Timing Correct - MPC Swing shootout: Do all MPCs have the s

konc3pt wrote:

Askia Shaheed wrote:There is something to laugh about.

and it's the following statement

Askia Shaheed wrote: I always felt the MPC 2000's timing was more solid

which mpc are you finding to be out of rhythm

I never said or implied that any MPC is "out of rhythm."

Lampdog wrote:

peterpiper wrote:nothing to lol about. closing your eyes will focus you brain to the ear and will result in other start/end points. This will give your beat a different groove.

peace

I don't know why they lol at this when this is a major part of the truth.

When I learned on my asr10 years ago, timing and swing was way different than when I learned the 2kxl. It's because the non-precise chopping on the older non graphical machines. When you have a waveform graph your chops are more precise when "looking" at the zero crosspoints. Instead of just laughing at this info P has offered, close your eyes and chop THEN look down at where you cursor is and how far away you are from the zero crosspoint. Now if you keep it like that AND then go ahead and add swing on the same tracks with the non-precise chops you'll get crazy swings. When different people chop that same exact way, everyone will get crazy swings that will never be precise or identical to the next man and it's all "magical".

"Looking" at the step edits may reveal the exact same swing on different machines.
But taking into account what I said above will possibly explain the "magic" that P and is speaking on.

Askia Shaheed wrote:Sample editing with or without graphic waveform editing has nothing to do with this topic.

You are mistaken.

I am doing a controlled test.

I first pointed out the fact that certain swing settings (50, 51, and 52%) do the same thing on the 2000XL. I am trying to determine if this is the case with other MPCs. Sample editing has nothing to do with this. You gentleman are introducing variables, guessings, skipping to the end to come up with your own conclusions, etc.

So its not about music but about the math behind the machines?
In this case all MPCs with 96 ppq will do the same on each % swing because the math is the same.
% of a specific number of ticks is always the same. The only thing that could affect a different shift in ticks would be a different interpretation of the decimal place. But why should the software do this even when you got 2 %values on each tick (in 16th swing mode)?

Just to make it clear again. Like you said: We're not talking about any musical thing here. This has nothing to do with groove or vibe or feeling of a track. You're interested if the is a difference in calculation of MPCs. So its about if one MPC interpretation of 52% swing is 1 tick and another MPC interprets it as 2 ticks. There is nothing between the ticks.

peace

Askia Shaheed wrote:

konc3pt wrote:

Askia Shaheed wrote:There is something to laugh about.

and it's the following statement

Askia Shaheed wrote: I always felt the MPC 2000's timing was more solid

which mpc are you finding to be out of rhythm

I never said or implied that any MPC is "out of rhythm."

in that case what do you mean by saying that you always felt the mpc 2000 timining more solid ? how do you percieve it ?

I am looking forward to seeing the test result. However, it is going to be not easy comparison test to conduct.

I heard that even MPC-2000XL would go slightly out of sync, when it has to deal (trigger) a lot of sequence data:

http://www.mpc2000xl.com/hiccups.htm

Logically, MPC-2000xl should handle "dense" sequence data better than MPC-3000 and MPC-60, since it has faster processor (But we would think which machine sound better is another story.)

I'm interested how Askia will conduct the testing.

JVC wrote:I am looking forward to seeing the test result. However, it is going to be not easy comparison test to conduct.

I heard that even MPC-2000XL would go slightly out of sync, when it has to deal (trigger) a lot of sequence data:

http://www.mpc2000xl.com/hiccups.htm

Logically, MPC-2000xl should handle "dense" sequence data better than MPC-3000 and MPC-60, since it has faster processor (But we would think which machine sound better is another story.)

I'm interested how Askia will conduct the testing.

^^This is off topic.

I thought this topic was about music, groove, feeling, vibe, too. But as far as I understand he just want to know/test if a MIDI note will shift the exact number of ticks on all MPCs if the % setting is the same. This is not about the actual sample and how or when it gets played. Its about the MPCs interpretation of the swing parameter or in other words: How does the MPC calculate how many ticks it will shift a MIDI NOTE ON message based on a % number.

I wonder why the engineers at AKAI used that 50%-75%. It's senseless if you have a fuction like the EARLIER/LATER

peace

peterpiper wrote:^^This is off topic.

I thought this topic was about music, groove, feeling, vibe, too. But as far as I understand he just want to know/test if a MIDI note will shift the exact number of ticks on all MPCs if the % setting is the same. This is not about the actual sample and how or when it gets played. Its about the MPCs interpretation of the swing parameter or in other words: How does the MPC calculate how many ticks it will shift a MIDI NOTE ON message based on a % number.

I wonder why the engineers at AKAI used that 50%-75%. It's senseless if you have a fuction like the EARLIER/LATER
peace

"Shift timing EARLIER/LATER" is different parameter than swing.

I didn't realize that this is such a touchy topic. This is under MPC comparison topics, so I don't think this is off the topic. Trying to find out (or trying to confirm) about it is great, the test result can be a "myth buster."

I personally don't believe MPC-3000 / 60's "magic swing" (they are great machines, nevertheless), but I do believe the fact that they do not have waveform display would contribute "different feeling" between MPC-3000 and MPC-2000.

to bust any myths on this ur prolly gonna have to actually measure the time between multiple hit's.
Maybe the step edit screans show equal values. While the actual signal output is just slightley "Off".. And perhaps newer models just work more precise, while both do the same maths?
If this is at all possible... that is...

JVC wrote:

peterpiper wrote:^^This is off topic.

I thought this topic was about music, groove, feeling, vibe, too. But as far as I understand he just want to know/test if a MIDI note will shift the exact number of ticks on all MPCs if the % setting is the same. This is not about the actual sample and how or when it gets played. Its about the MPCs interpretation of the swing parameter or in other words: How does the MPC calculate how many ticks it will shift a MIDI NOTE ON message based on a % number.

I wonder why the engineers at AKAI used that 50%-75%. It's senseless if you have a fuction like the EARLIER/LATER
peace

"Shift timing EARLIER/LATER" is different parameter than swing.

This difference is small. Technically the functions do the same. Shifting a MIDI message.
The differences are:
1. SWING only affects Notes on "offbeat" positions like 001.01.24 and 001.01.72 while EARLIER/LATER will shift every note no matter the position.
2. SWING can only shift a note later while EARLIER/LATER can shift in both directions.



MPC2000 manual
Swing %:
This appears only when Note Value: is set to 1/8 or 1/16. Shift
the note events of the even beats at the ratio set here to swing the
rhythm.

Shift Timing:
This shifts the note event back or forth the number of ticks set in the
amount: field below.
EARLIER shifts back
LATER shifts forth

amount:
This sets the volume of the shift timing. The maximum value varies
depending on the Note Value settings. For example, if the Note
Value is set at 1/16, the maximum amount is 12 which corresponds
to half of the time correction value (24 ticks).

Askia wrote:Here are my final results when Note Value is set to 1/16. When using the following settings, my drums were quantized to the following intervals:

SWING%: BARS.BEATS.TICKS:

50-52%: 001.01.00, 001.01.24, 001.01.48, 001.01.72
53-54%: 001.01.00, 001.01.25, 001.01.48, 001.01.73
55-56%: 001.01.00, 001.01.26, 001.01.48, 001.01.74
57-58%: 001.01.00, 001.01.27, 001.01.48, 001.01.75
59-60%: 001.01.00, 001.01.28, 001.01.48, 001.01.76
61-62%: 001.01.00, 001.01.29, 001.01.48, 001.01.77
63-64%: 001.01.00, 001.01.30, 001.01.48, 001.01.78
65-66%: 001.01.00, 001.01.31, 001.01.48, 001.01.79
67-68%: 001.01.00, 001.01.32, 001.01.48, 001.01.80
69-70%: 001.01.00, 001.01.33, 001.01.48, 001.01.81
71-72%: 001.01.00, 001.01.34, 001.01.48, 001.01.82
73-74%: 001.01.00, 001.01.35, 001.01.48, 001.01.83
75%: 001.01.00, 001.01.36, 001.01.48, 001.01.84

Swing 53% or 54% shifts the note on positon 001.01.24 and 001.01.72 one tick later.

LATER amount 1 will do the same (ok, like I said it will affect every note in the set timing range)


To me the "magic" has nothing to do with this absolute settings. There IS a shifting of samples but it is not in the scale of midi ticks but way shorter in time. For example: The time from one tick to the next on a 90bpm beat is 0,0069 seconds so lets say 7ms. The "magical" shifting as I see it is in the range of a few samples to something like 1ms and it's not adjustable by the user cause its (again this is how I understand it. I don't have enough insight on this to give 100% answers) caused by the internal clock, RAM, CPU, sample readout from RAM process. I doubt that there is an MPC that can play back measureable 100% exact in timing. Add the fact that your ear won't be able to cut exact on the first transient and/or the display probable won't show you really what the sample look like. All this factors affect how the beat will come out.

There are some strange things goin on in those old machines fo sure.
Examples?

1. MPC2000 end point is not shown correct. This can be tested easily with a drumbreak Set the startpoint to one hit and move the endpoint to the point where you even can't HEAR the next hit. Look at the display. Why does it show you that the endpoint of hit 1 is not the startpoint of hit 2??

2. s950: endpoint is not "absolute" but depends on the pitch you play a sample. Lets say you have a sample like "snare-hihat" and want to trim it to snare only. You set the endpoint to the poition where you even can't hear the hihat. you play it on the original pitch of the sample. Now if you've done this, play the sample at a lower pitch. Lets say -12 semitones. ooops, there is the hihat again

3. again s950: you've trim a nice snare sample. You've put it on a keygroup and edit the amplitude ADSR. The attack parameter should affect the way the sample comes in. Value 0 should sample with an instant 100% level. The higher the value of attack -> the softer the attack of the sample. But it happen to me a lot of times that the snares attack sounds harder if the value is set between something like 2 and 9. It should get softer.

If someone have insight and answers on these strange behaviours, please tell.

peace

peterpiper wrote:If someone have insight and answers on these strange behaviours, please tell.

while thats all very interesting, wouldn't that be off topic? ;]

here's a 1 second 44.1k impulse sample anyone can use for this test.

http://s000.tinyupload.com/index.php?fi ... 6851957640

(ive always been curious how the ad/da convertors on any given sampler handle unique signals like this. for example an xl does not reproduce DC easily.)

edit: here's a noise sample if that impulse gets garbled
http://s000.tinyupload.com/index.php?fi ... 4923075252

peterpiper wrote:

JVC wrote:"Shift timing EARLIER/LATER" is different parameter than swing.

This difference is small. Technically the functions do the same. Shifting a MIDI message.

The differences are:
1. SWING only affects Notes on "offbeat" positions like 001.01.24 and 001.01.72 while EARLIER/LATER will shift every note no matter the position.
2. SWING can only shift a note later while EARLIER/LATER can shift in both directions.

I wouldn't call it "small difference." I think you've just proved that they are totally different.

peterpiper wrote:There are some strange things goin on in those old machines fo sure.
Examples?

1. MPC2000 end point is not shown correct. This can be tested easily with a drumbreak Set the startpoint to one hit and move the endpoint to the point where you even can't HEAR the next hit. Look at the display. Why does it show you that the endpoint of hit 1 is not the startpoint of hit 2??

I've noticed that, and I do include the "next hit", on 2000XL because of the reason. I think MPC-2000 would play samples with very short fade out, even when fadeout is set zero.
I try not to look at the screen, when I am editing samples on XL, but I have to admit that it is useful.

Actually, it is already mentioned:

http://www.mpc2000xl.com/short.htm

tapedeck wrote:

peterpiper wrote:If someone have insight and answers on these strange behaviours, please tell.

while thats all very interesting, wouldn't that be off topic? ;]

here's a 1 second 44.1k impulse sample anyone can use for this test.

http://s000.tinyupload.com/index.php?fi ... 6851957640

(ive always been curious how the ad/da convertors on any given sampler handle unique like this. for example an xl does not reproduce DC easily.)

In that case I would call it thread hijacking

DC?? dirct current?? in that case I guess its get filtered before it hits the ad. A DC offset correction on a sample editor is just a highpass filter with a very low cutoff. The DC is handled like a peak of a very low frequency.

peace

peterpiper wrote:...in that case what do you mean by saying that you always felt the mpc 2000 timining more solid ? how do you percieve it ?

peterpiper wrote:So its not about music but about the math behind the machines?
In this case all MPCs with 96 ppq will do the same on each % swing because the math is the same.

The MPC 2000 was my first MPC. At the time, I always set quantize to 1/16ths, applied swing anywhere from 52% to 66%, and shift notes + or - a few ticks. When I applied the same settings to the MPC 4000, I just didn't get the same results. Based upon this, my perception is that they don't quantize/swing the same way. The purpose of this thread is to determine if this is fact or fiction.

We don't know if all MPCs with 96 ppqn perform the same until it is verified by someone posting the results like I did previously in this thread.

peterpiper wrote:DC?? dirct current?? in that case I guess its get filtered before it hits the ad. A DC offset correction on a sample editor is just a highpass filter with a very low cutoff. The DC is handled like a peak of a very low frequency.

yea thats exactly what im referring to. i generated some full amplitude dc samples and transferred them to the mpc via cf card. there must be a dc filter cause the resulting output of the mpc was definitely not what i put into it.

about the test, it might even be more interesting to look at an audio recording of the resulting swing ... while all machines might report the same tick values, i wonder how close they actually are.

didn't somebody do a really thorough test like this on the 5000?