Any tips for a more efficient use of the ER-301's CPU/memory?

ER-301
#1

Are there any tips for more efficient use of memory?

I haven’t had the ER-301 for long, but last night I was a bit concerned that the memory was already at 74%, though I had quite a few things going on: 3 or 4 mixers in stereo outputs 1-2, each mixer with a sample player, a filter and a fixed delay… plus an eq after the mixers… Trying to replicate one of my common uses for the Octatrack. That was without recording.

Not sure if I shouldn’t worry too much about the memory, leave my worries until the ER-301 starts to crash on me… After all, I never checked how much memory the Octatrack was using overall, but quite often I had 5 to 7 channels with samples and two effects each, leaving the last channel for further filtering and EQ, and I’m concerned that if I’m going to replicate that, I’m going to start having problems. What surprises me, because the ER-301 has much more powerful specs than the Octatrack.

Anyway, just want to know from more advanced users if you had any tips about the most efficient use of the ER-301’s memory.

Thanks…

1 Like
#2

May I confirm that you are talking about memory and not CPU usage? Also if you are indeed talking about memory then are your referring to sample memory or kernel memory?

#3

So sorry… I mean CPU usage, I think. Sorry, I’m still quite new to the ER-301. Maybe I need to read more about difference between sample memory and the kernel memory, but I assume those two have to do with the sample pool and the sample file size. But I also assume that all that also has an effect on the CPU. Is that correct?

#4

are you running at 96 or 48k?

#5

Ah, yes. I suppose that’s one of the first things I should know, if I’m worried about the CPU usage. My ER-301 came with version 0.1.6pre-patch-13, Release Candidate (96kHz). As far as I know, there is not an option to switch to 48k. Not that I want to switch to 48k, but nice to know I could change if the CPU is struggling.

In what other specific cases would you recommend switching to 48k?

#6

One of the ‘tips’ that I have come across so far would be: delete anything that you’re not using.

It might seem like an obvious thing, but in my case I had been using a 4 sample players template, and sometimes I only needed three. Deleting the one that is not in use frees quite a bit of CPU, no matter if it didn’t have a sample assigned to it or if the entire mixer was bypassed.

#7

Honestly, unless you’re DRASTICALLY pitching shifting audio files, 48k will offer almost 2x the performance and still sounds excellent.

The firmware currently is released in a 48k and 96k version, with a future plan to offer a software switch. Currently, you can change firmware as often as you need. They’re both the same other than the rate.

Also, If you’re not using stereo samples or FX, work in mono and you’l free up some CPU juice as well :slight_smile:

2 Likes
#8

and keep in mind, bypass is not disabling DSP … the unit is still running in the background but not passing audio. If you don’t want it impacting on performance, you currently have to remove it from the chain.

#9

The SAMPLE PLAYER unit does everything at sample-rate, with the highest possible quality, no short cuts. There are no “control rate” signals in its path. Speed changes (linear fm), pitch changes (exp fm), triggers and so on are all processed on each and every sample for the highest fidelity. That is why a SAMPLE PLAYER takes up 5% (or more) of CPU.

Simple playback of a sample that is already in the native rate (48kHz or 96kHz depending on the firmware you are using) without variable speed would take less than 0.5% of CPU. However, such simple playback has not been exposed to you yet. For example, the Piano16 instrument sampler that @anon83620728 and @Joe created could be redone with a much simpler playback unit that would take almost no CPU.

3 Likes
#10

Of all the words you type, @odevices, I believe my favorite one is “yet.” :tada:

3 Likes
#11

Thanks for the explanation. Looking forward to read more about the way the ER-301 works, in the manual, once it’s ready. This forum is also great, but sometimes it takes me quite a bit to find the information I want.

I’ve to say, I’m very impressed with the ER-301. The sound quality is excellent. I haven’t tried the 48kHz version yet. I was just a bit freak out the other night when I realised I was already using around 75% of the CPU, but recording, even in stereo, seems to take no CPU at all. Looking forward to the other sample players, like reading straight from the card, for accompaniment, which I assume would take barely any CPU at all.

My apologies if I keep comparing it to the Octatrack, since I sold my OT to buy the ER-301… but I don’t mean it in any bad way. I know the ER-301 is far more powerful and it allows me to do some crazier stuff. And if I remember correctly, the sample rate in the OT was only 44kHz. Though, to be fair to Elektron, the OT is quite old now. But even then, I don’t see them coming out with a machine as powerful as the ER-301.

#12

I know what you mean, but the search functionality is excellent and well worth exploring!

I come from a similar background with the Octatrack and had very similar thoughts along these lines, I still have my Octatrack though :wink:

I think there are some very important differences between the two!

Not least of which is the innate ability of the ER-301 to route and process signals in a totally flexible way. This is possible to some degree, but quite tricky and comparatively very limited on the Octatrack; I believe the performance differences such as they are, lie in making this flexibility available.

Quality is another important factor, the ER-301 is way more powerful in this regard, I mean, not just a bit, the difference is HUGE! You’e always going to eat up more CPU cycles when working at this level.

I still think the Octatrack is an amazing bit of hardware and I love it very much, ironically my current feelings are that it’s not a case of one or the other but both as it makes an excellent companion in many ways! I’ve no intention of selling my Octatrack despite this being an initial consideration when buying the ER-301!

#13

I sold my Octatrack to fund one of the early ER-301 prototypes (I think the 6th one).

7 Likes
#14

Do you miss it?

#15

Not really. The more I used it the more I wanted something that that would let me stay inside the modular. I’m not talking about satisfying an OCD to have everything in the same physical box. It was jarring to move from Octatrack to modular and back while not really having access to all the modular signals within the Octatrack and vice versa. I got the same “teleportation sickness” when using the DAW with modular via Expert Sleepers.

4 Likes
#16

Thanks for sharing this, I must admit it is precisely these reasons I wanted to make the move too!

I always knew the ES approach was too ‘jarring’ as you put it, it’s too much of a leap for me too, I know others get along with it very nicely and of course that’s great and totally viable especially these days.

However, I don’t find the Octatrack/Modular shift that problematic and am happy to accept that there’s a control signal flow from the Octatrack to the Modular (by way of Yarns or the Shuttle Control) and an audio signal flow back to the Octatrack from the modular via way of the audio ins. Yes, they are separate worlds with different operation structures, but that’s ok!

Actually, I like having these two worlds to hop between, one is particularly good for one set of tasks and the other good for another, for example, I really get along with the Octatrack sequencer style and in particular p-locks controlling many parameters in a completely independent way which while not impossible can get very messy very quickly and requires a large investment in utility modules in the modular hardware world.

So, the most exciting thing about the ER-301 to me is that it opens up this signal to destination dialogue without the heavy hardware requirements - virtual attenuverters and offsets on every parameter and the ability to add as many mixers, with whatever you like behind them to complement these, is one of the most powerful advancements in modular synthesis signal processing ever I think!

Being able to take multiple signals, and feed each of them to multiple places in the ER-301 and have completely independent treatment of those signals for each destination, is to my mind, p-lock functionality in the modular world on steroids … and without the mess!

It’s incredibly inspiring!!

** writing this out just made me realise that the Sample Looper could be used in the same way as the Octatrack records p-locks - @NeilParfitt has already demonstrated the methodology, but I hadn’t quite made this conceptual connection until now - amazing!!

:cat:

#17

You know a bunch of people, me included, will be hounding you to do a follow-up on this, right? :slight_smile:

#18

Ha! Me and my big mouth :stuck_out_tongue:

I think @NeilParfitt is the man for the job with one of his excellent videos!

The principle is the same as the envelope recording and the synced Looper triggering combined :wink:

I will think on and see if I can build something that demonstrates this in a meaningful way, I don’t think it’s a candidate for a custom unit though as it is essentially just using the Looper to record your manual modulations over a synced/fixed amount of time.

2 Likes
#19

Yes, just having a laugh :slight_smile: So nice reading about the great ideas people here come up with, will have to chip in with at least something when I get my unit!

1 Like
#20

I’m always up to doing more vids! :sunglasses:

3 Likes