MPC2500 Timing ***[[ JJ vs. Akai ]]***

Hey! over in the MPC5000 section they were talking about an interesting topic, the timing of the MPC! here is what they were saying:

From Avene:

"scd, just wondering if I could ask a huge favour? Is there any chance you could test the timing of this thing? I'd really like to know how 'sample accurate' it is. I'll tell you how.

Firstly find a really tiny click sample. Build one yourself if you have to. The shorter the better. Maybe as tiny 20 samples in length or less? It must peak at the very start. Normalise it.

Now assign this click to 16 pads. The same sample on every pad. Then trigger all 16 clicks simultaneously. All 16 pads triggered from the same time location, such as the first beat of a sequence, and all at full level. Maybe even on every beat of a sequence.

Hit play and record the result into a DAW.. Logic or whatever. Now when you play these 16 clicks simultaneously, do you just hear 'one' click as you should, or do you hear a small tone? Zoom right in on the waveform. Do you see just one click or more? If more than one, how many?

Essentially this will be the deciding factor for me as to whether or not the MPC5000 is worth buying. I did this test on the MPC4000 as it's timing always felt pretty sloppy to me. I triggered 16 tiny clicks as described above and recorded them into the computer. When triggering 16 of the same sample it should have just sounded the same as just one of that sample, but it didn't. It sounded more like a tone. When I checked the waveform on the computer I could see every one of those 16 clicks one after the other! Which proved to me that the MPC4000 really was a flawed machine, simply because it's unable to play back two samples in perfect sync. The delay between the first and last clicks a was a number of milliseconds. I later sold it partially because of this.

I also tried this test on the MPC500. Believe it or not it was better, but still not perfect. Instead being able to see all 16 clicks one after the other, I could only see 5 or 6. So the MPC500 can at least play back 3 samples in perfect sync.

With the MPC5000 being the new flagship machine, I would really like to know what it's timing is like. As far as I know nearly all software sequencers and apps such as Reason and Fruity are sample accurate. And you can hear it too, they sound tight. In this day and age, the MPC5000 should at least be able to match the timing of some cheap software apps like these."


From scd:

"Click test:

Hey Avene, good news, the 5k does a pretty good job

First I assigned your sample to every of the 16 pads. Then I recorded a 1 bar sequence on 120 bpm with a click on the first and third beat.

I first recorded a single pad in Logic, it looks like this (maximum zoom factors):
http://www.synthmusic.info/mp3s/single.jpg

Then I recorded the sequence with the simultanious hit of all 16 pads with same duration and velocity:
http://www.synthmusic.info/mp3s/16click.jpg

As you can see there are three samples visible (again with maximum zoom factors in Logic). Every "hit" looks the same in Logic as well.

Proof that the 5000 has indeed a very tight timing."

I'm not to sure if this method will give accurate results due to phasing possibilities but i went ahead and did the test!
I loaded up Akai os 1.23 and recorded audio at 0db digital out, then i loaded up JJ os 5.47 and recorded audio at 0db, the again at -12db. Not only are the samples different but the output volume is different too!
All audio was recorded into Logic with an Mbox2.
Chek out the results i got!



JJ has not only improved the features but the timing as well!
I cant believe how sloppy the Akai timing is! there is F@#$ing 10 samples! at least JJ has got it down to half that! And what is with the output volume's??? anyway, JJ rocks!
But now i have a question of my own, Do you think JJ could improve the timing further, making the MPC tighter in its timing???

Notice the slow attack on the Akai OS too! JJ OS has immediate attack. With Akai OS I would go as far as to say its only good for loops, but not accurate enough for individual drum samples.

B

trybtek wrote:Chek out the results i got!



JJ has not only improved the features but the timing as well!
I cant believe how sloppy the Akai timing is! there is F@#$ing 10 samples! at least JJ has got it down to half that! And what is with the output volume's??? anyway, JJ rocks!
But now i have a question of my own, Do you think JJ could improve the timing further, making the MPC tighter in its timing???

I think the volume differences you see are the layered samples you see. Nothing to worry about. The 2500 is indeed not as tight as the 5000.
I want to do this test tonight with the 4000 as well.

Boele

Also, it would be nice if you do the test again with the sample I used, which is VERY short and fast.

But interesting anyway!

hey scd i used the sample that Avene posted it but recorded it in stereo!

trybtek wrote:hey scd i used the sample that Avene posted it but recorded it in stereo!

Ah yes, that tricked me. Thanks

So JJ just fixed his own code

Some one should try the same thing with the MPC 2000XL.

We could really use the 960 ppq in the next JJOS2 for the 2500 JJ is this possible?

scd wrote:I want to do this test tonight with the 4000 as well.

Boele

No need to, unless you want confirmation. I saved the original audio files from when I tested it. This was recorded in Sony Vegas. The timeline is set to seconds and milliseconds. Notice you can see all 16 samples one after the other? Not only that, the distance between the first and the last is around 7 milliseconds. Way too much in this day and age of sample accurate software music software that will run on a $500 laptop. Plus, the distance between each is kind of random too.

Also, it's the same wav sample I posted up before too. There's just a longer attack in the image below due to the default amp envelope attack of the MPC4000 being set to 1.

From Roland MV8800 demos I've heard, I think it's even worse. Does anyone have an MV they can test this on too? Just for curiosity.

MPC4000 bad timing image -

i don't think i've ever heard 4000 users complain about timing
but that example shows different.

scd wrote:

trybtek wrote:
I want to do this test tonight with the 4000 as well.

Boele

On my 4000, this is what I did and what I am seeing:


Instead of using a DAW, I kept it all internal on the 4000, I used Step Edit to Make sure the 16 notes were hit at the same time, and used the MPC's internal analog In to record the resulting Wav. I don't see the same results that Avenue saw on his 4000 with 16 seperate hits. I'm actually seeing 3 different hits, like you were seeing on the 5000.

http://www.renegaderhythms.com/click_4000.jpg


Cheers,
Dave

renegadebliss wrote:

scd wrote:

trybtek wrote:
I want to do this test tonight with the 4000 as well.

Boele

On my 4000, this is what I did and what I am seeing:


Instead of using a DAW, I kept it all internal on the 4000, I used Step Edit to Make sure the 16 notes were hit at the same time, and used the MPC's internal analog In to record the resulting Wav. I don't see the same results that Avenue saw on his 4000 with 16 seperate hits. I'm actually seeing 3 different hits, like you were seeing on the 5000.

http://www.renegaderhythms.com/click_4000.jpg


Cheers,
Dave

But the test wasn't checking its inputs. Its checking the sloppy timing from its outputs. So to get an "unbiased" test, please do the test on its outputs and report back....

OJ Reem wrote:
But the test wasn't checking its inputs. Its checking the sloppy timing from its outputs. So to get an "unbiased" test, please do the test on its outputs and report back....

You might want to ask somebody with a DAW and a 4000.

But who ever does do that, remember that the particular method can be flawed
if there are issues of timeing on the particular DAW/Computer system. So somebody
with a nice Audio Card such RME and that have done the necessary tweaks that their particular OS needs....


Cheers,
Dave

i did it with my xl on step edit and got the same result on the single as i did with all 16 playing. it registered as a single hit. not multiple.

but really. this is a stupid topic. if you feel like the machine isnt keeping time for you. then sell it and move on. but if you need to blow up the wave form to milliseconds to be able to tell if its out of time. then you need to sell it too because your probably not making music with it anyway.

i did it with my xl on step edit and got the same result on the single as i did with all 16 playing. it registered as a single hit. not multiple.

but really. this is a stupid topic. if you feel like the machine isnt keeping time for you. then sell it and move on. but if you need to blow up the wave form to milliseconds to be able to tell if its out of time. then you need to sell it too because your probably not making music with it anyway.