"Truncate" changes sample BPM reading - is there a fix?

[ADEMIKS]

"Truncate" changes sample BPM - is there a fix?

Hello. I just recently got my SP-404SX and spent a lot of time assigning samples to pads and inputting correct markers. After all of this progress I had all of my samples with perfect start/end times set. I figured that I would truncate all of these samples in order to save space and allow for temporary marker changes in the future. To my surprise, after truncating, the BPM is totally wrong when read off of the screen [CTRL2].
[example: Original markers set the sample BPM to 78BPM / after truncating, the unmarked sample reads 112BPM even though the same piece is being played]

This is problematic for me because I have no way reliable way of matching this truncated sample's tempo with that of a pattern's BPM because 78BPM is now disguised as 112BPM and so on...

Is this just the reality of the SP-404SX? Do you recommend that I never truncate samples if I record them in and adjust the markers as opposed to copying from a PC? Is there a way to adjust a samples base tempo rather than just doubling it?

[hseiken]

Sample pitch match is purely comparing two numbers, not anything inherent in the audio itself and no, I don't think it recalcs when you trim. It still just does the calculations on the fly. It should, if there were a way to control it, change tempo instantly if you could change the length of the sample while it's playing (i.e. taking off bits of the end while the middle plays).

Basically, it's just doing some math:

Length of sample is x length and y bpm. If sample is z bars, then at y bpm it should be stretched to match that length. (i.e. think in terms of seconds and how long a beat is in seconds...then you start to figure out how it's doing it's math) If you change x, then you're going to affect y because it doesn't know that you trimmed the sample intending to calculate the space you just cut from the sample.

The workaround = resample the timestretched/tempo'd version then cut that shit with tempo match off. CONS = you gotta redo your chops if you decide to change tempo of your track later and it affects how your sample fits in.

[ADEMIKS]

Thanks for your reply. Just to make it more clear, when I import a wav from my PC to the SP-404sx via SD Card transfer (example: 2 bars / 8 beats perfect loop / 78BPM) the sp-404sx will recognize this sample as 78BPM. If I record in this same section of audio through line in with automatic start point and specify 78BPM before recording and hit rec at the end point while recording, I will get the same end result. However, after recording through line in in this fashion, "mark" will be lit and there will be a little extra time ignored after the endpoint. Once I truncate this sample the wav becomes identical in sample length as the file I imported via SD card (2 bars / 8 beats / same length in seconds)--yet, the sp-404sx will scramble the read BPM to something far off of 78BPM. (The sample will still playback the same)

Further, if I copy and paste this new truncated sample off of the SD Card and onto my pc I could then proceed to import it back onto the sp-404sx via SD Card and the sp-404sx will correctly identify 78BPM. Try this yourself and you will see that the math in the sp-404sx during truncate is flawed.

So moving forward, how is the truncate feature to be used if one cares at all about BPM matching / time stretching in the future? If I'm working with a true 78BPM sample that the sp thinks is 112bpm, setting the stretch to 108bpm in reality sounds more like 74bpm since the reference is wrong.

[hseiken]

Just to be clear, you're making the assumption that because the numbers don't match then the beats aren't going to match?

If this is the case, I highly suggest you ignore the screen and just use your ears. The numbers the 404 comes up with are meant merely as 'attempts to help' but they're not some super high tech feature. The 404 itself is a fairly lo-tek machine and anyone would agree that it's sequencer is quite garbage and this goes for anything on the machine pretending to be BPM based, in my opinion.

Again, if you're having issues that when you truncate your shit gets retime-stretched, then stop truncating and make duplicates (i.e. resample) of when the bpm was matched right, then take the bpm stretch off and just treat it like a normal sample. If you're going for perfection and you're using the 404, that's immediately a clash of ideals and the 404 win against you every time.

I hope this helps a little and if not, I'm quite sure another person will float through here and give your problem another pass with different ideas.


EDIT: And if you're truncating because you're 'trying to save space', get over that shit real quick. 404SX has up to 32GB sample memory. In practice what that means is you can sample a complete album to store on every pad at full quality and you're not going to run out of room. Again, stop truncating if that's the issue, and figure out a way to duplicate stuff without truncating. All I got.

[ADEMIKS]

Sure it is low tech, but the problem is that the truncate process not only scrambles the BPM displayed but it affects how time stretch will work. After truncated and after an incorrect BPM is displayed, I can no longer scroll the time stretch knob to the right and sync the BPM with the pattern BPM. Why?

Here is a scenario. The true sample BPM is 78 and it is showing up as 112bpm after truncation. The pattern BPM is set to 92pm. I go to my truncated sample and set the time stretch to sync to the pattern BPM of 92. Now, the sp-404sx tries to slow down the playback of the truncated sample to reach "92bpm" instead of speeding it up to reach 92 from 78. The result is now a sample that is super slow and noisey. So in total, it looks and sounds like the BPM shown is what the sp-404sx works with.

I'm not concerned with running out of space, but I would like the ability to reset my markers to change a long 4 bar loop to a 1 bar loop on the fly and than return to a perfectly cut 4 bar sample later on if I wish by deleting the marker. That's where I saw truncation coming in... Also, if I keep markers I may accidently hit a wrong button and alter a start/end point and have the sample cuts change without knowing right away.

So... my only work around seems to be to resample a marked sample and physically hit the record button at the right time to end the recording and deal with the fact that it's not as precise as truncate (especially when catching the end of a very high BPM sample or waiting for a very long sample take to reach the end)?

[ADEMIKS]

Sorry if this post seems strange...
This BPM problem is just driving me insane! It just seems like I'm able to record in samples and have it show the correct BPM, and than at a later time the BPM shown is scrambled. Some samples will resample and show the correct BPM each time I try, while others consistently show the wrong BPM...

That being said, some samples, for example a .wav at 83BPM (real tempo as verified by Ableton Live, FL Studio, tap metronome), will show up as 62BPM on the SP-404SX. On the other hand, a different sample might actually match up on the SP-404SX...

[urje]

Might be a bit late if you've taken a hammer to it already but my high mind figured it. All you goota do is keep jabbing the 'Mark' button whilst you have the pad on 'Loop' with the 'Gate' off. Make sure you hit it in time to what you think the BPM is then eventually the SP will correct itself. Might take 20 attempts but it shouldn't take longer than a few seconds in reality.

If you're close to sure two pads are the same length they should read a similar BPM. You can always check by running them out at the same time with no loop.

Dunno if any of that makes sense and I failed to read any above posts too clearly. Hope it helps.

[ADEMIKS]

Yes, I have become faster at chopping marker regions. However, the problem was that once samples are marked and than truncated afterwards, the truncated version will likely show as a different tempo than the original marked sample (a 16 beat sample when accurately truncated to 4 beats for example). The problem this has created for me is that the truncate feature becomes useless for tempo tracks (not drum shots etc) because now you cannot accurately time stretch them to match the bpm of other loaded samples. The bpm difference can be significant, sometimes over 30bpm. So now I resort to copying the full sample and only using markers to shorten them and be extremely careful not to accidentally change mark points.