Is DP's summing wack?

[qo]

Hi toodamn. Right, I agree with you that this wouldn't be a fair test. Perhaps a better test might be to to find two engineers; one, a widely acknowledged ITB wiz, and the other likewise with OTB, give them an identical set of files, and let them use whatever they wanted in terms of "outboard" while mixing the files in DP (plugs for Mr/Ms ITB and hardware for the other wiz) and then compare the end results. If I alone were to post results, then there's certainly a good possibility that I just don't know what the heck I'm doing when working ITB and that would be unfair to MOTU.

Along those lines, does anyone know of any widely available commercial releases (not movie soundtracks) that were done entirely ITB in DP? I know Gary's (Timeline's) latest was done at least partially in DP, but am not sure if it's entirely ITB (Gary?). I'm just curious to hear what DP is capable of in the right hands.

[jr213]

Check out The Forever of Now by The Illustrated Band on iTunes. Recorded, mixed, and mastered entirely in DP.

[Newsles]

I dont know about the latest DP version, but, without making signals too hot (something Ive always tried to avoid mixing in the "virtual" environment, then the sum passing through my cheap Fostex AD/DA I find the results clean, spacious and well-balanced - odd considering some adamant comments in this thread about the summing issue.

Maybe it's just my ears.

And just for the record, I have an absolutely hump mixing board - a Behringer MX of all things!

[daveyboy]

Itunes: artist-Hunter Payne, CD- "Nailed"

All done in DP, mostly live players, little MIDI. Drummer from Warrant (my brother) for you 80's metal heads....

[alphajerk]

i find the summing to be fine in DP. i dont know about the -6db level rule... but i definately like tracking in the -18 to -9db peak window. for mixing, i think the more you get right on the way in the better the mixes sound with less processing inside, but i dont find much improvement running OTB to hardware and back in for insert processing to be much of an improvement over plugins either with the extra conversions. i think if one wanted to go that route, it would make more sense to just mix OTB all together. but then you dont have anywhere near the creative power with automation and so forth...

its all subjective anyway... what is better or worse. i tend to ignore all of that and ask if im happy working the way i work [well, apple isnt making me happy right now with their idiotic change to PCIe making my PCI card useless... and OSX/4.6 bloating my poor dualG4 to where it just doesnt work anymore to mix with... but thats another complaint that the only thing i hope i can do is save the session as a OS9 version and go back to OS9/3.11 to mix]

so to the comment about hotter levels making a mix muddier? you think you could say mix the peaks at -6dbFS, print it, raise the master fader +6db so its peaking at FS and the mix automatically becomes muddier? ill have to try that out and see if it makes a difference. i wonder what would happen if you just pullled the master down -6db on a full scale mix?

i would think that refraining from too much use of the floating bit might help within the mix... or maybe it could be beneficial to the sound as well when the rounding back to 24 occurs.

i kind of like doing everything in the mix, you can instantly recall anyway if it doesnt sound right when you get out to the car to listen, but i find if you want to compress/limit... its better to hear in the mix how the mix is going to affect the comp/limiter in real time than later on. of course you might not be going for that insane level squash, but even in that case... arent you mixing for "final" anyway and mastering shouldnt really squish it just to make it louder but simply bring up the level to its most decent level?

i dont know.

[Fibes]

The -6db rule applies to all tracks but like anything audio there are no rules just one law.

Mebbeh i'll see you over Xmas.

[stephen1212b]

Looking at the published specs of the a to d chips you can see that the distortion rises almost one for one for the last highest three db.

It is also true that under specific circumstances some of the reconstruction filters required to convert linear PCM into analog can create inter-sample peaks up to six db higher than simply looking at the bit data would indicate.

Add to this the unfortunate truth that the analog input of many studios doesn't have sufficient headroom to handle the nominal 2 volt output of many D to A converters and it is no wonder that keeping the level down would sound better.

There are other reasons that summing in the analog domain might be preferred, such as the randomization of the timing errors and the dithering effect of the analog noise floor prior to current summing.

If you do master in the box it is preferable to retain the internal 32 bit floating point or greater through to the final dithered 16 bit file. If you choose to use another program to "master" print 24 bit from DP and add dither in the other program when going to 16 bit.

Inserting a masterworks plug in the bottom position of the master fader and selecting the dither option even if the plug doesn't do anything else is a good practice. Do not adjust the master fader from unity or the benefit of the dither will be lost.

[hsten]

Check out what Bob Katz have to say on the PSW forum (his second reply on this page):

http://recforums.prosoundweb.com/index. ... #msg_64944

[chrispick]

Check out what Bob Katz have to say on the PSW forum (his second reply on this page):

http://recforums.prosoundweb.com/index. ... #msg_64944

Well, if it's good enough for Mr. Katz, it's good enough for me.

I guess it's back to the ever-plaguing question: Why do I suck so bad?

[iMAS]

"I just received one of the most beautiful, dimensional, pure, spacious mixes I've ever heard. How was it done? In the Box, mixed digitally in Digital Performer." --Bob Katz

[alphajerk]

Looking at the published specs of the a to d chips you can see that the distortion rises almost one for one for the last highest three db.

and distortion is always a bad thing? i guess it depends on the type of distortion... most people who prefer analog like it because its a distortive medium, i mean in the true sense of the word... not in the cranked half stack sense.

It is also true that under specific circumstances some of the reconstruction filters required to convert linear PCM into analog can create inter-sample peaks up to six db higher than simply looking at the bit data would indicate.

i have heard that as well... but again, what are the consequences and what type of material is causing it? transient material suffers far less negative impact.

Add to this the unfortunate truth that the analog input of many studios doesn't have sufficient headroom to handle the nominal 2 volt output of many D to A converters and it is no wonder that keeping the level down would sound better.

There are other reasons that summing in the analog domain might be preferred, such as the randomization of the timing errors and the dithering effect of the analog noise floor prior to current summing.

there are many reasons but that example can also lead to a heavy accumulation of noise inherent in the analog summed tracks.... which IMO are far worse than any deleterious effects of summing within the box, where randomization doesnt occur and the only noise added together is inherent in the source files. i mean analog has as many [if not more] downsides due to the electrical components the signal travels through.

If you do master in the box it is preferable to retain the internal 32 bit floating point or greater through to the final dithered 16 bit file. If you choose to use another program to "master" print 24 bit from DP and add dither in the other program when going to 16 bit.

Inserting a masterworks plug in the bottom position of the master fader and selecting the dither option even if the plug doesn't do anything else is a good practice. Do not adjust the master fader from unity or the benefit of the dither will be lost.

id rather stay all the way @ 24 bit to the end of the print and deal with any src/bit later when assembling the songs retaining a "higher-rez" version of the mix.



the -6db "rule" is merely associated with the intersample example mentioned above.... plus the perception of human hearing only being capable of processing 20bits of data [at most]... and most convertors and analog gear only having 120db of dynamic range anyway [20bits worth] so using those last 4 bits only risks more error than one would gain [ptp]. if we are talking on the way in, unless we are recording some serious symphonic stuff, we deal with a far less dynamic window... a cranked half stack for example really only needs 3 bits or so to be captured.

when mixing, anytime you venture into the floating point bit land you risk distortion from rounding errors that are going to occur at the end of the line bringing it back down to 24 bits [or less if you added like the mwlimiter to 16bit] fortunately that error/distortion is exact and repeatable so if its to your liking, its going to sound the same everytime [which for discussions like this about something sucking but trying to do it the same way evertime is the simple term of insanity]... but if it floats your boat [again ptp]... then break that rule.


im just trying to point out there are inherent problems with ANY system you use but most are operator errors poorly learning how to use it within its "limits" and then knowing how to push it past those limits in a constructive manner. after all, reality is what you can get away with.

[alphajerk]

Mebbeh i'll see you over Xmas.

yeah... seems like its been years now.

[stephen1212b]

I mentioned 32 bit float because that is the engine that DP is based on. An argument can be made that 48 bit fixed point double precision would be better.

All I was really suggesting is that the fewer times you modify the data and print to 24 bit the better as each time some rounding errors are likely to occur and these errors are interactive with further processing and therefore more cumulative than you might expect. Careful monitoring of levels at every stage and leaving some headroom for effects processing and summing just as you would on an analog board will provide the best results.

[danny]

there are many reasons but that example can also lead to a heavy accumulation of noise inherent in the analog summed tracks.... which IMO are far worse than any deleterious effects of summing within the box, where randomization doesnt occur and the only noise added together is inherent in the source files. i mean analog has as many [if not more] downsides due to the electrical components the signal travels through.

heavy accumulation of noise? not with any decent analog system. let's leave mackie and behringer out of this., eh? ;) The only time i ever experience analog noise is when I patch in some older equipment (ie: eventide instant flanger WHOOSH). my console is very quiet, as are all pro desks. ITB guys who try OTB do tend to change their tune. For under $5k you can get a fantastic pro analog console used these days. Besides the sound, it's a hell of a lot of fun, you never deal with latency when recording, you tax your computer less, and you open a world of devices you can easily interface with.

Sure, you can do some fine mixes ITB, but the playing field is constricted in terms of what you can get away with, imo. OTOH, if it's a very sparse mix with all beautifully recorded elements, it'll tend to sound great either way. I mixed an African record both ways, and the dense midrange energy from all those percussion and vocal elements really taxed the digital system's resolution...it was a huge relief when i was permitted to remix OTB.

[Nama]

If my mix is foggy, it••™d sounds foggy anywhere, but you are saying the Ghost fixed this? Then, I believe you.

I just personally don't have any experience that my mix started sounding foggy after this, or after that. I like to track good signal, and I'm really fortunate that I don't work with bad players. If they play good, and I record them OK, everything sounds good, and it leads to a good mix, eventually.

I think gear is just a master of money, convenience, space, looks etc.
I "personally" never find DP's summing sound good or bad.