Input Levels: Hotter on the Right

I was thinking I should post this in one of the more generic forms but since I loaned out my 1000 to a friend I can't verify it on anything other than my 2500 so I'm here. I also didn't find anything in a search for "Input AND Level" that seemed to address my question so here goes...

I have noticed that when sampling into my MPC in stereo that the right channel is hotter than the left. I assumed at first that this was a mixer setting (I have a mixing console and patchbay that the signal passes through) so I reset the levels for the mixer channels I was sampling from and made sure they were identical. (gains were equal, all eq off, fader at unity. everything was the same with no effects or any external signal modifications). I fed it a simple kick drum sound from my MachineDrum perfectly panned to center. The right side was still hotter. I then directly wired my Machinedrum main outs to the inputs on the MPC (with different cables). Still hotter on the right. I went into Logic and used Ultrabeat to individually send a solo kick, then a snare, then a single bass note (c0). All hotter on the right. I used the TestOscillator in Logic to send a constant tone (as sine, square, and noise). All hotter on the right on the MPC.

I would expect that the signal would be equal in both channels until I start panning and appling dynamics but it is consistently off. The same tests on my FantomX8 have the expected results: equal signal in both channels.

So, my question is three fold. 1. Has anybody else noticed that the right channel is hotter on their MPC (any model but 2500 specifically may be more helpful here)? 2. is there a setting somewhere that I'm missisng on the MPC that may cure this? 3. Based on what I've done can anybody think of how this may be a problem with what I'm doing? I feel like I've done a pretty decent test (taken out all extraneous gear, changed cables, used multiple sound sources and types, verified expected behavior in other gear, etc) but I'm still open to this being user error here. I seem to remember this behavior on my 1000 as well but like I said, I can't test it right now. I guess most of the time when I'm sampling things I just figure that the reason for the signal stregnth discrepency I've seen is due to the dynamics of what I'm sampling but today I got curious when I noticed a beat I was sampling that I would have expected to be an equal signal wasn't.

Thanks for help or ideas.

From my experience with MPCs, this is common.

I just duplicated this test with a sine wave coming from my O2R. When I have the sine wave coming into the ANALOG inputs of the 2500, I get varying results..the left channel was a little hotter. I used the same cables and switch the right and left channels. The right channel was a little hotter. The one thing I didn notice is that if I kept the record gain of the MPC 2500 low, I get pretty even results...as I increase the gain, I get a little boost in either the left or right channel. I did this test using unbalanced cables.

I performed the same test using the digital inputs and got even results on both the left and right channel.

Oh, I should have mentioned that I had the record gain all the way down on the MPC. I try to boost the signal earlier in the chain so I tend not to use the MPC gain at all or very little.

Ideally I would use the digital output on my 828mkII and feed the MPC but it comes out at 24bit and that's not configurable so it just doesn't work right. I can always boost the weaker channel more on the mixer bus that feeds the samplers to get it lined up but was wondering if I'd done something wrong.

Thanks[/i]

zs941 wrote:
Ideally I would use the digital output on my 828mkII and feed the MPC but it comes out at 24bit and that's not configurable so it just doesn't work right.

So..are you having problems recording 24bit digital audio into the MPC 2500s DIGITAL SPDIF INPUT?

YES!

so im not crazy huh. i noticed this too, machinedrum panned centre, into mpc1k l/r... hotter on the right side. using the mpc's input gain seems to make it worse.

should we start complainin to akai, cause actually yeah this is annoying.

ps. im talking about anlog inputs here.

As for the 24bit digital output to the MPC's 16bit digital input that is always a problem. The MPC has no inteligent dithering for higher bit-rate input: it just chops 8 bits off the bottom so you lose data. I've emailed Akai about this in the past and that's how they explained it to me. Unless Akai added a bit converter to the 2500 (which is very unlikely. i've never seen mention of it) you'll lose data. Because the output of the 828 is true 24bit (meaning, it doesn't just pad 0's on either end of a 16 bit signal) I'd be lossing real information which isn't good. Sometimes this isn't a big deal but sometimes it is. Now, if I send a 24bit signal into the MPC it will record something (it doesn't just reject the data) but it's not the correct data. I did a bunch of tests with my MPC1000 quite a while back but I haven't repeated them with the 2500 so maybe I'm wrong.

You have a FantomX, Jahrome so maybe you could try the following test: sample something to the FantomX (or just resample/skip back sample something). Make sure you have no gain adjustment on the sample (in the Params page - it defaults to +6db so set it to 0 just to be sure it's pure). Replay the sample a few times and get it in your head the way it should sound. Then, sample that output on the MPC with the SPDIF input fed from the FantomX. Play that same sample on the MPC and it will sound wrong. The Fantom is a weird beast because it's internal sample engine is 16bit (which is good because you can easily move samples back and forth from the MPC with no problem) yet it's input and output are both 24bit (to allow enough headroom for the summed output with all voices played - easily a 24bit number). The Fantom will generally just pad 0's to the top of the outbound signal to fill the bits however the MPC doesn't remove those top 0 bits but instead removes bits from the bottom (because it assumes I guess that you want amplitude over low bits - makes sense, really). So, what you end up with is 16bits of data - the top half of which are 0's. Which means you now have an 8bit signal, really.

No fault of Akai's because they don't advertise 24bit input/output for the 2500 but it is something people shoud be careful with. Most audio equipment is 24bit these days so my guess is that Akai mainly figured that people would be sampling from the SPDIF outs on CD players which are still often 16bit. Also, using the 16bit output from the MPC to a 24bit input on another interface shouldn't be a problem - assuming of course that it doesn't pad the bottom pushing your sample up way too high!

(P.S. All of the above assumes you're dealing with 44.1kHz sample rate on both ends. If you start plugging 24/48kHz or higher outputs into the 16/44.1kHz digital input on the MPC it's gonna sound f*cked up...)

YES!

so im not crazy huh. i noticed this too, machinedrum panned centre, into mpc1k l/r... hotter on the right side. using the mpc's input gain seems to make it worse.

should we start complainin to akai, cause actually yeah this is annoying.

ps. im talking about anlog inputs here.

No use of complaining. It is really not that much of a difference and can't see why it would be annoying. In fact, not only the MPC does this. I have plenty of synths connected to my digital mixers. The L/R channels of these synths do not always register exactly the same on my mixers meterbridge. An I am using sytbs from Korg, Yamaha, and Roland.

I would try this test again using balanced cables..but the sound source would have to be balanced as well.

zs941 wrote:
Most audio equipment is 24bit these days so my guess is that Akai mainly figured that people would be sampling from the SPDIF outs on CD players which are still often 16bit. Also, using the 16bit output from the MPC to a 24bit input on another interface shouldn't be a problem - assuming of course that it doesn't pad the bottom pushing your sample up way too high!

I will attempt your test....

Most audio equipment is 24 bit?? Are you sure about that? The higher end gear that is 24 bit can dither down to 16 bits. As far as the MPC 2500, use its analog inputs/outputs. They are all balanced. I get exceptional results recording into the analog inputs of my mixers.

No use of complaining. It is really not that much of a difference and can't see why it would be annoying. In fact, not only the MPC does this. I have plenty of synths connected to my digital mixers. The L/R channels of these synths do not always register exactly the same on my mixers meterbridge. An I am using sytbs from Korg, Yamaha, and Roland.

I would try this test again using balanced cables..but the sound source would have to be balanced as well.

for my part i am using balanced cables. they're all high-quality, correctly shielded, and with gold connectors. i've also tried unbalanced cables and gotten the exact same result. personally i think it is worth reporting the problem. my other sampling gear does not have the same problem. *shrug*. i'm not terribly worried about it because i can use my mixer to boost input channels to the mpc individually but then i raise the noise floor in one channel which isn't ideal.

Most audio equipment is 24 bit?? Are you sure about that? The higher end gear that is 24 bit can dither down to 16 bits. As far as the MPC 2500, use its analog inputs/outputs. They are all balanced. I get exceptional results recording into the analog inputs of my mixers.

Yeah, I'm sure about that. Just about every device with a digital out spits data out at 24-Bit: high- or low-end gear. Where they begin to differ is sample rate. If you don't believe me just go read the specs for a handful or gear with a digital out. A search at Sweetwater for "16-Bit" and "24-Bit" returned 7 and 209 respectively so just read some of those specs. Can you give me an example of a piece of gear that has a built-in bit-rate converter that isn't simply that? I can't say I've ever seen a keyboard, sampler, or synth with digital inputs/outputs that could correctly convert a signal but that doesn't mean they don't exist.

Also, I was using the analog inputs/outputs. That is what started this thread. I thought it was clear in my original posts but just in case: I am using Analog Inputs and Outputs. I am not using the Digital Inputs and Outputs. All of my gear including my 828 use 24-Bit/44.1kHz outs which will not work on my MPC. I am using balanced cables in a completely balananced chain. Using Analog inputs is where the problem appears to be.

Do you use the digital inputs on your MPC to sample? Do you still get channel signal level differences? Also, what sort of digital mixer do you use?

jahrome wrote:

YES!

so im not crazy huh. i noticed this too, machinedrum panned centre, into mpc1k l/r... hotter on the right side. using the mpc's input gain seems to make it worse.

should we start complainin to akai, cause actually yeah this is annoying.

ps. im talking about anlog inputs here.

No use of complaining. It is really not that much of a difference and can't see why it would be annoying.

Um, well Jahrome, it is annoying because I'm trying to sample material which should have a kick or bassline panned dead centre, and when i listen to it the kick is off to the right- yes that's annoying!!

And to fix it maybe i have to tweak on my mixer or pan in the mpc- guess what, also annoying.

I am using Yamaha 02R mixing consoles. Minutes ago, I did the same test with the Fantom XRs sampler. I got the same results as I did with my MPC. What other sampling gear are you using?


If you are getting uneven results, it is most likely hardware related. I have noticed similar results in a lot of gear.

i have a fantomx8 and a vsynthxt. the fantomx8 didn't exhibit the same behavior and i haven't tried the vsynth.

the Yamaha 02R is 24/48, right? have you ever fed the spdif out to the the mpc input?

I tested both the Akai MPC 2500, Fantom XR, and the Fantom X6. All produced similar results.

I used the Oscilator of my Yamaha 02RV2 that produced a 1Khz Sine wave at -20dB. I sampled this in stereo into the MPC 2500, Fantom XR, and the Fantom X6. When I view the left and right channels of the samples in ALL of these units, the left and right channel of the sample were/are not identical. Also, while sampling..at no time did the left/right input meters of any of the samplers register at the same dB at the same time. I carefully monitored the peaks of each machine. The left/right channel of each machine peaked at the same dB level..but at no time were the channels of the incoming samples at the same dB.

The 02R sample rate is 44.1/48Khz. The analog stere outputs of the 02R is 20 bits balanced.

I'm curious about something. Do you sample into your 2500 via SPDIF from any source with an output bit rate over 16 or do you mainly use the analog inputs? Do you find the results to be ok? If so, what sort of stuff do you normally sample that way. It would be good to have some examples of samples types that may work well.

zs941 wrote:...The Fantom will generally just pad 0's to the top of the outbound signal to fill the bits however the MPC doesn't remove those top 0 bits but instead removes bits from the bottom (because it assumes I guess that you want amplitude over low bits - makes sense, really). So, what you end up with is 16bits of data - the top half of which are 0's. Which means you now have an 8bit signal, really...

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.

zs941 wrote:
I'm curious about something. Do you sample into your 2500 via SPDIF from any source with an output bit rate over 16 or do you mainly use the analog inputs? Do you find the results to be ok? If so, what sort of stuff do you normally sample that way. It would be good to have some examples of samples types that may work well.

I usually sample thru the analog inputs. Previously, when I sampled from CDs, I used a Denon CD player w/ SPDIF digital out. Now, I just use the CD-M25 installed in the 2500 for sampling from CDs.

Since the start of this thread, I have been looking more into sampling digital sources at different rates. It appears that the MPC 2500 can dither samples and do a very good job at it.

My test: I recorded a Jay-Z vocal into Nuendo 3 at 24 bits and a sample rate of 48Khz. I sent this track into my 02R which is set in 24 bit mode and a sample rate of 48Khz. The 02R and my PC is locked via BNC cables. When I sampled this into the 2500, it recorded the vocals without degrading the signal but did transpose the sample down a semitone.