Duh, I meant tracks in 24/96, host based processing .... Sorry about that !

Paul.

Paul, it is possible, but 40 tracks with plugins can be asking a lot of host-based - I really doubt even a 1.2GHz PC could do 40 tracks of 24/96 - 24/48 yes, but I doubt 96k. Hard disk streaming is another consideration - this is the equivalent of 80 tracks at 24/48, which, AFAIK, only Paris can do on PC and Mac equally, and PT on Mac - yes, PT is PC also, but I haven't heard good recommendations on the NT version - just check it first out if you go that route. Both Paris and PT are hardware hybrid to help processing. Paris provides 4-bands of excellent EQ per channel, dedicated, plus comps/gates/etc. for most every track (every track if you don't use any of the built-in reverbs, etc). Paris doesn't support 96k yet, and I don't know about PT. For what you are trying to do these are the two best options, and really Paris if you want to stay PC. Nuendo is a host-based option, but probably not with that many tracks at 96k. For 24/48 or 44.1 your options are more open.

Slight correction - since you want live inputs - Paris can do this with comp/EQ on each input, as I believe PT can also. 96k is still in question either way. Some host-based programs now support live inputs with processing (Logic for one), but for 32-40 tracks, you would be better to go with a system like Paris or PT.

Cedric,
thanks a lot for your info. I already was afraid that 40 tracks at 24/96 including all the goodies (baddies ?) is asking for too much. I have this slight feeling that 24/48 is not the way to go, if you want to compete with good quality analog mixing desks.
You see, I'm thinking about upgrading sooner or later from my current analog desk, which is of humble origins. Automation would be nice since I think my music will benefit from it.

16/48 didn't cut it and 24/48 looks like an intermediate solution to me. I hear far to many cd's that sound like shit, you see.
Upgrading more later than sooner looks like a possible solution to me, as processing power is increasing almost on a monthly base now and is getting relatively cheaper.

EQ's and compressors and gates YES, reverbs NO, they will surely eat up too much processing power. Outboard reverbs is the way to go, IMO. That's one of the reasons I mentioned 40 inputs, you have to count in the sends / returns to and from these outboards .... I'm doing no more than 24 tracks at once usually, say max. 8 FX returns, and a little extra for expandability, that's why I mentioned 40 tracks.

Anyone else who wants to share some insights ?

TIA,

Paul.

[ March 20, 2001: Message edited by: Paul Jenner ]

You say EQ and Noise Gate on every track--and it sounds like you already do some recording. How do you do the noise gating and EQ right now? Have you considered tracking through your outboard EQ & noise gates?

Also, although this might not work for your music production, most sequencers will allow you to group channels--so, you may be able to get away with making a few groups and running batches of tracks through them to take a load off your CPU.

Any decent system right now would be able to handle 40 tracks running at the same time, with automation.

Would it waste too much time to track your 24 channels, then mixdown your EQ/effects before continuing with more tracks? Or do you tend to do a lot of post-tracking tweaking with the settings to make your cornucopia of channels mesh together well?

--Gabriel

Paul,

Other than 96k, Paris fits the bill with 8 Aux sends/returns, expandable I/O, and I am sure eventually 96k/surround sound too. There are some standalone hardware units (Mackie, Fostex, Akai, etc) that say they support 96k, but as far as I know, true 96k converters are expensive - Mackie says 96k I/O cards are coming for the HDR24/96, but it will only do 12 tracks at 96k. Similar limitations will exist for any HD based system. The converters are the real challenge, as is getting your beautiful 24/96 mix down to 16/44.1 for CDs, unless you plan to stay DVD audio all the way and hope consumers follow suit. Many of the bad sounding CDs out there can be attributed to other factors than sampling rate and 24-bit. Good engineers and get some nice sounding mixes into 16/44.1. Bad engineers can make 24/48 sound like crap. 96k is great, but marketing is pushing it faster than pro audio engineers and consumers. Just my opinion - it will be here en masse eventually, but for now is more of an investment than a PC DAW. I would suggest looking into a system that is modular enought to allow you to move up to 96k, but provides high sound quality now and the mixing features you are looking for. There is more to a good mix than the sampling rate/bit depth - not to diminish the advantages of 96k, but you will get more going from 16 bit to 24 than you will from 48k to 96k today. Next year, or the year after that will start to change with faster HDs and CPUs.

Hello Gabriel, thanks for the reply.

I now record directly to two-track stereo via my 24/8/2 desk. EQ on all channels off course, just one or two seperate noise gates.

I.m *NOT* looking for multitrack recording on a pc, just mixing (shall we say "live feeds") inside the pc. Then HD recording it to stereo off course, maybe 5.1 in the future.

I have had a quick look at Nuendo, it seems like everything I need is inside there. Haven't done extensive listening tests tho.

I just wondered if, given the current state of affairs, a pc could handle *mixing* those 40 tracks with all the goodies in 24/96. If this appears to be way out of reach for a pc, then why bother with a digital system, I just might be better off with a better analog board.

I do not quite undestand what you mean by busing a few tracks. Buses are nice, sure, but only after individual EQ-ing ...

Thanks,

Paul. :confused:

Hello Dedric, sorry about that "C" last time ...

I still have my doubts about spending big bucks (for someone with my budget, yes, it's BIG bucks) for "only" 24/48.
Some of the higher end EQ plugins upsample to 96 k first, then perform the EQ calculations, then downsample back to 48 k. There must be a reason for that. Some well known engineers indeed have reported that these plugins DO sound better, despite all the up / down juggling going on. Sure, there is more to a digital equalizer than the sample freq they work on, algo's for instance, I know.
Maybe to you guys I come across a little blase talking about this 24/96 thing all the time when I'm in fact still a humble home recordist, but if and when I'm ready to upgrade the homestudio, I want something I will be happy with for longer than the first few weeks. As good converters come standard with this 96k sampling freq and software like Nuendo can handle it ..., you catch my drift ?

Again, taking my time to make this decision might be best, like I said. In a year or so our precious little discussion may well be obsolete ...

Good thing to discuss such a topic on this forum anyways, as "host-based 24/96" seems to be the thing we are all heading for in the not so far future :) Thanks for chiming in !

Paul. :)

Paul,

I think it is wise to plan ahead. What that means really depends on your goals and aspirations, and you do seem to have your eye on 24/96. There are some options now, with varying degrees of flexibility for the future. Nuendo with an RME Hammerfall or Dig96 card would seem like the best of the host-based options - although the analog I/O is still 48k. For midrange - the standalone DAWs I mentioned earlier support 96k in software with 96k I/O options coming down the road. That seems like a bit of a risk given the closed platform. Anything else is getting into a more expensive system up front. Hardware I/O will come for all of the systems eventually and software support is there now with most everything from Cool Edit Pro to high end standalone DAWs. Even if you started with Nuendo and an RME card, your investment isn't monstrous, so if you grow out of it even once 96k support comes along, you aren't too far in to change course. You are right though, this discussion may well be moot point in a year.

The audio community "debate" over 96k isn't really whether it is theoretically better (math and Nyquist theorem tell us that on paper even without hearing it) but rather, is it worth the tradeoffs now in a consumer market that may well force broad acceptance of 96k to wait another 5 years or so? It comes down to investment justification.

Hello Dedric,

Even if a consumer format at 96k is not going to hit it big time as the next standard, 96k has it's advantages -as I see it- even when transferring to 44.1/16 for the end product.
I have my doubts about whether DVD will be used for enhanced sound quality. With the ever decreasing quality of the average domestic playback system, mp3's and all, the tendency of the record buying public to look upon music as a very disposable product, who is going to hear / appreciate the difference ?
My idea is that the traditional analog master tapes have always been at a level say 5 times the quality and resolution of the end product, vinyl or cd. Now, when moving towards digital format master "tapes", I see no point in changing that attidude: I'd say, why not keep your digi master / intermediate stages at some 5 times better quality, before squeezing it onto a consumer format, like it's always been ? After all, working with good quality gear is always easier than working with so-so quality gear (don't ask me how I know that for sure :) )

Regards,

Paul.

Hi Paul,

If you really want to go the 96kHz way right now, there are other products on the market doing the job with a lot of comfort.
The pyramix-workstation from Merging Technologies for ex. is already running up to 384kHz for realtime DSD-to-PCM conversion - I think it's one of the first worldwide doing this.
I'm myself working more and more with 96kHz; the DSP-drain on the mykerinos cards is bigger, of course, but the last software version will allow you to link up to 4 boards together - giving you, by the way, the total of 64 physical ins & outs in adat format.
EQ's and dynamics by Daniel Weiss, some more great stuff by Vincent Burel (especially his reverb Aphro V1 & a killer multiband-comp called C10), surround up to 7.1, DirectX-support (some of the Waves-plugs run very nice at 96kHz), on-board SMPTE, VITC, LTC & Wordclock reader & generator, and a very powerfull & intuitive editor.
I'm running a dual-card setup on W2000 and a fast SCSI-controller - smooth like sugar.
The stuff is of course quite more expensive then Nuendo & a Hammerfall card, but not so overpriced like a big protools-rig.

Cheers
Benoit

Hello Benoit.

>If you really want to go the 96kHz way right now, there are other products on the market doing the job with a lot of comfort.
The pyramix-workstation from Merging Technologies for ex. is already running up to 384kHz for realtime DSD-to-PCM conversion<

Would you care to elaborate a little more on this ? I don't have the slightest idea what you're talking about :) Is this "pyramix" system host based ? :confused:

> ... giving you, by the way, the total of 64 physical ins & outs in adat format.<

Ha, but you need two ADAT style inputs to work in 96k, using bit splitting techniques, right ? If so, is this cumbersome in any way ?

> ...(some of the Waves-plugs run very nice at 96kHz)<

Aha ! Like for instance the equalizers or what ? I'm very curious to learn more about what works clearly audible better in 96k !

>The stuff is of course quite more expensive then Nuendo & a Hammerfall card, but not so overpriced like a big protools-rig.<

The latter is a good thing. I must add that, although Hammerfalls are *good* (I have one), a *great* system needs the better outboard converters, like the RME ADI's ?

Thanks for sharing your info so far, Benoit.

Paul.

Hi Paul,

Sorry for not having been more specific...
Pyramix is a DSP-based system running under W2000 or NT4, made by a small swiss company called Merging Technologies (http://www.merging.com), hold by some ex-Studer/Nagra-Kudelski engineers. They don't seem to be well-known in the US... ;-)

"... giving you, by the way, the total of 64 physical ins & outs in adat format.<

Ha, but you need two ADAT style inputs to work in 96k, using bit splitting techniques, right ? If so, is this cumbersome in any way ?"

You're absolutely right: you'll have to split those 64 inputs into 32 for 96kHz operation. If you need more then this, I know that the soft will support up to 8 boards in a next release (8x16 ins/outs at 44.1/48kHz, the half at 96); there is also a MADI-version of it, offering 64 ins/outs pro board.

"Aha ! Like for instance the equalizers or what ? I'm very curious to learn more about what works clearly audible better in 96k !"

To me, the biggest audible advantages in working at 96kHz are in fact the eq's. Thanks to the Nyquist theorem, the filters sound much more sweet and analog (especially in high frequencies) running at 96kHz. I don't even talk about the same plugs roaring at 192kHz! ;-)

"The latter is a good thing. I must add that, although Hammerfalls are *good* (I have one), a *great* system needs the better outboard converters, like the RME ADI's ?"

Oh yes! I have a Hammerfall running with Nuendo too - definitely the best in this price-range.
I'm using 2 Swissonic and 2 ADI ADC's for the moment; I've been trying some super-expensive stuff like apogee's or the merging converters called "Sphynx" - I'd say that the price gap is not proportional to the sound-quality difference - but there is one, that's clear.

A DAW by itself has a sound of its own - not as audible as with analog gear -, but the plugs used are making a big difference. A lot of the grand-public VST & DX stuff doesn't sound that good to my ears...

What are you going to hook up to your DAW? A 2" analog machine?

Cheers

Benoit

[ March 22, 2001: Message edited by: Benoit ]

Originally posted by Paul Jenner:
I now record directly to two-track stereo via my 24/8/2 desk. EQ on all channels off course, just one or two seperate noise gates.
I.m *NOT* looking for multitrack recording on a pc, just mixing (shall we say "live feeds") inside the pc. Then HD recording it to stereo off course, maybe 5.1 in the future.

I could be wrong here (not thinking of all the possibilities) but your method is:
instruments->current recorder, then recorder->mixer->stereo tracker.

To mix within the computer, I believe you would have to transfer your files to digital:
instruments->recorder, then recorder->computer (aka mixer & stereo tracker).

Correct? It sounds like you're looking at PC solutions (as opposed to Mac) which I'm not as familiar with the specifics, but it is entirely possible to mix within the computer that way. You would most likely end up having to transfer your files to the computer (essentially, multitracking them, but you would be multitracking from your current recording system as opposed to multitracking from live musicians), and then mixing down once you got everything in the box.

There's another thread going on, over in the Pro Audio section of these forums I think, about the pros/cons of using your DAW as a mixer.

The main problem with mixing inside a PC will be: once you load in 40 tracks and EQ them all individually, and add in other filters (such as noise gating), real-time playback will produce stutters. However, once you click on "export/mixdown audio," your computer will have no inherent problems throwing down any amount of tracks, with any amount of effects on them, without complications.

I just wondered if, given the current state of affairs, a pc could handle *mixing* those 40 tracks with all the goodies in 24/96. If this appears to be way out of reach for a pc, then why bother with a digital system, I just might be better off with a better analog board.
I do not quite undestand what you mean by busing a few tracks. Buses are nice, sure, but only after individual EQ-ing ...

Yeah, that pretty much answers what I was thinking. If it's reverb, you could make 2-3 reverb patches and run your channels through there, but that's not something that yields good results when EQing.

I think, to paraphrase simply:

A PC will not effectively replace your mixing board. It will offer you a lot of extra possibilities, but the technology is geared towards: input all your tracks->work on your near-unlimited tracks inside the computer, mixing, effects, etc->export two tracks or burn a CD.

With more than, maybe, 16 tracks of 24/96k you're going to have problems with real-time playback.

Does that help clear anything up?

--Gabriel!

Hello Benoit, this is great info, thanks !

>Pyramix is a DSP-based system running under W2000 or NT4, made by a small swiss company called Merging Technologies (http://www.merging.com), hold by some ex-Studer/Nagra-Kudelski engineers. They don't seem to be well-known in the US... ;-)<

That is clear, I have never heard people talk about it on RAP ...

> ... you'll have to split those 64 inputs into 32 for 96kHz operation. If you need more then this, I know that the soft will support up to 8 boards in a next release (8x16 ins/outs at 44.1/48kHz, the half at 96); there is also a MADI-version of it, offering 64 ins/outs pro board. <

Good.

>To me, the biggest audible advantages in working at 96kHz are in fact the eq's. Thanks to the Nyquist theorem, the filters sound much more sweet and analog (especially in high frequencies) running at 96kHz. I don't even talk about the same plugs roaring at 192kHz! ;-)<

YESSS !, I knew it ! Ahum, in fact, I knew nothing but came to expect something along these lines as mentioned in my previous post.
I don't quite get what old Mr. Nyquist has to do with it, but I do understand the part about "give those bit crunching algorythms enough meat to sink their teeth into". There seems to be a general consensus about the "more bits" part. I'd say: throw in some extra kHz to make the algo's happy who perform their very magic in the frequency domain in the first place ! That's more or less why I mentioned 24/48 to be an intermediate step (see previous post), not because I'm a spoiled rich daddy's kid ...

> Oh yes! I have a Hammerfall running with Nuendo too - definitely the best in this price-range.
I'm using 2 Swissonic and 2 ADI ADC's for the moment; I've been trying some super-expensive stuff like apogee's or the merging converters called "Sphynx" - I'd say that the price gap is not proportional to the sound-quality difference - but there is one, that's clear.<

Very nice, Swissonic and RME ADI seem to be competitors in the same price range, right ? Now we are all here in this cosy RO place, if you can spare the time, any comments on the Swiss vs. RME ? Both are getting good press, so I've been told.
Apogees *should* be clearly audible better, as they cost about 4 times an RME. I've heard more rumours about the RME breathing down Apogee's neck. Let's leave out the Apogees. They are WAAAAYY to expensive for me and I doubt if my music would benefit from them (more on this below ...). But, if anyone wants to give more details about why Apogees are King, please do so !

>A DAW by itself has a sound of its own - not as audible as with analog gear -, but the plugs used are making a big difference. A lot of the grand-public VST & DX stuff doesn't sound that good to my ears...<

Yeah, I know: sad story. Like 1 + 1 = 2 in every country of the world, this doesn't seem to be the case in DAW land ... I'm told they all do the math differently. This stinks ! (Who farted in the console ? This mix stinks, you know ....) ;)
But, may I presume the bad VST mixes you hear have more to do with the people operating those systems than with the actual piece of software ?

>What are you going to hook up to your DAW? A 2" analog machine?<

Ha ha, get ready for a big disapointment Benoit !
Hell no, the only analog reel to reel I own is a humble Revox B77 mKII (in a like new condition I must add).
No, my music is solely made with samplers and synthesizers, controlled by Cubase on an Atari, to have them play all together, hence my "live feed 40 channels" - concern. Why all the fuss then about striving to have an excellent mixing console one day, you might ask ?
Well, let me try to explain. Imagine an expert acoustic guitar player, gently plucking the strings, thoughtfully mic-ed with a Neumann, top notch pre-amps and all, this gives us an increadable amount of subtlety we have to preserve during the recording and mixing process, right ?
Now, say a lead sound from an analog synth is never going to be that subtle, I know that, you know that. However, in my experience synth recordings need a godawfull amount of processing in order to sound right. A lot of EQ, a lot of dynamic controlling, to name a few, going on at my place. And gating would be nice. Those Roland analog chorusses do sound fantastic, but are also very noisy, you see ?

Why is that ? Just a theory:
With mic-ed acoustic recordings a skilled tracking engineer (not me !) can fiddle around with pick up patterns, mic selection and mic-placement to his heart's content. If done properly, these tracks virtually "mix themselves" (well, it is a catchy phrase, isn't it ?). Not a lot of mic placement going on when recording synthesizers, ay ?
Also, say, a bass guitar: Left to right, front to back, top to bottom, bass guitars are designed to crank out one thing and one thing only: bass notes.
Not so with a synthesizer. Although synthesists like me have their own preferences about which synth to use when time comes to program bass sounds (Mini Moog, Roland Juno etc.). Although these puppies are known as the "kings of the lower octaves", they are what they are: General purpose instruments / tone generators. If you're not processing carefully, Roland Juno bass sounds turn out to resemble a Juno strings sound a lot. Not very nice when mixing, I can tell you !

Still not fallen asleep, guys ? I'm glad we do not have a moderator yet ... Apologies about the long posts, Benoit.

That's more or less why I am concerned about the sonic quality that processing in a DAW will give me, even if it's "just" sampler / synth based music. O.K., I admit it, I can be a perfectionist at times. My current setup (and my chops !) will not allow me to crank out killer sounding tracks at the moment. However, it's getting better all the time.

Paul.

Hello Gabriel.

(Look at my reply to Benoit's post, I just forwarded it.)

>I could be wrong here (not thinking of all the possibilities) but your method is:
instruments->current recorder, then recorder->mixer->stereo tracker.<

No, sequencing all tracks so they play together, thus "live" feeds to the computer, which will only be used as an automated console (and writing / burning the two track afterwards, but that's the easy part ...). Not a lot of stress on my hard-drive though, just looking for how much stress is too much for a Pentium / Athlon processor :).

>There's another thread going on, over in the Pro Audio section of these forums I think, about the pros/cons of using your DAW as a mixer. <

Thanks, I'll take a look. These threads belong in this forum, don't they !

>The main problem with mixing inside a PC will be: once you load in 40 tracks and EQ them all individually, and add in other filters (such as noise gating), real-time playback will produce stutters.<

I already was afraid that this would be the case given the current state of affairs in pc world, with these amount of tracks and processing going on ...

>However, once you click on "export/mixdown audio," your computer will have no inherent problems throwing down any amount of tracks, with any amount of effects on them, without complications. <

I like to EQ etc. in realtime, balancing the eq'ed track against the others, don't you ?

>Yeah, that pretty much answers what I was thinking. If it's reverb, you could make 2-3 reverb patches and run your channels through there, but that's not something that yields good results when EQing.<

I spoke too soon :) !

>I think, to paraphrase simply: .....
A PC will not effectively replace your mixing board. It will offer you a lot of extra possibilities, but the technology is geared towards: input all your tracks->work on your near-unlimited tracks inside the computer, mixing, effects, etc->export two tracks or burn a CD.
With more than, maybe, 16 tracks of 24/96k you're going to have problems with real-time playback.
Does that help clear anything up?<

Yes, that was my main concern. Will your story be different after you've read how I work (see above, sequencing and all) ?

I thought it would be wise to ask over here, cos' Nuendo-ish dealers will only tell " this software package does 200 tracks, no problemo !" :(

Anyway, the more we talk about it, the more I (and maybe others) will learn. This looks like the right place to ask such a question, much more peacefull than RAP.

Thanks Gabriel,

Paul.

Hi Paul,

RME vs. Swissonic: sonically, we're talking about half-hairs. Not really relevant. Handling-wise: I'll give a plus for RME - the box has just more usable options (like TDIF connections & transfer between ADAT & TDIF-format for ex.). The Swissonic is very simple to use. The thing that bugs me with the ADI are the symmetrical jack-connections, which are far too close from each other.
It might be interesting to test both of them in a wordclock-slave situation, and see how good the PLL's are! ;-)
They both have good clocks - far better then the ones of an Adat, for ex. (which is running around 500ppm out of specs at 44.1).
I had some problems with the resistances put on the TOS-link outs of the Swissonic - they were also out of specs, I had to change them. I think that the guys corrected that since.

"But, may I presume the bad VST mixes you hear have more to do with the people operating those systems than with the actual piece of software ?"

Sure, but...I did bad VST mixes too! ;-))
I started with analog gear, and I guess I had to change my habits quite a bit...I wouldn't go back anyway, even if I still like to track stuff on a 2", transfer it into a DAW, and sometimes mix it on a big board.
And I know some developpers who work hard at making _better_ sounding plugs! ;-)

"And gating would be nice".
Sure! ;-) But listen: I'm recording mainly accoustic music - jazz ensembles, classical music, contemporary, salsa, rock, you name it.
Over the years, I tended to use less and less gating, but tried to use the spill in order to make the mix more alive. If I run into problems (like noise & hum), I'll usually try to use expanders instead of gates - sounds smoother to me - , or even better, to edit the tracks in my DAW - what is nearly as fast, and most of the time much more musical.

"controlled by Cubase on an Atari"
Yesss! I've been doing so many things with this duo...actually, those boxes had a charm that I miss with the actual machines.

"Not a lot of mic placement going on when recording synthesizers, ay ?"
...mmm...if you have a good room, you could try to put a good speaker system in it, feed it with some of your synths, mike the place, and record it back...I do this quite a lot with sequenced drums - it brings a new kind of life in it! :-)

">The main problem with mixing inside a PC will be: once you load in 40 tracks and EQ them all individually, and add in other filters (such as noise gating), real-time playback will produce stutters.<

I already was afraid that this would be the case given the current state of affairs in pc world, with these amount of tracks and
processing going on ..."

Well, THAT's the reason why I love working with DSP-solutions. First, they operate much more smoother then native systems; and then, you might put some more cards in your system as soon as you start getting short with processing power. I'm not talking about the far better stability and latency behaviour...

"With more than, maybe, 16 tracks of 24/96k you're going to have problems with real-time playback."
With a native system, that's for sure! :-)

Cheers
Benoit

Hi Benoit,

>RME vs. Swissonic: sonically, we're talking about half-hairs. Not really relevant. Handling-wise: I'll give a plus for RME - the box has just more usable options (like TDIF connections & transfer between ADAT & TDIF-format for ex.). The Swissonic is very simple to use. The thing that bugs me with the ADI are the symmetrical jack-connections, which are far too close from each other.<

Channel separation issues or just ergonomics ?

>It might be interesting to test both of them in a wordclock-slave situation, and see how good the PLL's are! ;-)
They both have good clocks - far better then the ones of an Adat, for ex. (which is running around 500ppm out of specs at 44.1).<

RME claims to have state of the art clocks and PLL circuits. Would clocking from, say, an Aardvark be beneficially, with regards to the added expense ?
I've heard about the ADAT sync problems. After a few tracks they are reported to be significantly out of syn ... I think ADAT's days are over now. Alesis sure made a helluva lot of money out of it, no need to look back, though ...

>I had some problems with the resistances put on the TOS-link outs of the Swissonic - they were also out of specs, I had to change them. I think that the guys corrected that since.<

So you do a little soldering yourself then ? Ever felt like upgrading the RME's analog front end, like opamps, power supply and such ?

>Sure, but...I did bad VST mixes too! ;-))<

Happens to me all the time, but I'm getting better at it !

>I started with analog gear, and I guess I had to change my habits quite a bit...I wouldn't go back anyway, even if I still like to track stuff on a 2", transfer it into a DAW, and sometimes mix it on a big board.
And I know some developpers who work hard at making _better_ sounding plugs! ;-) <

Is this a convenience issue, or are you really satisfied with the sonic results you get from your DSP stuffed DAW these days ?
What exactly do these DSP's actually do ? The maths like the Pentium would do on a host based system ? You have to rely on the proprietary algorythms then. Are these any good, compared with say, Nuendo or Samplitude plugins ? Are there any third party plug-in developers for your system ?

>"And gating would be nice".
Sure! ;-) But listen: I'm recording mainly accoustic music - jazz ensembles, classical music, contemporary, salsa, rock, you name it.
Over the years, I tended to use less and less gating, but tried to use the spill in order to make the mix more alive. If I run into problems (like noise & hum), I'll usually try to use expanders instead of gates - sounds smoother to me - , or even better, to edit the tracks in my DAW - what is nearly as fast, and most of the time much more musical.<

I think we can agree about this. In some genres gating sucks the life out of your mixes. But for my applications, no room sound and all ....

>"controlled by Cubase on an Atari"
Yesss! I've been doing so many things with this duo...actually, those boxes had a charm that I miss with the actual machines.<

I see no point in sequencing on an other platform. Atari is fine, timing is rock solid, never had a single crash in three years of use ! Why change a winning team ?

>"Not a lot of mic placement going on when recording synthesizers, ay ?"
...mmm...if you have a good room, you could try to put a good speaker system in it, feed it with some of your synths, mike the place, and record it back...I do this quite a lot with sequenced drums - it brings a new kind of life in it! :-)<

I knew you were going to say this :) For the apps you mentioned, it seems like a good approach. I have to try it myself sometime, but at the moment, I don't feel the urge.

>Well, THAT's the reason why I love working with DSP-solutions. First, they operate much more smoother then native systems; and then, you might put some more cards in your system as soon as you start getting short with processing power. I'm not talking about the far better stability and latency behaviour... <

As was to be expected ... BTW, I think you mean smoother etc. than *Host* systems ?
What do these cards actually cost and how many would I need for 40 channels ? How do you get 24/96 from an RME into these cards, what kind of connector / interface ?

Just curious Benoit, where are you located ?

Paul.

Particualrly at the moment and for home studios, 96kHz is a bit of a red herring, little more than marketing hype. There are definately advantages to 96kHz but nothing that would be particularly noticable in the average home studio. A more noticable improvement would be to stay at 44.1/48kHz and add some quality outboard EQ, preamps, clock source, reverb, etc.

For all practical purposes 96kHz is the future, but it's not the present. IMHO, the current batch of 96kHz systems are at the lower end of the market and cannot compete with the high end 44.1/48kHz professional systems.

Greg

>Particualrly at the moment and for home studios, 96kHz is a bit of a red herring, little more than marketing hype. There are definately advantages to 96kHz but nothing that would be particularly noticable in the average home studio. A more noticable improvement would be to stay at 44.1/48kHz and add some quality outboard EQ, preamps, clock source, reverb, etc.<

I think we can agree on that one to a certain degree, Greg. 24/96 on the spec sheet alone says nothing. But Benoit and I have been talking about RME converters, which seem to be above entry level, sonically, are reported to have a stable clock, clocks between RME's can be slaved / mastered and there is always the option of adding an Aardvark or something. Converters without a good clock are useless, I'm pretty much aware of that ...

>For all practical purposes 96kHz is the future, but it's not the present. IMHO, the current batch of 96kHz systems are at the lower end of the market and cannot compete with the high end 44.1/48kHz professional systems.<

Would you subscribe RME ADI outboards as as lower end ? I'm curious to learn about your opinion. Main concern is to give the algo's enough meat to sink their teeth into, like I mentioned in previous posts, especially when EQ's are concerned. I couldn't care less about the expanded bandwidth.

Thanks Greg,

Paul.

Hi Paul,

"Channel separation issues or just ergonomics ?"

...ergonomics.

"I think ADAT's days are over now. Alesis sure made a helluva lot of money out of it, no need to look back, though ..."

...mmm...I've got 3 XT's and a BRC laying around for months here...somebody interested? ;-)

"Is this a convenience issue, or are you really satisfied with the sonic results you get from your DSP stuffed DAW these days ?"

...convenience/speed first. I'd say that the "louder" the music, the more difficult it gets to mix only in a DAW . I've been working many years on an ADT, a SSL G4000+ and a Neve Legend, and know what comes out of such boards. Digital only is getting close, but not quite there...on the other hand, I'd never start recording classical stuff with analog gear again! For jazz: well working too. When I do rock stuff, I patch my DAW in the board for summing. Best of both worlds! :-)

"What exactly do these DSP's actually do ? The maths like the Pentium would do on a host based system ?"
...basically yes - the big difference is the efficiency of modern DSP-cards. For ex, a Philipps TriMedia chip running at 150MHz is as powerfull as a PIII 500 for audio applications. Now put up to 8 of those chips in a DAW! I'm sure you see the picture...

"You have to rely on the proprietary algorythms then. Are these any good, compared with say, Nuendo or Samplitude plugins ?"
...oh yeah! The eq's and the dynamics were developped by Daniel Weiss, one of the gurus in the world of digital audio.

"Are there any third party plug-in
developers for your system ?"
...yes: the great developper Vincent Burel is porting a lot of his stuff on pyramix. And: any DirectX plugin will run with the system - not on the DSP-cards, but on the host-cpu.

"What do these cards actually cost and how many would I need for 40 channels ?"
...one card with adat interface is around 3200.- swiss francs...and I guess you'll need 2 or 3 of them for your purpose (you could sync your atari to pyramix, and record your tracks in 2 shots or more). I know, quite a lot of money, but count what you'd have to spend for an similar protools setup - which, btw, can't run at 96kHz at this time.

"How do you get 24/96 from an RME into these cards, what kind of connector / interface ?"
...adat TOS-link splitted. You could go with MADI or AES-EBU as well by using other ADC's.
But hey, just have a look at http://www.merging.com ;-)

"Just curious Benoit, where are you located ?"

...switzerland. Yourself?

Cheers
Benoit

[ March 23, 2001: Message edited by: Benoit ]

Hi,

>...mmm...I've got 3 XT's and a BRC laying around for months here...somebody interested? ;-) <

Keep them for location recording ? Rent them to others, they're just lying around after all ...

>"Is this a convenience issue, or are you really satisfied with the sonic results you get from your DSP stuffed DAW these days ?"

...convenience/speed first. I'd say that the "louder" the music, the more difficult it gets to mix only in a DAW . I've been working many years on an ADT, a SSL G4000+ and a Neve Legend, and know what comes out of such boards. Digital only is getting close, but not quite there...on the other hand, I'd never start recording classical stuff with analog gear again! For jazz: well working too. When I do rock stuff, I patch my DAW in the board for summing. Best of both worlds! :-) <

And use the Weiss EQ's and other digital tricks ? Only leveling on the analog board ?

I've heard people complaining about summing problems with digital buses before. I don't understand it, but I take your word for it. Is this a matter of a somewhat mushy bus sound when driven hard, or is it a subjective matter ? Digital sounding cleaner when driven harder is a fact but remains "unnatural" for our ears ...
Would going D > A, two track real to real and / or tube stage, A > D, solve the problem, or is this just a poor workaround (Did I hear someone say: "Magneto" )? :)

>"What exactly do these DSP's actually do ? The maths like the Pentium would do on a host based system ?"
...basically yes - the big difference is the efficiency of modern DSP-cards. For ex, a Philipps TriMedia chip running at 150MHz is as powerfull as a PIII 500 for audio applications. Now put up to 8 of those chips in a DAW! I'm sure you see the picture...<

I do !

>"You have to rely on the proprietary algorythms then. Are these any good, compared with say, Nuendo or Samplitude plugins ?"
...oh yeah! The eq's and the dynamics were developped by Daniel Weiss, one of the gurus in the world of digital audio.<

I have no more questions :) !
I take it you are satisfied with your EQ's ...

>the great developper Vincent Burel is porting a lot of his stuff on pyramix. And: any DirectX plugin will run with the system - not on the DSP-cards, but on the host-cpu.<

Very thoughtfull of them Pyramix guys ....

>...one card with adat interface is around 3200.- swiss francs...and I guess you'll need 2 or 3 of them for your purpose (you could sync your atari to pyramix, and record your tracks in 2 shots or more). I know, quite a lot of money, but count what you'd have to spend for an similar protools setup - which, btw, can't run at 96kHz at this time.<

And of which numerous people have stated that it (PT) yet has to produce the first good sounding track ! I'll look up how much 3200 Swiss Francs is, duh ... :(

>...adat TOS-link splitted. You could go with MADI or AES-EBU as well by using other ADC's.
But hey, just have a look at http://www.merging.com ;-) <

I did, not an awfull lot of info out there. I'll look again, though.

> ...located ?"...switzerland. Yourself?<

Holland.

Anyways Benoit, I really appreciate the time you take to share your knowledge and experience. People who own such boards and keep a Nuendo system for kicks in their toilet are usually very busy ...
Give my regards to Boris Blank and Dieter Meier, when you see them this weekend at the pub ?

Regards,

Paul.

"But hey, just have a look at http://www.merging.com ;-) <
I did, not an awfull lot of info out there. I'll look again, though.

...I know. They are only a few people there, and the promo staff at digidesign is at least as big as the developper crew at merging. But feel free to mail them if you have any questions, they are very helpfull. Or have an eye at http://www.media-assistance.com ; those guys market pyramix in Germany; you might find more infos...

"Give my regards to Boris Blank and Dieter Meier, when you see them this weekend at the pub ?"

...well, the only pub that is going to see me this weekend will be an airless and overheated control room - & I'll have to take my beer with me! ;-)

Cheers
Benoit

[ March 25, 2001: Message edited by: Benoit ]

<< Main concern is to give the algo's enough meat to sink their teeth into, like I mentioned in previous posts, especially when EQ's are concerned. >>

The best EQ plugins I've heard are probably the McDSP ones. However, if we are talking about the finest quality EQ, the argument isn't between 48 or 96kHz, it's between plugins and outboard gear and at the moment there is no contest, high quality outboard EQ is still considerably better than plugins. This will change eventually, in fact I recently heard Digi's Reverb1. It's the first plugin reverb I've heard that deserves a place with high end outboard reverbs.

<< Would you subscribe RME ADI outboards as as lower end ? >>

I've never heard them so I can't really say. What I am saying is that the plugins on host based systems are not yet generally as high quality sonically as say the best TDM plugins, even those I've heard running on 96kHz systems. This is the situation at present and why I feel the current batch of 96kHz systems aren't as good as the dedicated DSP 44.1/48kHz systems for mixing.

Greg

>>The best EQ plugins I've heard are probably the McDSP ones. However, if we are talking about the finest quality EQ, the argument isn't between 48 or 96kHz, it's between plugins and outboard gear and at the moment there is no contest, high quality outboard EQ is still considerably better than plugins. This will change eventually, in fact I recently heard Digi's Reverb1. It's the first plugin reverb I've heard that deserves a place with high end outboard reverbs.<<

Point taken, Greg. I would not dare to compare the EQ in a Nuendo-ish system, or even in a dedicated dsp system for that matter, to a top notch outboard EQ box. Not even when that host/dsp based system is running 24/96.
Given the undoubtedly big price difference, it isn't fair to expect such quality in the first place.
The point is, like I was discussing with Benoit, if for instance the EQ's on a host / dsp based system would sonically benefit from running at 96 k. Benoit states that is in fact true, 96 k does something to max out the sonics in host/dsp systems, like I was more or less expecting.

__Would you subscribe RME ADI outboards as as lower end ? __

>>I've never heard them so I can't really say. What I am saying is that the plugins on host based systems are not yet generally as high quality sonically as say the best TDM plugins, even those I've heard running on 96kHz systems. This is the situation at present and why I feel the current batch of 96kHz systems aren't as good as the dedicated DSP 44.1/48kHz systems for mixing.<<

I hefta take your word for that. TDM's have been around for a longer time, right? They probably have matured more. I don't expect 96 k to be a magic trick, where suddenly a mid-level system outperforms the Big Buck systems, just to clarify my thoughts.

Thanks for the insight Greg,

regards,

Paul.

"there is no contest,high quality outboard EQ is still considerably better than plugins"
...mmm...are you talking about digital EQ's?
If yes, which ones? ;-)
Don't forget one thing: many fx's (especially EQ's) are based on the same - or very similar - algorithms: the diffence in sound is induced by a different user interface, and another way of reacting to the controls.

Cheers
Benoit

Well Benoit, if I may post a reply before Greg answers ...

When designing an EQ algo, you can at least go two ways: Doing it straighforward, like in "mathematically correct", or recreating a famous analog EQ character, with all the anomalies that came with these. One could in fact appreciate these as more "musical" sounding, although they are mathematically incorrect. Such efforts are more likely to come from people who were already involved in designing analog eq circuits, like maybe George Massenburg ?

Maybe Greg meant something else. Your turn, Greg ? :)

Paul.

I'd like to clear up a point. The advantages of 96kHz over 44.1/48kHz are not directly due to the higher sample frequency, they are due to the filters. In 44.1kHz the filters are very sharp brickwall filters to keep all the frequencies below the Nyquist point (half of the sample frequency). The problem being with 44.1kHz that the Nyquist point is very close to the upper range of human hearing so the filters have to have a very sharp slope. With 96kHz the Nyquist point is 48kHz, well beyond human hearing and therefore the filters employed have a nice gentle slope. The laboratory tests I was invited to attend about a year ago left me with the impression that 96kHz gave a more defined and richer bass and more defined stereo image. I have to say that while there was this noticable improvement I'm not sure how noticable the difference would be in a home studio. The listening tests were carried out in a very high end monitoring environment.

Baring in mind the improvement of 96kHz over 44.1/48kHz is due to the brickwall filters in the AD conversion, I can't see how feeding 96kHz to a plugin is going to make much of a difference.

<< Would you subscribe RME ADI outboards as as lower end ? >>

going back to this question briefly: I heard some Lucid Audio converters not long ago, I was blown away, best ADCs I've ever heard. Then I was told the price, $11,000 for the two channel unit. Now that's high end!! :D

Greg

>>going back to this question briefly: I heard some Lucid Audio converters not long ago, I was blown away, best ADCs I've ever heard. Then I was told the price, $11,000 for the two channel unit. Now that's high end!! <<

Thanks for ruining my day, Greg ! :) really, I'm sure units like the Lucid make RME's pale by comparison. I do not expect the RME's to perform at that same level ...

Now, in the way you describe things, going to 96 k and processing at that rate will only be beneficial if the end product will be released at 96 k, like SACD or DVD. Or do I hear you wrong ?

And yes, I expect the subtlest of sonic improvements to remain inaudible in a typical home studio. Still, there's nothing wrong with trying to max out the sonic performance of a given setup. Discussing topics like these could eventually lead to a better understanding of things, which is nice, cos' I'm always interested in learning about studio-technology.

Regards,

Paul.

<< Now, in the way you describe things, going to 96 k and processing at that rate will only be beneficial if the end product will be released at 96 k, like SACD or DVD. Or do I hear you wrong ? >>

That's about the size of it! If you record and at mix at 96kHz then convert down to 44.1kHz the only possible difference I can think of would be the filters used by your software to perform the conversion rather than the filters on the ADCs going in.

It may be time to pay a visit to Stephen Paul's forum. He would hopefully be able to tell you for sure whether my assertion above is true.

Greg

Hello Greg.

<>

I'm aware of that thread. So far we're waiting for an in depth answer by Stephan.

Still, some EQ plugins for setups working at 44.1 / 48 do upsampling to 96 k first, then perform the calculations and sample down to the original format. Have you heard about that ? BTW, what are you using ?

Paul.

Hi Paul,

<< Still, some EQ plugins for setups working at 44.1 / 48 do upsampling to 96 k first, then perform the calculations and sample down to the original format. >>

Thinking about this logically: You've tracked at say 48kHz so your ADC's filters have removed all frequencies above the Nyquist point, which is about 24kHz. You now stick in an EQ plug that upsamples to 96kHz but there are still no frequencies in the track above 24k. How can you cut/boost or have any effect on frequencies that simply don't exist in the original material? The plug then downsamples to 48kHz, once again employing the harsh brickwall filters but within our 24k limit. If anything, logically, this has got to be worse than just using a straight 48kHz plug.

Greg

"If anything, logically, this has got to be worse than just using a straight 48kHz plug."

I think that there is some wrong aspect in your explanation. I'll try to explain, but for now: have you ever done the test by yourself? If not, take a good file at 44.1 or 48, play it through the best 96-compatible EQ that you might have, convert your media to 96kHz, open a new project at 96kHz, play the song with the same EQ-settings you had before, and just LISTEN... - without converting back to 44.1 for the moment... ;-)

Cheers
Benoit

Hello Benoit.

Could you describe the effects you hear when performing the test you mentioned ? And when you convert back to 44.1 after EQ'ing at 96 k, did you test it that way yourself, and with what results if I may ask ?

Paul.

Hello Paul,
That "upsampling" stuff you wrote about isn't exactly working the way you described it. However, there is a situation where a e.g. 16-bit word could benefit from 32 or 48 - bit processing, as it happens in most high - end TDM plugins. Whenever you change anything in a 16 - bit sample (level, frequency contents, etc), it is likely that you'll need more than 16 bits to describe the result - hence the need for a bigger processing reservoire insude the plugin. When the processing is done, you'll need to downsample it to main mix bus word length. Good plugs use dither and noise shaping here to keep the distortion low. Bad ones simply truncate those bits that carry the finest information about sound.
That's why McDSP and Waves Ren sound goooood!

Branko

p.s. I'm mixing feature films in TDM for more than a year. I'm very happy with the sound I get from this system, and I'm even happier since I've installed Rosendahl Nanosyncs clock and bb generator. A huge difference!

Hello Branko.

Thanks for your post. I take it this Rosendahl sync is more stable that the built in clock in your converters ? How does this translate sonically ? Better imaging and stereo separation perhaps ?

Paul.

Hi Benoit,

I haven't done the test myself as I am a PT TDM user.

EQ is the cutting or boosting of frequencies in the signal. As the 44.1k signal does not contain any frequencies above about 20k there should not be a difference between EQ'ing 44.1k material at 44.1k or 96k. If there is a difference it's a matter of great concern because either the conversion process or the EQ plugin must be creating frequencies not in the original material.

Greg

Hello Paul,
As I'm doing all of my work in Dolby stereo and Dolby Digital (4-track matrixed and 6-track discrete), I don't feel the difference in stereo imaging. However, there is a nice, not so subtle difference in sound transparency and definition. I can hear more details in sounds now, to describe the effect in short. Definitely extended the life of my 888/24 boxes! I was already looking for some Apogees, when I discovered this solution.
Branko

Hello Branko.

This is good to hear and hardly a surprise. After all, what good is a top-of-the-line converter without a good clock ?

Was this upgrade from buying the seperate clock cheaper than buying a new converter ?

Paul.

Hi Greg,

"EQ is the cutting or boosting of frequencies in the signal. As the 44.1k signal does not contain any frequencies above about 20k there should not be a difference between EQ'ing 44.1k material at 44.1k or 96k."

Of course, but I'm not talking about the _spectrum_ contained in the file itself, but about the different way the FILTER sounds - and works - at 96kHz. Means: a 5dB boost at 15kHz for ex. won't sound the same with an EQ working at 44.1 and the same EQ working at 96kHz, using the same audiofile.
EQ curves at 96kHz - having the Nyquist limit frequency 2x higher then 22 or 24 kHz - look much more natural, more similar to analog ones at high frequencies - not pushing the maths to the max around this crucial point.

Cheers
Benoit