Pioneer has everything running capped at an artificial hard limit around -5dBFS, possibly to reduce inter-sample errors. The SPIDIF will never go past this.

It’s actually a bit below. Ean had it just a little too hot in the video.

Don’t take my word for it… if you have a chance to play on one, try it out 🙂

I suggest you look up stuff on signal chain theory and unity gain. What you’re describing is exactly what you’re not supposed to do. Equipment is made to work at unity gain. Best S/N ratio is at unity gain. If you running hotter, what headroom do you have left when two signals are mixed together or when you apply FX? None. You get distortion and digital distortion is not pretty.

Unity gain is not a setting. It’s what 0db of amplification is called. I’m fairly sure that you can set a Xone DB4 to produce 0db of amplification, no?

The problem is not just clipping but it’s mainly gain structure. For the best S/N ratio, your signal chain should always be at unity gain at every stage.

Because a lot of DJs nowadays don’t know what they are doing (even pro ones) and like to run everything hot, manufacturers like Rane, Pioneer and others attenuate the signal to give the mixer some headroom. On DJM900 it’s a setting that club owners can access and put, for example, 6db of headroom so the DJ can’t screw it up. DJ sees everything light up yellow/red and thinks that they are playing ‘louder’.

But of course, they’re not. The mixer is attenuating by 6db. So the club has to amplify the signal again to get the signal back at unity gain. They do that in their amplifiers before it hits they compressors/limiters.

Result: The signal chain is not optimal. It doesn’t sound as good as it could.

This is not an opinion, btw, it’s how signal chains work. Hope it makes things a bit clearer.

At the levels I was playing there was no digital distortion or un-desirable artifacts, however when REALLY pushing it hard on the redline it became pretty crunchy – however most mixers do that as well.

Please see my above post. Also, when did it artifact in the video? Unless Ean had the master level pot ALL the way up with the mixer red-lining hard, there was no chance of the mixer clipping.

Thanks for the link:) That article is in regards to using typical mixing boards but regardless my analogy still holds true. They mention “comfortable level” below clipping. This “comfortable level” is just to give you some space to add water without worrying about clipping the signal. This may be necessary when mixing a live band but as DJs we’re mixing fully mastered songs with bass, drums, vocals etc.

The hotter the input level is without clipping the lower your signal to noise ratio is going to be. As long as you don’t clip any of the gain stages there is no problem with distortion and you’re achieving the hottest and cleanest signal on the output. Regardless of how many channels of audio your mixing you want the level of all the audio to stay consistent. The glass should be full at all times.

If you are running as hot as possible without clipping on channel 1 and you bring in a track on channel 2 you need to make vol and/or eq adjustments on channel 1 as to not spill the water over the edges of the glass (distort the signal). Yes.. if you put both channels at full volume and blend them together water is going to be all over the floor.

I use Traktor and mix four decks.

Running at just under clipping gives me a reference point regardless of how many tracks I put in the mix and that reference point is as hot as it can possibly be without clipping. If at anytime I hit red I know I need to turn down the inputs as to keep the overall volume consistent and not distort the sound. That reference point could be lowered to, say, the first yellow LED but then I’m not maximizing the signal to noise ratio and I’m achieving a lower output.

Of course this discussion is totally different when using the MP2015 as its impossible to clip the inputs to distortion. The ONLY way to distort the audio on the MP2015 is to turn up the main volume all the way to 10. You could have every input knob turned all the way up and it STILL won’t distort if your main volume knob is lower than 10. The headroom on this mixer is insane.

I’m stoked you picked one up and are enjoying it:)

False.

Intersample errors begins around -3dBFS.

Thanks for the numbers Shawn. As I was saying earlier, I understand why you guys scale the gain internally but I still think it would be better to educate users to set up proper gain structure rather than display misleading metering in order to let them have all the yellow lights on 🙂

Thanks Shaun for the more specific details which I didn’t have.

Hmm… it sounds like either the float to fixed point conversion is way offset to give a ludicrous amount of real output headroom and the designers are now worried either about losing low-level significant resolution digits in the conversion or in underutilized analog total dynamic range, and/or there’s some fudging of the recommended gain usage now to assuage those unjustified complaints about the mp2015 headphone amp, and/or there’s a limiter’s threshold triggering somewhere in the yellows. On the latter, I can see the meters bounce less dynamically when they are run as you say, which is consistent with a compressor/limiter but I suppose could be just an issue of meter response if there is no limiter.

Either way, if the meter says Overload and that’s not actually its 0dBFS, which is what the meters should be reflecting since every output on the unit, digital or analog, must conform to that fixed-point scale, it’s unprofessional meter application. Furthermore, recommending everyone run every peak at exactly the same spot fundamentally misunderstands the varying RMS of musical content and is bad advice for DJing produced & mastered music. If you have a limiter in the DSP code signal path, please add the option to disable it in the next firmware. I also recommend you add a digital input attenuator setting for the SPDIFs, say, 0dB, -6dB, and -12dB. And the utility option to make the isolator audible in the headphones program signal would also be nice.

Same here. I have tried that approach. The problem with it is that you lose your line level output gain, lets say for a record/session out of the mixer will be well below par. The truth is in order to get the full signal path out of the mixer it is forcing us to ride that one -10 light. Which as I have seen in my tests you can turn the input level around 2 notches and hear an audible change in of about 5 db yet and still be on the same -10 light. Songs come in louder/lower even though this one light is lit.