Input Levels: Hotter on the Right

That's not dithering Jahrome, it's just truncating the 24 bit data and the pitch change is from the MPC not realising it's 48k and reclocking the sample.

Dithering is a process where an algorithm is applied to a data stream which reduces the bit depth in a psychoaccousticly pleasing way, reducing inharmonic tuncation errors.

morgan wrote:
That's not dithering Jahrome, it's just truncating the 24 bit data and the pitch change is from the MPC not realising it's 48k and reclocking the sample.

Dithering is a process where an algorithm is applied to a data stream which reduces the bit depth in a psychoaccousticly pleasing way, reducing inharmonic tuncation errors.

Dithering is a mathematical process where a random noise is
added to the least significant bit of a digital word. With very low level
signals, the quantization error becomes correlated to the signal level. This
creates a measurable amount of distortion. By adding dither, the correlation
between the signal level and the quantization error is cancelled, allowing
the digital system to encode amplitudes smaller than the least significant
bit. If you change the word size as a signal passes from one digital system
to another, being able to add dither allows you to maintain a high quality
signal.


I wouldn't say that the reason for the pitch change is due to the MPC 'not realising' that the incoming signal is 48K. The MPC 2500 has done an excellent job of recording the digital signal recorded at a higher bit and sample rate....'dithering' the 48Khz/24 bit audio to 44.1khz (at whatever bit rate). I get the same results when I set a Nuendo Project at 48K and import a 44.1khz sample. The pitch of the sample will be altered.

morgan wrote:That's not how digital audio works at all. Binary data is the same as decimal in that the most significant figure comes first and each digit afterwards is of lesser significance. If the Fantom really did pad 16 bits up to 24 by adding 0s at the start it would distort the signal really badly, especially seeing as the first digit indicates whether the signal is going positive or negative. You're right about how 24bits gets truncated to 16 just by chopping the last 8 digits off though.

Actually, that's exactly how binary anything works as far as I know. Wether or not it's big endian or little endian is basically irrelevant to my point but let me be clear about what i meant. When I said the fantom x will "pad 0's to the top of the outbound signal.." I meant that all bits above 16 are set to 0 in a 24 bit representation of that number. Padding the higher order bits will not affect the data at all because at the end of the day it's the same number. That is to say, anything above 2^16 / 65,536 / 1000000000000000 and up to 2^24 / 16,777,216 / 100000000000000000000000. So 100111000100 would be 00000000100111000100 - which is the exact same number: 2500. Using this example: If I sent 00000000100111000100 to the MPC this is what I would get: 000000001001 which is 9. Yeah, there are 16 bits there but only 4 that are representing a number.

Writing this all out and thinking about it again has made me realize that I have been using the wrong terms in previous posts. I'm not talking about "bit rate" I'm talking about "bit depth" but I hope most people got the point.

If I'm smokin' crack here lemme know.[/i]

Jahrome: Nice cut and paste, straight out of the 02R manual. Doesn't mean you understand it though.

The pitch is altered because your digital audio player, be it nuendo or the mpc, just plays one sample after the other, not knowing at what rate they were recorded at. When you play something at the wrong sample rate you get 44 100 samples playing per second instead of 48 000, which pitches the sample down. No amount of dithering will change that.

Anyway, this is too off topic, lets get back to the analogue input levels.

Why don't your post the link to the site you cop'ed that definition of dither from

The MPC did not dither you data, man. If it had correctly dithered it you'd end up with no audible difference in the data. F*cking the pitch is not adding noise it's doing exactly what morgan just said...

d'oh, morgan posted right before me!

off topic indeed. the point remains however, you can't feed any signal other than 16/44.1 to the MPC or your data will not be right. period.

as for the analog level issue i guess i'm suprised that the signal would ever vary that much and if it's happening in other gear then perhaps there is an anomaly in your setup, Jahrome? i don't see that with all my gear but at the same time i haven't tried it all out yet...

Yo zs, the thing with padding out the data with zeros is that they always go on the right hand side of the number, where the data is less significant. That way when the MPC truncates it you don't lose any data.

It's like having a number 0.4592 (simple four bit decimal example) when you want to pad it out to eight bits you go 0.45920000 rather than 0.00004592

Ya dig?

so did we figure out if this problem actually exist. And if it does is it hardware based or OS (software). Cause mine should be in at alto soon!!!!!!!!!!

i'll do some more tests between my fx8 and the mpc and see if i can't narrow this down any more. maybe i'll just pm you since mrnymns is buggin' out and jahrome is playing professor with his mixer manual. in any case, the point about not being able to sample from my 828 remains true since there is no padding taking place: data would be lost.

mrnymms, i'm thinkin' it's a problem but i'm gonna do some more tests to try and narrow it down before i make that claim for sure. can you try and do some tests with your setup and see what you results you get? maybe try a few configurations?

I don't have my 2500 yet not at least for a week. My 2kXL samples fine; I mean according to the probably not very accurate meter of my mpc it might be off a sliver, but not drastic. IS the supposed hot channel a lot or just a little?

Yall are cracking me up.

You can feed any signal you want into the MPC. Even if it was recorded at a higher sample rate..bit rate..bit depth..little bit Yes, the 48Khz sample was transposed a semitone when I sampled it via the digital input..but after I went into the program edit menu and adjusted the pitch....guess what?? The sample played back perfectly. No apparent loss in data or quality of sound

And back to the analog test....I seriously doubt it is software based. If you are having problems, it may be specific to your set-up (gear, cables, power supply, etc).

zs941, you sound like a mad scientist. Let me repeat myself...there is no apparent lost in data or quality of sound when you send a 48Khz sample into the MPC 2500. It is very UNTRUE saying that you can't. Also, why hide your findings in a PM?? I know you are looking hard for a 'bug' to name after yourself But I have found a major one (which I have not posted yet because no one has used their MPC long/extensive enough to find it ). So keep digging

As far as the 'anomaly' in my set-up??

Anomaly-something different, abnormal, peculiar, or not easily classified

My set-up works perfectly. I am not the one having problems I can't solve. If I am the only one not having problems..that I guess you are right...its an anomaly specific to my set-up.

what i meant was the person having the "problem" was it the actual 1/4 plugs that are faulty on the 2500? I just want to know cause I don't want to spend money or time or the phone tag to get mine sent back in to fix faulty connections on the machine

jahrome wrote:
You can feed any signal you want into the MPC. Even if it was recorded at a higher sample rate..bit rate..bit depth..little bit Yes, the 48Khz sample was transposed a semitone when I sampled it via the digital input..but after I went into the program edit menu and adjusted the pitch....guess what?? The sample played back perfectly. No apparent loss in data or quality of sound

think what you want, professor but you're wrong. you may not be able to hear any difference but you have lost data. it is not the same data you sent it because bits have been removed.

jahrome wrote:
And back to the analog test....I seriously doubt it is software based. If you are having problems, it may be specific to your set-up (gear, cables, power supply, etc).

i doubt it's a software based thing, too. never said it was. i still haven't said it's a "bug" or a "defect" simply that it's a problem that could use more lookin' into. could well be a power issue and that's what i'm gonna play with next but it's strange.

jahrome wrote:
zs941, you sound like a mad scientist. Let me repeat myself...there is no apparent lost in data or quality of sound when you send a 48Khz sample into the MPC 2500.

apparent and actual are two different things. you've actually lost data but apparently you can't tell. which is fine. it's all about the ears, right, so do what your ears tell you is good. seriously. but don't kid yourself or others into thinking that pitch-shifting a f*cked up sample either dithered or replaced the data you lost.

jahrome wrote:

It is very UNTRUE saying that you can't. Also, why hide your findings in a PM?? I know you are looking hard for a 'bug' to name after yourself But I have found a major one (which I have not posted yet because no one has used their MPC long/extensive enough to find it ). So keep digging

i never said you can't record different bit depths or sample rates into the MPC (i actually said you can). the MPC will record what you send it, no doubt, but what you get is not a digital copy as you may expect. what is very UNTRUE is to claim that the MPC correctly records that data, that is has a bit converter, that it has any form of dithering, or that the data is there and you just need to pitch-shift it. You clearly understand very, very little about digital audio so do everybody a favor and hit the books. of course i'll post my finding in the forum but discussing this with you is a waste of time because you keep claiming to know all sorts of stuff that you clearly don't. just admit you don't have a clue, ask some quesions, do some reading, or shut up. For somebody who gets so uppidy when somebody makes false claims about the MPC it's odd that so far in this thread you've made at least four.

jahrome wrote:
As far as the 'anomaly' in my set-up??

Anomaly-something different, abnormal, peculiar, or not easily classified

My set-up works perfectly. I am not the one having problems I can't solve. If I am the only one not having problems..that I guess you are right...its an anomaly specific to my set-up.

that's pretty funny. Touché....

i've yet to find a problem i can't solve either but i'm not arrogant enough to think that there are no problems with my setup.

anybody with a 2500 wanna try and reproduce the tests that jahrome and i have done? still curious to see if there are others with different results.

mrnymms wrote:what i meant was the person having the "problem" was it the actual 1/4 plugs that are faulty on the 2500? I just want to know cause I don't want to spend money or time or the phone tag to get mine sent back in to fix faulty connections on the machine

look man, i'd just buy one. i doubt there is a hardware problem with the 2500. seriously. now, i'm not saying that there isn't a problem BUT it is really unlikely you'll have to send you unit back to the factory and whatnot. it's a great machine and i really like it. my recommendation as a user (and not as a seller - i don't sling mpc's) is that you should buy one.