RCA vs Digital
Curious if anyone has ever noticed a difference on a major PA. Normally, I work/DJ small/medium raves and everyone just uses digital. I’m assuming it’s cause it’s easier and sounds better (duh) but now I’m working a major festival with a very nice PA. The back line guy was using RCA and when I asked why, he said it doesn’t really make a difference. I figured, of all places, a major festival would have used digital. Thoughts?
Holy crap. It can make a huge difference. Think about feedback alone with using a data signal vs. analog signal. Digital outs are also balanced as opposed to the unbalanced RCA alternative. This just hurt my brain. What festival was it?
I think he means connection between the CDJ’s and the mixer, not mixer-FOH. I can easily imagine the difference is not audible, even on a festival PA.
That I can agree with.
Correct. And interesting… I’m currently working Outside Lands
RCA is a plug format, not a type of audio.
You can transfer digital over RCA plugs.
Back when I first put together a full CDJ-2000NXS and DJM-900NXS setup I plugged both the regular Line RCA and Digital to the mixer and to my ears there was a tiny difference in which the Digital sounded a bit more pronounce on the highs but hey those were my ears and not a RTA. I haven’t done any comparisson on my new NXS2 setup but I have both line and digital hooked up in my coffin at all times.
But to agree with others above, regular Line output can be susceptible to noise as it is an unbalanced signal. Digital can still be susceptible to ground loops and RF interferance but the odds of it happening are very very slim.
Good info to know. Thanks guys
If you use the digital, you lower latency also as you’re cutting out two conversions - D>A out of the CDJ and A>D on the mixer
That is a very good point as well. Especially with the already heavy load on the DSP inside of DJM series mixers. It’s not a huge difference, but still a cool observance. 
Ah yes, the latency. EDM DJ’s number one enemy. 
I would dare to say it’s a almost unmeasurable difference, I would love it if somebody proves me wrong.
Anyway, can we conclude that if you buy top-of-the-line, 6000$ worth of gear, it makes zero sense not to invest in a few digital cables?
Sonically, you will not hear a difference, even on the biggest sound systems at major festivals…
…However, there is one key point to note for why you should ALWAYS spring for digital connection when possible on pioneer gear:
When hooking up gear to DJMs with digital connections (either S/PDIF or USB) it is NOT POSSIBLE to clip the channel gains on the mixer. You can (but you shouldn’t) turn the gains and the EQ’s up 100% on the channels well into the “red” and it still will not be physically clipping (only on the channels, the Master Out & Booth Out is different. More on that in a minute).
I’ve contacted Pioneer reps directly on this to confirm and tested myself.
The reason is because of the 32bit floating point DSP (on NXS2, I think it’s 24bit on first gen 900nxs and djm2000) makes the headroom so huge that it is pretty much impossible to clip the gains on a digital signal.
When I set up gear for international artists on tour at major festivals I always use digital inputs, have the Master Attenuation at the lowest setting it will go (-12db on NXS2, -6db on DJM900nxs) and also booth attenuation on nxs2 the same. As well the Peak Limiter turned on in mixer settings.
This means that I can safely not worry if a DJ “redlines” (which happens WAY more than you think with major artists) because I know the signal leaving the mixer will not be distorted no matter how loud the DJ turns it up and I don’t have to play the silly game of explaining to them that if they turn it down, I’ll turn it up on the Front of House mixing console running the show sound system.
It also has the added benefit that if the previous/next DJ uses a Serato box (because they STILL haven’t gotten Serato Club Kit expansion in 2017 🙄
or some other analog sound card device, you don’t have to worry about unplugging a live signal cable as only the cdj’s will be digital.
So in summary, always use digital when possible simply because you can drive a mixer louder (which you don’t need to do, but DJs never learn…) without distortion, and it makes changeovers with other pieces of gear into the same mixer a breeze
Not only on Pioneer gear.
And I would count that as a con, not a pro argument. Why would I enable lousy DJ to pass as competent one? 
Isnt that Pioneer DJs entire business model?
I think the difference can be heard. I find less grunge on the highs and a clearer, more phase-accurate midrange. A little harder to tell on the bass.
Older digital Pioneer DJMs were 32bit DSP. NXS2 is 64bit. Other than their choice of dither to the output stages, it shouldn’t make any difference. None of their filter Qs are narrow/tight enough to get ringing errors in 32bit domain. Doesn’t hurt, though.
You’re right that using the SPDIF inputs prevents input clipping, but it’s to do with the ADCs being before a variable analog trim/gain stage rather than the type of DSP. Those trims on the Pioneer digital DJMs are analog controls when using the analog inputs, and digital domain when using the digital inputs. EQs and everything further down are digital domain for everything, obviously. That said, the DSP stage also cannot be clipped regardless of the input due to the floating point math. Pioneer added those Clip LEDs because the channel meters below don’t actually represent the input stage but the DSP section. You’re right that the outputs can be clipped, since they become fixed point with a real, finite scale. The Clip LED on the master, though, is just an extension of the master meter that previously served its purpose just fine. Seems they wanted to stick with a theme and keep it obvious.
The attenuation settings for the outputs, however, do not give you more headroom prior to the DJM output stages (sample depth/rate conversion, DACs, digital outs). All they’re for is conforming the DJM outputs to outboard gear’s input needs, nothing more. The max digital output level has always been strangely 5 or 6dB down from full scale, anyway.
I find it strange Pioneer still hasn’t allowed for the master out volume knob to have unity at max, as some other brands have allowed in their settings. But sticking it at about 1:15 and putting a blue speakon cap and tape over it does the trick. You could just turn the master out even lower to give yourself more master out headroom, but there’s nothing preventing the DJ from just crushing the DSP section even harder and getting to the same peaks as before on the outputs. You might as well just stick the knob at unity so the master out shows the same levels as the channel outs. Otherwise, if you’re running the master out at max, unity on the upfaders becomes 7 instead of max on those. That was more intended for the when Pioneer still had the rotary option so those guys could run the mixer without fussing with the trims.
A compressor limiter is also not all it’s cracked up to be in this kind of situation, either. It’s one thing if you’ve got a rock band, analog sound board, and a guy to babysit the levels. Adding one there can be a nice safety for someone going heavy metal caveman and smashing an electric guitar. Compressors are very useful obviously for production purposes and absolutely essential for the output stages of power amplifiers. Without a limiter on an amp’s output stage, you can get distortion peaks that are many multiples of the amp’s rated power, since the rated power is usually just the max output below 1% THD. Without a limiter on an amp’s internal output stage, the only thing preventing higher distorted outputs is the amp’s power supply and the venue’s circuit fuses that even will allow many amps more than they’re rated for a few milliseconds. Some of these cheap mega amp power supplies lack power factor correction caps and could potentially try to draw nearly 100 amps if they were unlimited and distorting!
A digital DJ mixer, however, is already a brick wall peak limiting device. Unless you’re running unprotected tweeters right up against their handling AND are confident your front-end source gear will never have issues (cables disconnecting, software glitches, etc) AND the content music never has any square wave-like aspects (good luck with that), having the limiter on does nothing advantageous except be more forgiving of stupid DJing habits by sounding better when they push the mixer too hard. That might be useful recording a one-off DJ mix in the studio where you want to smooth over mistakes and not ruin it, but in a live big-sound environment you’re just inviting its abuse and that can spell death to woofers by heat as a result of crushed dynamics over normal crest factor of program content. Even if you don’t cook your woofers, you’re still incentivizing less musical dynamics that can make people think the sound system is poor. If the DJ clips overtly, everyone knows whose fault that is. Whether they trigger a compressor limiter or hard clip digitally, dynamics are crushing either way. Let it sound like shit when they clip and they’ll do it less.
This made me chuckle 
dsquareddan is right about this.
If you’re running CDJs into a Digital mixer, you should be using the digital IO.
The ADC and DAC noise & artifacts are really subtle. CDJs supposedly use the same DACs as Bryston hi-fi CD players (which are pretty high end), and I can’t imagine the DJM is using ADCs that are worse. They’re probably not absolutely top of the line (Prism Sound, Dangerous, Lynx, etc.), but they’re good. And you will not hear flaws in the converters at a DJ gig. So, that’s largely academic. But, why introduce noise & flaws when you don’t have to?
The big one is that a lot of modern tracks are mastered so hot that they cause intersample modulation distortion. You can tell if it’s going to happen by running the tracks through an oversampling meter (there are a lot of them available for free), but unless there are features on the CDJs I don’t know about, there’s no way to avoid it except by keeping the signal digital through one more gain stage (which the mixer easily handles).
It’s fewer cables to screw up.
It keeps the line inputs clear for other DJs using the same mixer that may need them (e.g., controller setup not using the DJM’s sound card or other hardware like synths, samplers, etc., maybe if someon is using a DVS box that the mixer can’t replace for some reason).
If you’re at a festival with a sound guy, just let him do his thing as long as you can play your set. But, if you’re actually running sound, the digital IO is a better choice.
Before anyone challenges me on the cables…this is not the audiophile monster cable vs. coat hanger thing. I’ve heard it most obviously comparing George L guitar cables to pretty much any regular guitar cable. The George Ls just flat-out sound brighter, and you’d have to be basically deaf to miss it. That doesn’t, however, mean that they’re better…I don’t like the brighter sound and compensated for it with the tone controls on the guitar or the amp back when I owned a few. But, it’s very clearly different. All single-ended cables will exhibit this effect to a degree, though how much it affects things depends a lot on the level of the signal (mic, instrument, line, speaker), the length of the cable, and how the cable was built.
It happens because single ended cables literally are a resistor (nothing transmits electricity perfectly) and a capacitor (a hot and ground near each other separated by an insulator), which is exactly how you make an analog low pass filter. Fortunately, they’re crappy low pass filters. But, it still happens.
They’re also crappy antennas because, again, that’s how you build antennas. It’s easiest to hear that with a long, unbalanced mic cable. You have to apply so much more gain that the radio is often audible after the preamp.
That’s why analog, single-ended cable runs are supposed to be short. It minimizes all of those effects.
Using a 2-foot or so cable like you’d use to connect a CDJ to a DJM, neither effect is drastic. But at least in an academic context, it’s there. And if you can avoid it by using hardware that’s already there, why not?
I have to disagree with point 2. Overly-hot masters (above -3dBFS) contribute to intersample clipping within the digital domain during aggressive DSP post-processing by the end user. It’s unlikely, though possible, that the CDJs have this issue during their master tempo processing, which is the only processing that occurs on them when activated and when the pitch is not on zero. Not sure what DSP architecture newer Pioneers use, but I think even old Denon media players used 32bit float, though some of those had a lot of effects, too. If CDJs did have such an issue, though, it would affect a CDJ’s output equally regardless of whether a SPDIF connector was used or analog RCAs, since SPDIF is not floating point and has just as much a hard clip limit as the DACs have once you bother processing the signal at all.
FYI: Pioneer CDJs are bit-perfect over the SPDIF when master tempo is off or pitch is at zero and no DSP processing occurs. Now, as I said previously, there is absolutely a utility in avoiding clipping on the ADC end of the mixer by going digital, especially considering Pioneers’ variable-gain-trim inputs into those ADCs and the digital domain meters that are of limited usefulness… to say nothing of the obvious advantages (however subtle they may be for some people) or cable management convenience you mention.
Still, I am curious if some media players have issues with intersample clipping, especially those with a lot more processing and at budget prices. I’m also reminded by your comment of the fact that Pioneer drops its digital signals on the digital DJMs by about -5dBFS prior to the fixed point stages of the sample rate converter & SPDIF and the DAC. Perhaps that is to prevent intersample distortion in that last 3dB of headroom, not that DJs should be utilizing that headroom. I’ve wondered about that decision.
The problem is that intersample peaks clip the DAC itself.
I believe they are using floating point math internally, which means the DSP happening inside the CDJ shouldn’t cause intersample peaks as long as they do something sane before the output. The only reason to use floating point math in that instance is for virtually limitless dynamic range, and it should be easy for them to figure out if the timestretching algorithm would cause an increase in gain and ajust for it. So, that’s a moot point.
The DAC itself is fixed point, just like every other DAC. That’s where the clipping happens.
If you have a track that runs all the way up to 0dBFS (but isn’t clipped itself), transfer it digitally (zero gain), then turn it down to have enough headroom before it hits a DAC, you prevent the intersample peaks.
Which is exactly what happens if you run digital into the DJM and then don’t hammer it’s outputs.
As long as they didn’t do anything stupid, the processing inside the CDJ should be virtually meaningless, at least in this regard.
You mean the oversampling processing stage on the DAC? Maybe an issue with very hot masters. Ok, I can buy that as perhaps another free benefit from using the SPDIFs when master tempo is off or you’re at zero pitch. The NXS2 DAC is 32bit and state of the art, though, not some el cheapo. Perhaps its own headroom’s already accounting for that. The master tempo is also an oversampling processing, and the SPDIF is also fixed point and requires its own transcoding & anti-alias filtering. Pioneer doesn’t drop the signal at all on their CDJs in the digital domain. Maybe they should or at least provide the option? Admittedly, any actual interpeak clipping from master tempo when away from zero pitch would be like a bird fart in a hurricane. In contrast, if we want to talk about preventing clipping by using the SPDIF on the CDJs, the ADCs on the DJMs are variable-gain-trim and the channel meters are useless without that “Clip” word above them for making sure the analog inputs are good.