and some DAWs don‘t offer each and every processing option. They are oriented at the demands of general user expectations.
If you want to experiment with higher resolution, nothing stops you from doing that...
Except that 90% of current quality synth/fx engines already did use higher internal sample rates.
And it is really hard to tell the difference in a final mix.
I had that „problem“ ages ago and for my humble ears it just wasn‘t worth the effort.
Someone with the same system did the switch to 96khz throughout and sometimes publishes a mix or two.
None of these mixes ever triggered a „wow, that‘s great...“ with me, they are nice mixes... not an another dimension of sound.
But I second any attempts to make your very own experience. Conditions can be tricky sometimes, as this thread shows.
Yes.
And that means, that you cut out aliasing and other stuff between the plugins - which is what you want when creating music, but what stops you from analysing sound and effects of aliasing.
Why should an iDevice not allow sound creation and analysis as is possible on a desktop?
This has absolutely nothing to do with the creation of music.
This is just a topic around sound processing on iDevices.