Help me understand sample levels in Rytm

Hello. Please help me understand what’s going on with the sample levels on ARMk2.
Here I have a pretty normal kick and rumble sound that I made with a kick, some overdrive, delay and reverb. It sounds great on the Rytm and exactly the same when recorded through the main out into Ableton.

Levels are:
BD level: max
Amp level: 100
Compressor: 0% mix, volume 100

The recorded waveform looks like this in Ableton:
original

Once I import the recorded wav back into Rytm, its level is way lower than the original. I’ve set it as a sample, disabled overdrive, delay and reverb while leaving all other settings the same. (edit: SYN is also disabled when playing the sample.) Once I record that again through main out into Ableton, the resulting waveform looks like this:
sampled

Why is the level so much lower? The waveform looks the same as the original in the SMPL window on Rytm but it plays way less loud. It doesn’t matter much if I set the sample level to 100 or 127, it’s still super quiet.

The sound was never clipping as you can see in the pictures. The imported sample doesn’t seem to clip in the SMPL view either. So what happened to the imported sample? Are there some tricks to make recorded samples sound like the original?

I understand that samples are mono on the Rytm, but that doesn’t explain the low level. I have also played the original recording in mono in Ableton and the level is unaffected, so this can’t be some out-of-phase issue either.

Finally, I’m recording into Ableton to show the different waveforms, but this issue persists when resampling the original sound internally on the Rytm: the level of the resulting sample is way lower than the original.

This seems signifiant.

Also, what velocity does the trig have? It will need to be 127 to reply at intrinsic level.

It could do if the phase of one side was slightly off compared with the other. I think the sides are summed in the stereo-mono conversion. Convert it to mono in Ableton before exporting.

Well, the overdrive is already “baked” into the sample when recorded through main out, so not disabling it when playing the sample will cause clipping and distortion. Even with that, the level is considerably lower. I have set trig velocity to 127 when recording the original sound as well as when playing the sample.

1 Like

Fair enough. I’m out of ideas!

Here you can see that it’s not a phase cancellation issue, either.

I summed the original recording to mono, normalized it and exported as 16bit 48kHz wav:
normalized

Again, disabled effects and played back the sample. This time both sample level and amp level turned all the way to 127:
sample-100

As you can see, the sample has lost around 8dB of volume.

How are you moving the samples back to the Rytm? Resampling via Audio In or Overbridge? Or importing with the Transfer utility?

If the former, it must be gain staging (although the Rytm automatically normalises the sample after recording IIRC, so shouldn’t cause sample to be softer).

If Transfer, what bitrate are you exporting from Ableton? Transfer might be converting the sample to the bitrate supported by Rytm, and that might explain it? I don’t know for a fact how Transfer would handle that case, just trying to think of things that might be causing it.

I am rendering the sample as a 16bit 48kHz wav file, which is supported by the device according to the Rytm manual. After that, I import the sample using Transfer. I have tried both with normalization and without, seems to make no difference.

I’ve tested it to some extent and it seems that there’s no issue with the sample being copied over to the rytm, if you download back the transferred sample it’s identical, so there’s no issue with the volume, so it’s part of the signal path that makes it quieter.
it seems that if you set the velocity to 127 for the imported sample, SRC level to 127 and AMP on 100 you get approx 3db reduction, if you start pushing the amp to ~115-116 you get to the same levels as the original export, I’m not sure where exactly the volume reduction happens.
I did something with recording a sample playback rather analog engine and after normalizing the sample and transferring it to rytm it had same levels…
can’t say I’ve reached some conclusions except maybe it has to do with the filter, if I switch from LP to HP I notice same sample gets automatically ~+1db, so it’s probably something to do with the signal flow, I thing it’s best if you normalize the sample before importing it back to get more consistent level…

1 Like

See my post above. Normalized sample, imported through Transfer as 16bit 48kHz wav, played back with sample level 127 and amp level 127, it still plays about 8dB quieter than the original sound.

Track level is maxed out and compressor is (still) at 0% mix, level 100.

1 Like

yeah I didn’t touch the comp or track levels too, and have no such difference in volume, also tried doing 16 and 24 bit exports, all playing pretty much same level…

have you tried creating fresh project to see if still the same? because on my mk2 this is not happening more then 3 db and I’m pretty sure that’s the track level being on 100…

Finally found the culprit. My audio interface had a +6dB boost on the Rytm input. Without the boost, I can get the levels pretty close, but it’s still a bit fiddly.

For instance, recording with the same settings as above but the gain set at 0dB on my audio interface, Ableton will show the peak volume of the sample as -9dB. Normalized and imported to the Rytm, its level is lower than the original if SMPL level is at 100, but higher if set to 127. This makes me think that the levels on the Rytm are not simply attenuation but there’s a boost when you go over 100.

With some trial and error, I can match the level of the re-recorded sample with the original in Ableton. It still doesn’t sound quite the same (as if it lacked some low-end), but I guess that’s to be expected.

4 Likes

I use my ears because its going in the master bus anyway.