Discuss the various methods you use in music production, from compressor settings to equipment type.
By sifis Sun Apr 10, 2011 12:04 pm
I think i get you now illgreen.. It's all about experimentation... you tell us to find our eq settings.. find our copmpressor settings and make our sound .. and maybe you ll get some cancellations on your signal... that's ok .. it;s acceptable..

But i have one more question that i realize...

Remember told you about sample numbers?? (st:0 end:470,431)

so, i pitch shift the "compressed" drum break just about 10 cents to get a little bit different frequencies... but now the samples are (st:0 end:490,591)

So the sample is bigger in time...

Anyone knows something about that??? Is it the algorithm you choose ?? or what??
It's not a time stretch !!! it's a pitch shift !! am i right??
User avatar
By nogginj Sun Apr 10, 2011 3:06 pm
No, phasing happens when two signals are of the same frequency therefore canceling out vital sounds that would've otherwise enhanced the sound, has nothing to do with timing. If you was recording guitar in one room and drums next door at the sametime and the drummer is banging so hard that his sound leaks into the guitar mic then phase will occur dumbing down the drums because the waveforms are the same.

Sorry but this is not entirely true. To anyone that doesn't know, or thinks they know, take this as a moment to educate yourself on what exactly is happening.
Two waves, of the exact same frequency, played at the EXACT same time, will sum, thereby becoming louder..
Two waves, of the exact same frequency, played at the different times, could cancel each other out at times, but it depends solely on the timing and how offset the waves are from each other.
This is physics, there's no way around it.

It actually IS all about timing, because as soon as they start playing at different times, the samples of the wave might interact in such as a way as to cancel each other out. If you have two waves, same frequency, and play them at the same time, they will double in loudness. If you play them such that the peaks of one wave coincide with the valleys of the other wave, then they will cancel.

It is dependent on lots of things, but timing is a HUGE part of it.

Also your example about drums leaking into the guitar mic is not entirely accurate, and can't even be summed up that simply. Depending on the mics used, mic placement, air in the room, etc etc, it could do any number of things, but definitely not simply 'dumbing down the drums'. If anything, it will make the drums seem fuller as they are coming in on more tracks. In this case, it comes once again all down to timing. If the recordings of the guitar mic and the drums are played at exactly the same time, it will likely make the drums louder (by a tiny tiny amount). If however there is a small delay (delay in the sample range...sub-milliseconds), it could produce phasing. But yea, in that case, it's actually all about timing. The 'dumbing down' you hear is because the drum sound is hitting the drum mics and the guitar mics at different times, and the timing offset causes phase cancellation.

You can do this experiment yourself in your daw.

Set up a test oscillator ringing at 220 hz on track1.
Set up another test oscillator ringing at 220 hz on track2.
Make your DAW output mono (this is very important).

Play both, and you will notice you still here the wave perfectly fine, maybe a little loud, maybe even distorted.

If you adjust the phase of the oscillator, start doing it, you will notice the main output tone start to get quieter.
If you cannot adjust the phase of the oscillator, you can use a sample delay....but it has to be a true sample delay where you can adjust the delay rate by 1 sample at a time.

If you have two oscillators at 220hz, playing at the exact same time, and you delay one by 110 samples, they should completely cancel out.

Of course, all this depends on the concept of sample accuracy...and if your daw, or gear (like the mpc 5000), cannot do sample accurate operations, then it's not going to work.
User avatar
By nogginj Sun Apr 10, 2011 5:44 pm
^^ that might work....there have been similar tricks used w/ mixing desks and daws...basically using 'null' fx so the delay between both channels is the same. sometimes just inserting an effect on a channel is enough to add the same amount of latency as an effect being used on another channel.

nice post.
By dustymaestro Sun Apr 10, 2011 5:46 pm
Sample accuracy/synchronization
Not as much a specification as an ability. Since independent digital audio devices are each run by their own crystal oscillator, and no two crystals are exactly the same, the sample rate will be slightly different. This will cause the devices to drift apart over time. The effects of this can vary. If one digital device is used to monitor another digital device, this will cause dropouts in the audio, as one device will be producing more or less data than the other per unit time. If two independent devices record at the same time, one will lag the other more and more over time. This effect can be circumvented with a wordclock synchronization.


When you shift forward or backward you have to find the sweet spot between loudness, flanging, flamming. It's used to make a sound fuller all the way to a flam.The MPC 5000 has 960 ppqn. The timing is spot on. You guys are mixing up the definition of sample accuracy.
User avatar
By nogginj Sun Apr 10, 2011 5:57 pm
960 ppqn

'ppqn' is going to change depending on tempo. 'Samples' are absolute and do not change lengths dependent on tempo.
Sample accuracy is finer than ppqn accuracy, again, depending on bpm.

But PPQN doesn't have anything to do with samples, as that is a function of the sequencer ;]

And I do agree, shifting forwards or backwards is the key to getting this to work right...can the mpc5000 shift that precisely?
Also, someone above said that maybe shifting wont even work as they claim the MPC triggers things slightly off, and randomly at that (at the sample level...not according to what your sequencer says).
I can see how that would make sense, sense the MPC is not designed for sample accurate processing of multiple tracks.
By dustymaestro Sun Apr 10, 2011 6:39 pm
PPQN (Pulses Per Quarter Note, sometimes Parts Per Quarter Note)
The timing resolution of a MIDI sequencer. PPQN indicates the number of divisions a quarter note has been split into, and directly relates to the ability of the sequencer to accurately represent fine rhythmic variations in a performance, or to recreate the "feel" of a performance. Older sequencers were capable of 96 PPQN (sometimes even less), which often resulted in a stiff "quantized" feel to the music (even if it hadn't actually been quantized). Current versions can reach 768 PPQN or even higher resolutions, which is more than adequate for most musical applications. Note that the resolution of the sequencer is especially important at slower tempos. If your sequencer is limited to a lower resolution, one trick is the double the tempo of the song, then perform the parts in half time. This effectively results in a doubling of resolution.

960ppqn at double tempo is overkill.
By dtaa pla muk Sun Apr 10, 2011 6:45 pm
a waveform sample accurately placed on top of itself is in phase and the volume will double.

1 + 1 = 2.

noggin wrote:To anyone that doesn't know, or thinks they know, take this as a moment to educate yourself on what exactly is happening.
Two waves, of the exact same frequency, played at the EXACT same time, will sum, thereby becoming louder..
Two waves, of the exact same frequency, played at the different times, could cancel each other out at times, but it depends solely on the timing and how offset the waves are from each other.
This is physics, there's no way around it.


yup.
and if you copy a sample, invert its phase, and then play it atop the other, 1 - 1 = 0 and the sound will null.
By dustymaestro Sun Apr 10, 2011 6:45 pm
I think the sample accuracy statement applies when there are multiple devices to be synced. If your just using a mpc with multiple tracks, the tracks dont drift. You might get some jitter with the processor but it should be sample accurate.
User avatar
By nogginj Sun Apr 10, 2011 7:21 pm
dustymaestro wrote:I think the sample accuracy statement applies when there are multiple devices to be synced. If your just using a mpc with multiple tracks, the tracks dont drift. You might get some jitter with the processor but it should be sample accurate.

Sample accuracy can be within a machine...it does not only apply to different pieces of gear.
And again, it's completely and wholly unrelated to PPQN...PPQN is tied to the sequencer and samples are audio...PPQN changes dependent on tempo, samples do not.

I agree with you, 960ppqn is very precise, but it's just wholly unrelated to the idea of sample accuracy within a device. Where talking audio and PPQN is sequence information.
I also agree it SHOULD be sample accurate, but is it in reality? I get the impression that the numark MPCs are aimed more at the consumer vs professional level, so this scientific precision is something that could be totally absent from the machine. Hell, are any mpcs really that accurate?

Having said all that, I don't have a 5000 to test on, and testing itself would be difficult for most people unless they have a digital connection between their MPC and recording software, be it a timepiece or a digi out. We're talking very very small chunks of time here, so precision is very important.

I can't make a claim as to whether or not the 5000 is sample accurate or not, but I can say that it could be a reason the parallel compression is not working properly (effected signal is delayed by small amount), and I can help dispel some of the myths about phase cancellation that are going around.
Last edited by nogginj on Sun Apr 10, 2011 7:38 pm, edited 1 time in total.
User avatar
By Ill-Green Sun Apr 10, 2011 7:29 pm
Yeah, you're right noggin. My bad mp3, when you was talking about timing, I was thinking about 4/4 measures and such, but the timing in this case is that both sounds must be perfectly aligned with each other to get that boost.
By dustymaestro Sun Apr 10, 2011 7:57 pm
I think everybody here knows what phasing sounds like. I have used phasing as an efffect also as an effect on a turntables. I thought it was because it was spot on. Its hard to tell a millisecond drift, so it seems spot on. Most people dont really care because its very hard to tell. The scientific formula that you guys are mentioning changes when you start doing more math 1+1+1+1=2 stereo tracks that have phase problems. PPQN and sample accuracy are two different things that can cause similar effect. Im not saying there the same thing. Im pointing out that sample accuracy is getting confused with other things. Dont go from scientific formulas to an impression. As far as the definition of sample accuracy I can only go by whats written, not my own meaning.

http://www.eqmag.com/article/whaddya-me ... ug-03/2102
User avatar
By Fung45 Mon Apr 11, 2011 2:27 am
A sound played at the exact same time as a duplicate of itself with phase inverted will cancel itself out.
A sound played at almost the exact same time as a duplicate of itself will produce a slightly offset echo effect, I think some call this phasing effect or phaser, perhaps that's causes the confusion.
User avatar
By otobot Mon Apr 11, 2011 5:38 am
Y'all confusing the MPC with DAW software. In any DAW the MIDI implementation is sample accurate.. in any hardware sampler it is NOT because of the physical limitations of the MIDI protocol. That's why you have unpredictable phasing when you layer let's say two snares exactly at the same time in any hardware sampler.
The seperate samples are triggered by two different MIDI events. MIDI is a serial protocol and two events at the same time will never go out like 100 % exactly at the same time. First one of the events, then the other one.

That's not a "problem" of only the MPC 5000... all hardware samplers will act like that when you layer two sounds triggered by a MIDI sequencer (it's own seq or an external one).
User avatar
By krush777 Mon Apr 11, 2011 6:08 am
Now that was a debate from which everyone learned something and progress was made! This was a great thread to read!