12 bit vs 16 bit

Ill-Green wrote:Doesn't the 1000 have BIT GRUNGER?

yep

Funkydrops wrote:

Ill-Green wrote:Doesn't the 1000 have BIT GRUNGER?

yep

Reading other posts though, it dosnt seem to come close to a real low bit resolution sound. More like a daws way aggressive bit crusher. I would just jjos's bit conversion, it's still no s950 but it's a lot more convincing than the bit grunger.

Hell sometimes I just use iTunes for batch bit/sample rate reduction, not 12bit though. In fact I've never seen a single piece software do 12bit exporting/conversion.

m56p87 wrote:

Funkydrops wrote:

Ill-Green wrote:Doesn't the 1000 have BIT GRUNGER?

yep

Reading other posts though, it dosnt seem to come close to a real low bit resolution sound. More like a daws way aggressive bit crusher. I would just jjos's bit conversion, it's still no s950 but it's a lot more convincing than the bit grunger.

Hell sometimes I just use iTunes for batch bit/sample rate reduction, not 12bit though. In fact I've never seen a single piece software do 12bit exporting/conversion.

Ok, the point here is that convert to 12bit a 16bit file cannot get the same result as in recording the same source at 12bit. The data gap between the two things cannot be calculated in abstract, only given the 16bit version. Not talking about the different sound that different A/D converters have. Just the fact to sample at a certain bit-depth influence in a not-linear fashion the way the signal is translated.

There is no straight-forward algorithm that can do that.

16bit are 4 group of 4bit, while 12bit are 3 groups of 4bit: bit-reducing skips the last group of 4 in the 16bit file, much like skipping the decimal part of a real number.

16bit===> Say you have 79,6.
Bit reduce ==>Skip 0,6
you have 79.

Now sample the source at 12bit: YOU GOT 80!

Some software emulate older samplers tonal quality and offer an approximation: Decimator is recognized as being pretty decent at it.

Funkydrops wrote:
16bit are 4 group of 4bit, while 12bit are 3 groups of 4bit: bit-reducing skips the last group of 4 in the 16bit file, much like skipping the decimal part of a real number.

16bit===> Say you have 79,6.
Bit reduce ==>Skip 0,6
you have 79.

Now sample the source at 12bit: YOU GOT 80!

Would this be true for all 12bit down converting? And if you were doing 16to lets say 8bit would it just drop the last 8bits too or would it distribute the loss of data differently?

m56p87 wrote:

Funkydrops wrote:
16bit are 4 group of 4bit, while 12bit are 3 groups of 4bit: bit-reducing skips the last group of 4 in the 16bit file, much like skipping the decimal part of a real number.

16bit===> Say you have 79,6.
Bit reduce ==>Skip 0,6
you have 79.

Now sample the source at 12bit: YOU GOT 80!

Would this be true for all 12bit down converting? And if you were doing 16to lets say 8bit would it just drop the last 8bits too or would it distribute the loss of data differently?

Bit-reducing, per se, it's just dropping less significant bits. Which has an audible effect on the signal represented that you may also like.

It's just that. I guess AKAI Bit-Crusher combines that reduction with a sort of distorsion, which also adds harmonics to the sample, while I think that JJOS bit-conversion does not add any distortion.

You can 'simulate the sound' of a 12-bit sampler, playing with bit-reduction/bit-crushing, slight ring-modulation and filtering... sort of.

Or just go with the Decimort: http://www.d16.pl/decimort

Instead of trying to force the mpc into being something its not why not just sequence the rack with the mpc and just sample directly with the sampler? Makes more sense to me. When I wanted a 12 bit sound I just got a 12 bit sampler and used it how it was meant to be used. I kinda miss it I may have to hit ebay and get another.

KevWestBeats wrote:Instead of trying to force the mpc into being something its not why not just sequence the rack with the mpc and just sample directly with the sampler? Makes more sense to me. When I wanted a 12 bit sound I just got a 12 bit sampler and used it how it was meant to be used. I kinda miss it I may have to hit ebay and get another.

This.



All MPC's (other than the 60) have a data rate of 44.1kHz, 16-bit. Even a "12 bit converted" sample will be saved at the 44.1, 16-bit specification.

resample on the mpc4000 allows you to change bits data rate and other parameters as it's resampled

Funkydrops wrote:

m56p87 wrote:

Funkydrops wrote:
16bit are 4 group of 4bit, while 12bit are 3 groups of 4bit: bit-reducing skips the last group of 4 in the 16bit file, much like skipping the decimal part of a real number.

16bit===> Say you have 79,6.
Bit reduce ==>Skip 0,6
you have 79.

Now sample the source at 12bit: YOU GOT 80!

Would this be true for all 12bit down converting? And if you were doing 16to lets say 8bit would it just drop the last 8bits too or would it distribute the loss of data differently?

Bit-reducing, per se, it's just dropping less significant bits. Which has an audible effect on the signal represented that you may also like.

It's just that. I guess AKAI Bit-Crusher combines that reduction with a sort of distorsion, which also adds harmonics to the sample, while I think that JJOS bit-conversion does not add any distortion.

You can 'simulate the sound' of a 12-bit sampler, playing with bit-reduction/bit-crushing, slight ring-modulation and filtering... sort of.

Or just go with the Decimort: http://www.d16.pl/decimort

Yeah I'm familiar with that side of things, I was just wondering if there are any inherent differences in the algorithms used to encode data at different bit rates. (other than the loss of data you mentioned).

jonestown massacre wrote:

KevWestBeats wrote:Instead of trying to force the mpc into being something its not why not just sequence the rack with the mpc and just sample directly with the sampler? Makes more sense to me. When I wanted a 12 bit sound I just got a 12 bit sampler and used it how it was meant to be used. I kinda miss it I may have to hit ebay and get another.

This.



All MPC's (other than the 60) have a data rate of 44.1kHz, 16-bit. Even a "12 bit converted" sample will be saved at the 44.1, 16-bit specification.

So what if its saved in 16bit 44khz.. When people go after a lofi sound they aren't doing it to save disk space.. You can't create data that isn't there so if you took a 12bit sample @ 26khz and sampled it into an MPC 16bit 44khz it would still sound exactly like the original 12bit sample.

m56p87 wrote:

jonestown massacre wrote:

KevWestBeats wrote:Instead of trying to force the mpc into being something its not why not just sequence the rack with the mpc and just sample directly with the sampler? Makes more sense to me. When I wanted a 12 bit sound I just got a 12 bit sampler and used it how it was meant to be used. I kinda miss it I may have to hit ebay and get another.

This.



All MPC's (other than the 60) have a data rate of 44.1kHz, 16-bit. Even a "12 bit converted" sample will be saved at the 44.1, 16-bit specification.

So what if its saved in 16bit 44khz.. When people go after a lofi sound they aren't doing it to save disk space.. You can't create data that isn't there so if you took a 12bit sample @ 26khz and sampled it into an MPC 16bit 44khz it would still sound exactly like the original 12bit sample.

Theres more to an ADC/DAC than that... There is going to be coloration, aliasing and other artifacts than the "resample" function is not going to give you. You are using 16 bit digital copy of an analog sound, and down converting the 16 bit digital sample into a 12-bit emulation, still in 16-bit. See whats happening? In a 12-bit sampler, the analog sound is converted straight to a 12-bit digital "image". Its like recording a movie with an HD camera and using an 8mm effects plug-in VS. using an 8mm camera. That being said, there are probably good 12-bit emulations out there, Im not up on the best software......


There is also a resample function on my 2KXL. If I sample the same sound into the S900 and into the XL, use the resample function with the exact same settings in the XL, they dont sound the same. If you sample the same sounds into an EMAX and a S950, they dont sound the same either.....both 12-bit, feel me?

I was not speaking on disk space, although 12 bit is obviously going to take up less. Most hardware samplers have upgrades that make disk space a non issue anyways.

MoreBuck$ wrote:resample on the mpc4000 allows you to change bits data rate and other parameters as it's resampled

The 60 and the 4000 are not 16 bit....

jonestown massacre wrote:
Theres more to an ADC/DAC than that.....

No doubt, I was speaking purely about the quantifiable data encoding not the abstract coloring of sound that occurs with the conversion. I'd easily say that the a/d conversion and the code specific to akai in that conversion is the major coloring factor, if you simply just drop bit rates (not sampling rates) all your doing is adding noise - nothing else.

m56p87 wrote:if you simply just drop bit rates (not sampling rates) all your doing is adding noise - nothing else.

I dont think so. search for a-law and µ-law and you know what I mean.

peace

jonestown massacre wrote:

MoreBuck$ wrote:resample on the mpc4000 allows you to change bits data rate and other parameters as it's resampled

The 60 and the 4000 are not 16 bit....

My 4K is set up to sample at 16bit and then i can resample with a lower data rate and bit rate to achieve Lo-Fi sounds.