Is DP's summing wack?

[qo]

i dont know how much pink or white noise is goign to tell you. there are also impulse tests to consider as well. system recovery etc.

Good suggestion to test impulse as well. Yes, you're right that pink/white may not tell us anything. Or may tell us a lot, depending on whether we get complete cancellation. We'll know in a couple days!

[alphajerk]

why not just take say a kick hit... push it up +6.02db on the fader of the track, take down the fader on the master track -6.02db and see what happens. or take a sum of a song instead of white noise.

basically you are testing the rounding "error" on the floating 8bits... which is going to happen the same every time, its not random. that is the big different between analog and digital summing, analog will pary a little bit each pass due to quantum physics of the electronics. its not pure like digital math.

one thing i noticed last night while rough mixing a song i am working on and this -6db "rule"... if i record with peaks in the -12 to -9dbFS range, even with the individual fader at +6.02db [highest the fader can go] it still peaks that channel at -6db [assuming all the gainstaging is proper on any plugins [which there arent any at this point]

[danny]

there is no duel... but name one PRO board under $5k. i dont care about any of your points otherwise.

There are lot's of bargains to be had out there. I have a Studer console with 24 inputs. It cost me $4k from a gear broker, originally $120,000 and looks and sounds like every penny (of the latter figure;). I spent about $1400 on customization and repair, so it's a little over $5k, including adding in some modules I have, such as a Neumann parametric. Many great boards by MCI, Soundcraft (TS series is great), D&R, etc, can be found in this price bracket. A friend just got a big D&R, like 32 inputs + returns in virtually unused mint condition for 3000 euros. With the whole world going ITB, analog prices have hit bottom. Neve & API didn't have a monopoly on good sound, do your research (http://recordingconsoles.net/consoles/consoles.htm might be helpful). So, thanks ITB guys! ;)

[zandurian]

Ditto about the great deals out there.
You have to think of an old mixer as a giant effects processor. The same way you would pass a singers voice through a tube mic to "enhance" the raw tone, or (this is an extreme example) a hammond organ through a leslie speaker, you can pass an entire band through some of those vintage pre's/eq's and it will do wonderful things.
I know it's ALL subjective but when I went from an MI level board to the MCI everything suddenly sounded big, fat, warm and punchy as opposed to (gulp, dare I say it?) cold and sterile. Really, it was night and day and though some clients thought it was strange to see this beat up old board in my "modern" studio, even the newbies could hear the difference. Admittedly, I was recording on early ADATs at the time (the MCI did a great job of smoothing out the early digital's harshness) but it's only gotten better as converters have improved.
The sound of some of the old boards really is "pro", and by that I mean the pre's are super fat. Definitely not totally accurate (or dead quiet) but they color the tone in (what most listeners consider to be) a very pleasing way (trying to stay politically correct here...) To me it's definately worth some maintanence expense.
This may seem to be a bit off topic, but I struggled for so many years trying to get a great sound happening (and I'm sure these modern plug-ins can do a lot) but when I did the vintage mixer thing it was like a magic wand - stuff would come back sounding GREAT with no effects at all and the rest of the process became much easier when starting from that vantage point.
Just my 2 cents, FWIW.
All that said, I DO appreciate these proposed tests and the quest for accuracy here because once I get that rich, perfectly distorted, fat, musical, goose-bump inducing performance captured I don't want anything messing it up down the line. Make sense?

[danny]

I don't agree that it's about adding distortion or about mic pre's, necessarily, nor do I think of a console as an FX processor. I think of it as a way to blend audio signals without molesting them ;)

Even if you've got just a stereo music mix and some vocals and blend them OTB and ITB you'll hear an obvious difference as to how big both elements can be made to sound. No shootout (especially one which we can't even read about) could convince me about ITB, as I get the chance to hear both ITB and OTB every single day (before I break out the tracks to seperate outputs). Do what you like, though!

I do a lot of automation in DP, and though I don't have total recall of my mixes, I do get paid to bring them back up so it's not hurting me any. OTOH, some people's work depends upon recall, and for those folks, it's either got to be an automated console or a summing box set at unity.

I can't afford Apogee or Lavry converters (which go up to about $8k for ONE AD or DA channel), but I don't pretend my MOTU is just as good. or spin elaborate, hostility-laced justifications defending that proposition. Pretending that mixing numbers sounds better than mixing voltages because one doesn't want to own a proper console or has insufficient experience with them amounts to the same thing.

[qo]

Here are two pics of DP's mixer with pairs of white/pink noise stereo tracks at different levels and pannings. All tracks are going to the Master fader and, as you can see, the Master fader in the first pic is not showing any signal. In the second pic, I've changed the level of track PO3 from +5.80 to +4.00 (PO3 is short for pink original 3, and is the second to last track). Here, the Master fader begins to show signal.

The 6 aqua tracks to the left of the (green) Master fader, labelled (WO1/WI1, WO2/WI2, WO3/WI3) are the white noise original/inverted tracks (the soundfiles were inverted using DSP-Quattro's Invert process). Likewise, the pink tracks to the right of the Master track are the pink noise tracks.

alphajerk, my feeling was that since white/pink noise contains all frequencies, just at different weightings, if the summing math has any frequency-dependent aspects, then summing all freqs would reveal any anamolies. Though, I can also see your point about trying normal program material as well and will do that later.

But, I wanted to post this initial result, which looks very good indeed. I'm going to try a bounce to disk and see what happens. Then import the bouce and zoom in to see if there are any deviations from the expected ruler-flat DC signal.

[qo]

You may have noticed in the first pic that the left channel of track PO3 is showing no signal. This appears to be a DP bug. After moving the fader, the track started showing signal. I experienced this twice during this test.

[Timeline]

Hey,

Might be better to use a storage scope and sona pulse.

This will give phase and transient at all frequencies.

Do it before a freeze and after canceled out of phase.

A kick is not all frequencies but better pink.

Just a thought.

[alphajerk]

qo, what IS the deal with PO3 on the top one? was it just the meters that failed to show signal on the left side? iow, could you hear signal soloing that track coming from the left? looks from the master fader on that despite the meter, the signals are cancelling evenly.... of course you need to look for low level noise.


danny, WHATEVER you decide to use to sum signals, you are molesting them in some sense or another... and recalling them isnt in the sense you are thinking of... its about a different way of working. its not about mixing a single song, printing it... moving on to another song. its about mixing 15 songs all at once, in order, varying order, throughout a album. cant do that on an analog board period [unless you are CLA and running mixing through your cookie cutter template]... but i have been to that site many times and used quite a number of analog boards for a quite a long time before i went ITB. i even remember using something called magnetic tape to record with.... before computers were even capable of recording audio and digital was still put on tape as well.... im not however old enough to remember a time when digital wasnt at least an option.

and honestly, i dont find [gasp] lavry to be very good convertors either, at least their idea of how the sound going in comes back out doesnt agree with my ears... nor does apogee really. it actually took me a while to figure out what and why i didnt like the lavry's. convertors are simply just another color one has to take into account when recording.

zandurian, your description of your mci sound is basically my statement before... take your adjectives and apply it to the source sound. why is it "big" [distortion], or "fat" [distortion], or the best, "warm" [distortion... especially the attenuation of higher frequencies, probably from teh old caps in teh board... hey i got an old bassman that sounds not a lot like a typical bassman and i refuse to make a single change to it cause i like the sound it DOES have].... if it works for you, GREAT! but it still seems silly to claim something else is "broken" because it doesnt sound the same as something else. is an API's summing broken because it doesnt sound like a Neve or even a sphere? which one is right and the others broken?

im just saying i think a lot of people take things too far overboard [ptp] with recording in the quest for some sonic utopia.... you mention recording on ADATs [bashfully] but i have heard a lot of nice recordings done on ADATs. its the indian, not the arrow.


back to qo, why not take a 8 track mix and make two mixes of it... even use automation and all in it.... take those 8 tracks and run them to the master fader and then on the right side of the master fader, duplicate those tracks and invert them with the trim plug and see what happens. although im not entirely sure what exactly what you are testing except phase cancellation.

have you heard the DAWsum CD? all DAWs seem to sound a little different which technically mathmatically they shouldnt. and fwiw, DP sounds more aggressive in the mids than logic or pt does which sounds more scooped, especially int he lower mids. the analog board sums just sound muddier than any of the DAWs.

[qo]

Hey Gary, what the heck is a storage scope and sona pulse? I'd try it if I had a clue, doh! heh.

alpha,

although im not entirely sure what exactly what you are testing except phase cancellation.

The intent was to eliminate DP's summing math as a candidate for why I felt the Ghost sounded better. If things cancelled perfectly (which they did, at first blush), then I was going to move into other territory to try to explain the difference. Next step for me (after BTD) is to add e.g. Eqium, non-bypassed, with a flat curve to several tracks. I'd expect that DP's PDC would result in everything lining up (temporaly) perfectly and that we should still have complete cancellation.

have you heard the DAWsum CD?

I have several of Fuston's CDs but didn't bother ordering the DAW CD since I'm of the belief that DAWs (being software), running on OSes (also software) undergo continuous change and make such a CD obsolete days after it's released. I'm sure it gives a nice snapshot of whatever versions he was using, and probably does serve to make a point about DAWs being different (or similar). I dig his hardware CD's (Ribbon Roundup, 3D Pre, 3D Mic), except for the suspicion that he's trying to convert it's listeners to Christianity with his choice of material (I don't need or want to be converted, sigh!).

[qo]

qo, what IS the deal with PO3 on the top one? was it just the meters that failed to show signal on the left side? iow, could you hear signal soloing that track coming from the left? looks from the master fader on that despite the meter, the signals are cancelling evenly.... of course you need to look for low level noise.

Opps, forgot about this. Yeah, this was definitely cosmetic. There was no signal hitting the DA converter so things were cancelling perfectly dispite the meter. Maybe we should start another thread entitled "Is DP's metering wack"

[qo]

Just tried inserting Eqium (flat EQ curve) on a track and still get cancellation. So, I tried modifying the curve and, as expected, the Master showed signal. I saved this preset (dragged it to the Desktop) and inserted it into the other track of the pair and still got complete cancellation. I'll probably try this procedure with other EQ plugs (Masterworks, Firium, etc). Maybe this is a decent way to test whether a given plugin is working with DP's PDC (at least for linear, non-time-based plugs)?

I then bounced this to disk. Below is the result (zoomed in all the way in DP's internal editor, check the scale on the left...):



If anyone is interested, I could create a disk image of this project and post it on my site. Might make for a nice way to regression-test new DP releases?

[qo]

So, what to make of this? Part of me agrees with danny/zandurian, that there IS an audible difference between ITB/OTB and that, personally, I think I prefer OTB. But, I really don't have enough experience with them side by side to know. Maybe over the next year or so, I'll gain that. Though, having been through this bit of testing, I agree with alphajerk that ITB is not adding/removing anything (at least nothing measurable, or at least nothing that I've been able to measure). So, in this respect, it does seem completely transparent. I do trust it more now...

I'm thinking perhaps a test using the Ghost, to see if the normal/inverted whitenoise waveforms cancel each other when sent to two channels on the Ghost that are zero'ed out. Maybe, as some of you have already hinted at, it's the Ghost that's adding distortion (not necessarily "distortion" in the literal sense, but some subtle non-linearity, aka voodoo aka mojo) and I just prefer the result?

But, regardless, OTB suffers from non-recallability (at least in the price bracket in which I live) which is a downside. Livable for some projects, but definitely not for others.

Finally, just for the record, I used imported 44.1/16 bit .wav files for the above tests. So, now we can wonder if the same results would be obtained with 88.2/24

[Splinter]

A little off topic, but with the advent of Tritone Digital's Colortone, I have toyed with the idea of instantiating it on each track and the master to add the color of a particular console, just to see how all the non-linearitiies would interact and if it would sound more "analog." This would possibly get us to the bottom of some things here, which I'm sure Qo will find - the color of the console is what he prefers more so than the actually summing. DPs apparent colorlessness lends no musicality to what is there unless told to do so, where the Ghost just sounds sweet. And boy does it ever. Some of my best mixes where done on the Ghost. Now I just need to figure out how to make DP sing the same way. I believe it can be done, but it's just different. Start with a good frontend (mic, pres, and A/Ds), get good levels (not TOO HOT), use quality plugs... all off these will get you there. It does rest on the engineer, but, if you're like me and you learned old school, you need to teach this dog some new tricks.

[iMAS]

So, what to make of this? Part of me agrees with danny/zandurian, that there IS an audible difference between ITB/OTB and that, personally, I think I prefer OTB.

I prefer OTB as well.

Maybe, as some of you have already hinted at, it's the Ghost that's adding distortion (not necessarily "distortion" in the literal sense, but some subtle non-linearity, aka voodoo aka mojo) and I just prefer the result?

Definetly.

But, regardless, OTB suffers from non-recallability (at least in the price bracket in which I live) which is a downside. Livable for some projects, but definitely not for others.

I know you already use the Ghost for summing, etc...but you may want to check out that new 16 ch. summing device --> Neve 8816, it has total "recall" via USB. I'm defintely saving up for this.

Btw, thanks for taking the time to do all these testing...very informative...much appreciated.