mpc2500 internal mixer

poundaproblem wrote:Same with the 1k. Kinda annoying actually because when a kick is a touch too loud, turning it down makes it too low.

Yup, that's one thing the XL has over the 2500, nice sensative faders. That and it's real easy to "select all" faders in the mix window on the XL. But the thing with JJOS that makes up for the select all feature is the global params settings. However, concerning the sensativity of the volume, the volume adjustment in the param screen is pretty effective. So I use that mostly.

The resolution of a sample is lower when the volume is lower - summat like each 6db below 0db is 1bit depth loss, I think...

I haven't a clue how the MPC goes about doing the math for the mixer & output section tho...

If you sample summat 'perfectly' at full volume the loudest part is recorded with 16bits of info - if you record the same sample at half volume it's recorded with a lower bit depth and so on...

It might still be a 16 bit file or output but it only has say 8bits worth of 'real' data

Spirit wrote:The resolution of a sample is lower when the volume is lower - summat like each 6db below 0db is 1bit depth loss, I think...

I haven't a clue how the MPC goes about doing the math for the mixer & output section tho...

If you sample summat 'perfectly' at full volume the loudest part is recorded with 16bits of info - if you record the same sample at half volume it's recorded with a lower bit depth and so on...

It might still be a 16 bit file or output but it only has say 8bits worth of 'real' data

That sounds inaccurate. The gain should merely change the output volume, not the characteristics of the sample. It's like a digital picture. If you got an 8MP picture and zoom out to 50%, the resoluting of the picture doesn't change, it just shows up smaller on the screen.

Spirit wrote:The resolution of a sample is lower when the volume is lower - summat like each 6db below 0db is 1bit depth loss, I think...

I haven't a clue how the MPC goes about doing the math for the mixer & output section tho...

If you sample summat 'perfectly' at full volume the loudest part is recorded with 16bits of info - if you record the same sample at half volume it's recorded with a lower bit depth and so on...

It might still be a 16 bit file or output but it only has say 8bits worth of 'real' data

I really wonder where you guys learn stuff like that. The bit rate determines with which resolution you can save volume (or amplitude in other words) information of a sample. You don't "loose" bit depth just because that volume value is lower.

What you probably mean is when your sample amplitude corresponds to the maximum value that 16 bits can represent, which is 65535, a sample amplitude corresponding to 8 bit, which is 255, is accordingly lower in volume. Now it could be that this is exactly 48 dB less, but I have never before heard of that relation.

Am quite happy to be wrong on this one - it was just summat off the top of my head that I remembered, perhaps incorrectly, from way back...

...it may have been an 'equivalent to' or related more to quantisation/distortion.

or just summat from the bit below out of remixmag applied in reverse

bit resolution. Whereas a greater sample rate allows more of a sound to be sampled over time, a higher bit resolution allows for a better dynamic range of that sound, and the two are inextricably linked.

Think of bits as representing the maximum range of dynamics possible. Every bit yields approximately 6 decibels (dB) of dynamics.

Therefore, 24 bits yield a maximum of 144 dB (6×24) while 8 bits yield only 48 dB. That is a huge difference! So, for example, with a lower bit depth, a sound with a long fading decay might get truncated much too soon due to the inherent lack of possible dynamic range. And that dynamic-range limit happens consistently across each individual sample recorded. If you multiply that one instance by every sound in a given song, you begin to see the problem.

So 16bit = 96dB ...and 8bit = 48dB - which is 48dB less


Anyways, a sample recorded at a lower level will still only have X bits worth of 'real' data... You don't lose bit depth as such, it's still a 16bit file., but everything above the value of 255 doesn't contain anything useful or meaningful as far as I (mis)understand it

The only way the camera analogy works is if you take another picture from further away i.e. 'quieter' zoom out in real life and then tried to zoom in afterwards.

A better analogy would've been to compare a normal digital camera with a HDR camera as the MP is more comparable to the sampling frequency than the dynamic range of the light it can capture.


Buy hey, who knows, I ain't a mathematician (clearly) so feel free to shoot me if I'm wrong

I was referring to zooming out in like photshop or windows picture viewer or somethin, where the picture file has already been recorded. I wasn't talking about using the zoom feature on the camera.

Anyway, I thought you were originally talking aout lowering the volume of a sample, not recording it at a lower volume. Either way, if you record something at -10dB or at 0dB both at 16 bit, al the same sound data will still be there. The only difference I really see is that the one recorded at he lower volume may have more audile air/white noise. But othere then that, the resolution will be the same.

Spirit wrote:but everything above the value of 255 doesn't contain anything useful or meaningful as far as I (mis)understand it

Not quite...no matter if a signal is at a value of 255 or 65535, it's not more or less useful. There is still the same difference in volume per "step" (like from 255 to 256), only at a lower volume.

What your source tries to explain is more the difference of different bit depths of a sample altogether. The volume changes per "step" is much higher in an 8 Bit sample than in a 16 Bit sample. And in addition, the last step, from 0 to 1, represents the lowest volume the sample can contain - i.e. 1/255 of the maximum value. In a 16 Bit sample it would be 1/65535. You see the difference. This is what is called the "dynamic range".

Not quite...no matter if a signal is at a value of 255 or 65535, it's not more or less useful. There is still the same difference in volume per "step" (like from 255 to 256), only at a lower volume.

Not at all, this all depends on how the information is quantised and whether it is liner or non-linear quantisation... e.g. a linear quantisation error from 2.5 to 2 is greater proportionally than from 99.5 to 99 (20% and 0.005%)

This is why I was saying I don't have a clue how Akai are doing their math in the mixer as there are many ways of going about the same thing with different results and so much of it is inextricably linked - liked showing you the 48dB relation you'd never heard of but you seem to want to ignore now it's been shown to you...

If you're only actually using 12 or 8 bits of dynamic range then it doesn't matter if the system has enough headroom to cope with 16bits of dynamic range - you are still only using 12bits of meaningful data

Guess we'll simply have to agree to disagree on this one

Spirit wrote:Not at all, this all depends on how the information is quantised and whether it is liner or non-linear quantisation... e.g. a linear quantisation error from 2.5 to 2 is greater proportionally than from 99.5 to 99 (20% and 0.005%)

This is why I was saying I don't have a clue how Akai are doing their math in the mixer as there are many ways of going about the same thing with different results and so much of it is inextricably linked - liked showing you the 48dB relation you'd never heard of but you seem to want to ignore now it's been shown to you...

I don't ignore anything, I try to understand by thinking myself, not reading some explanation which I still don't need the source of. Of course you can quantize non-linearly, but then I guess you will also get another dynamics-per-bit relation, whether it is 6 dB or something else.

My point is only, as stated by others as well, that reducing volume does not change your samples bit rate.