MPC 3 -- Let's GO!!

heygorilla wrote:Haha sorry not familiar with the history, but I assume 3.0 doesn't support direct to disk sampling, so it's a flat 5 minutes? Or is this fact just unknown at the moment?

I'd really love a hardware device that could multitrack record long jams in a box with some ok effects and other fun features. Tp7 is just a costly recorder and nothing more, with no screen. 1010 bluebox is the closest but it's a real dead end with no usb audio or usb file access, _and_ no features other than mix and record. Keen to hear any ideas!

+1

cocosushi wrote:

MPC-Tutor wrote:What db909 wants was standard on the old ensoniq esq-1 & eps synth / samplers:
while the seq. was looping one could add /stuff to the seq. til satisfaction set in.

So this is an incremental overwrite? I.e. after looping (say 4 bar seq) it goes back to the start and begins progressively overwriting the existing events, but if you press STOP after 3 bars, the 4th bar is left as per the previous pass?

Ensoniq had a brilliant team of engineers. Despite what most people think, their sequencer( EPS16/ASR-10) was actually pretty good and had a standout feature that is severely underrated - The Song track in song mode. The ability to record an extra track in Song mode OVER a group of sequences.

How this feature is not any modern Sampler (to my knoweldge) is beyond me.

Just did a quick disk streaming test with an MPC Live Mk 1 on an 8 layer piano using 4 sample chains (each chain = 74 slices). I used a busy jazzy piece with lots of trills and fast sections and sustain pedal.

Tested with Samsang evo 850 SSD, internal factory, a fast Integral 512 SD and a 512GB Sandisk USB 3.0 stick.

As expected, absolutely no problems playing pattern from RAM

All other options involved the MPC dropping some notes during the twiddly bits at the start (3.5s in) and end (29s) - this is where the sustain pedal is also being used. Demos here

https://on.soundcloud.com/TYr8eJuzrNxKMNST9

I removed the sustain pedal automation and all notes were triggered without a hitch. Extending some of the bass notes manually also did not impact the streaming. So sustain pedal seems to be the main contributor to any issues, I guess increasing polyphony"

Would imagine a typical 'electronic music' performance would likely be fine.

I also tried streaming with the original 'non sample chain' version (over 270 individual samples, took 30 minutes to set them all to stream). No difference to the sample chain version, so sample chains rule once again

And very good news - if you save this as a 'track' rather than an XPM file, the steaming settings are stored in the track file (which can be manipulated using scripts as it's just a compressed text file, so could automatically batch set hundreds of samples to streaming if need be). Tracks can be distributed just like XPM files, and also contain macros ()but need a separate trackData folder as well).

heygorilla wrote:Haha sorry not familiar with the history, but I assume 3.0 doesn't support direct to disk sampling, so it's a flat 5 minutes? Or is this fact just unknown at the moment?

I'd really love a hardware device that could multitrack record long jams in a box with some ok effects and other fun features. Tp7 is just a costly recorder and nothing more, with no screen. 1010 bluebox is the closest but it's a real dead end with no usb audio or usb file access, _and_ no features other than mix and record. Keen to hear any ideas!

I regret selling my old Akai DPS 12. SCSI2SD installed ran silent and record 6 tracks at the same time. Pretty sure they did a 24. FX were never an issue as have plenty of outboard here.

Sat unused for years. Kept it around for it's time stretch which had a nice vibe.

heygorilla wrote:Oh great thanks. Anyone know the record time limit?

cocosushi wrote:the amount of unused MPC RAM is the limit - for a single audio stereo file it's about 5 mins, iirc
edit: if 3.x would support sampling to disk, the RAM amount wouldn't matter anymore, cause playing back samples / audio from disk is supported finally in 3.x

heygorilla wrote:Haha sorry not familiar with the history, but I assume 3.0 doesn't support direct to disk sampling, so it's a flat 5 minutes? Or is this fact just unknown at the moment?

There is some misinformation here. In 2.x, the sample recording time is 20 minutes. The audio track recording time is limited to 5 minutes in RAM - the rest of the audio is truncated when you save.

In Force 3.x with disk streaming (and MPC 3), the sample time is also 20 minutes. The audio track recording is no longer limited at all due to disk streaming. If you've got a large enough storage attached, you can record for hours and hours and hours.

There is literally NO reason why you would need "sample to disk" on Force or MPC - to "sample" is literally to take a small/short snippit of audio. If you need to record for longer than 20 minutes - that's not sampling - that's just flat out recording. As such, you should be... using audio tracks and recording to disk.

EnochLight wrote:There is literally NO reason why you would need "sample to disk" on Force or MPC - to "sample" is literally to take a small/short snippit of audio. If you need to record for longer than 20 minutes - that's not sampling - that's just flat out recording. As such, you should be... using audio tracks and recording to disk.

Oh really, you think mpc 3.0 will let me record eg 3 stereo tracks... an hour long... simultaneously? I want to archive live jams, in a place where the good bits can be lifted out and used for new fun things.

heygorilla wrote:Oh really, you think mpc 3.0 will let me record eg 3 stereo tracks... an hour long... simultaneously? I want to archive live jams, in a place where the good bits can be lifted out and used for new fun things.

Yes, of course - as Force does this without issue. But my Force also has an SSD and I'm using a Roland MX-1 as the audio interface, so I've got enough inputs. MPC 3 works the same.

MPC-Tutor wrote:Just did a quick disk streaming test with an MPC Live Mk 1 on an 8 layer piano using 4 sample chains (each chain = 74 slices). I used a busy jazzy piece with lots of trills and fast sections and sustain pedal.

Tested with Samsang evo 850 SSD, internal factory, a fast Integral 512 SD and a 512GB Sandisk USB 3.0 stick.

As expected, absolutely no problems playing pattern from RAM

All other options involved the MPC dropping some notes during the twiddly bits at the start (3.5s in) and end (29s) - this is where the sustain pedal is also being used. Demos here

https://soundcloud.com/mpctutor/sets/mpc-disk-streaming

I removed the sustain pedal automation and all notes were triggered without a hitch. Extending some of the bass notes manually also did not impact the streaming. So sustain pedal seems to be the main contributor to any issues, I guess increasing polyphony"

Would imagine a typical 'electronic music' performance would likely be fine.

I also tried streaming with the original 'non sample chain' version (over 270 individual samples, took 30 minutes to set them all to stream). No difference to the sample chain version, so sample chains rule once again

And very good news - if you save this as a 'track' rather than an XPM file, the steaming settings are stored in the track file (which can be manipulated using scripts as it's just a compressed text file, so could automatically batch set hundreds of samples to streaming if need be). Tracks can be distributed just like XPM files, and also contain macros ()but need a separate trackData folder as well).

Nice, thanks.
Did you see how the CPU was acting?
Maybe automation in it self is the problem, as i understand, extended notes with no automation was fine. Is there any automation on non busy parts?
I can not hear the track, it’s not there.

Sorry, here's the correct link:

https://on.soundcloud.com/TYr8eJuzrNxKMNST9

I'll investigate the automation aspect in more detail tomorrow, but sustain only used around the problematic areas, . CPU was fine BTW, didn't seem any different to RAM streamed.

MPC-Tutor wrote:Sorry, here's the correct link:

https://on.soundcloud.com/TYr8eJuzrNxKMNST9

I'll investigate the automation aspect in more detail tomorrow, but sustain only used around the problematic areas, . CPU was fine BTW, didn't seem any different to RAM streamed.

Yes, you can”hear” the notes drop out.
Thank you for looking in to it.
Do you think it is a hardware or software issue?
I do not know the tecnical stuff, but you should be abel to do more than just play back the sample? For exempel automation, right?
You only have four sample tracks, some dude, here on the forum, (i think his name is enoch) had 16+ tracks streaming on the force before there where any problems. I did not find your track too busy, at least not 16+ audio tracks busy. Maybe im wrong, but when you removed the automation it play from disc as from ram, no problem.
Nice to hear that the CPU was not affected more than ram usage.
Again, thanks for taking the time.

EnochLight wrote:In Force 3.x with disk streaming (and MPC 3), the sample time is also 20 minutes.

So how was able record 45 minute cassette tape samples?

In the Force sampler you can even set the max record to 60min. I've never went pass that, but others have.

So did you mean to say, "without" disk-streaming? Or did I miss something?

Someone on the Elektron forum pointed out how unpleasant and unwieldy managing a project with a ton of different sequence variations is. I'm dealing with this at this very moment lol. I don't think the arranger looks good right now, but a solution to this particular problem would be very welcome. Modern music is dynamic in a way that isn't easily implemented on the MPC imo, it'd be great if they somehow squared that circle.

what do you mean by "dynamic"? dynamic arrangement? Isn't that an argument pro clips or sequences?

I think the max number of sequences I used was around 14 with maybe 6 - 10 unused ones,
but usually its less, sometimes only 1 or 3.
I think I onky needed a second grid set once or twice.
I give them names like s1, s2, s1b, s1b 5 up, in other words no really meaningful names.

Its getting confusing sometimes, especially when there are sequences in the project that dont belong to the project at all, but the alternative is to have all in one single timeline.
I certainly prefer to have clips/sequences.

The length of the sequences I use varies from classic 4 bar snippets to complete track length takes.

Sometimes I reuse a project for consistency of sound and out of lazyness, so unused sequences pile up.

B-Wise wrote:

EnochLight wrote:In Force 3.x with disk streaming (and MPC 3), the sample time is also 20 minutes.

So how was able record 45 minute cassette tape samples?

In the Force sampler you can even set the max record to 60min. I've never went pass that, but others have.

So did you mean to say, "without" disk-streaming? Or did I miss something?

Yeah sorry - it was capped at 20 minutes. Seems it's now 60 minutes (though it still samples to RAM). I wasn't aware you could sample past 60 minutes, though. It's been a minute since I fired up my Force, so I'll tool around with it in a bit.

IvoryBlack589 wrote:Someone on the Elektron forum pointed out how unpleasant and unwieldy managing a project with a ton of different sequence variations is. I'm dealing with this at this very moment lol. I don't think the arranger looks good right now, but a solution to this particular problem would be very welcome. Modern music is dynamic in a way that isn't easily implemented on the MPC imo, it'd be great if they somehow squared that circle.

The arranger has no real bearing on a project containing many individual sequences as the arranger is just an alternate way of managing a single sequence. What they need to work on is improving song mode, which is what the 'arranger' should have been IMO, i.e. a place to piece together and 'arrange' all the different sequences into a song, with multi-track view, 'blocks', ability to record audio tracks & automation across a song (without having to 'destructively' convert it to a single sequence), and so on.