Dubstepforum.com

PHASEten wrote:

Macc wrote:

PHASEten wrote:Its got all the way to three pages and no one has insulted anyones mother yet, that in itself must be some kind of record lol.

your mum

oh well...... it was nice while it lasted.

!

TeReKeTe wrote:^ use a bunch of different samples and layer them based on velocity-- or just use battery, etc for things like this.

having a background in an instrument will help... like, if i wanted to make my ride sound like philly joe jones, i'd have a very strong, almost metronomic pulse on the ride, the hihat w/ foot ever so slightly ahead of 2 & 4, and have the 2nd note of the 3-note swing pattern on the ride be the quietest.

if i wanted to do danny richmond, i'd put the 2 swung notes of the ride cymbal really close together, and constrain the dynamics.

Yeah. I guess I should really start to do more stuff played in live then edited rather than just entered into piano rolls.

Need to play more with velocity and filter automation, too, I guess.

PHASEten wrote:
oh well...... it was nice while it lasted.

Eyy man, you know I was kidding!

Good discussion indeed.

Macc wrote:

PHASEten wrote:
oh well...... it was nice while it lasted.

Eyy man, you know I was kidding!

Good discussion indeed.

Macc wrote:

Rob Sparx wrote: You know how I work Macc for certain tunes I keep everything loud so that when a frequency steps out of line I can hear digital distortion and know somethings not right which doesn't happen when your working at a quieter volume, same as your stressing the mix technique. I do all my final mixdowns at -6 like you recommended and for many tunes I work with peaks around -6 but use a limiter on the master volume until the premaster recording. As you know when you've got a good mixdown there really is very little difference in sound between the quiet mix (and amp volume wacked up) and clipped/limited mix (amp volume down) but when you've got a dodgy mix the sound can sometimes audiably clip at anything over 0db - mine will usually go up to +6db before you can actually hear any clipping.

Excellent post.

It's a good way to work when you know what you're doing - though it's perhaps a little personal and it is definitely rather advanced for (with all due respect) a lot of the other engineers on here. I know how that sounds, sorry peeps. But it's definitely important to really and truly understand (not just know) how and why you're using 0dB in that way, and that takes experience which Rob obviously has.

FWIW I always kept * things low and used the 'super massive ridiculous compressor' job to check anything like that, stuff poking out etc, but it amounts to the same thing really. Lots of roads lead to Rome! Hope you get where I'm coming from Rob!






* just realised I said that in the past tense.... it's been soooo long since I built a mix!

I deffo get where your coming from Macc and I do prefer the quality of sound when its not in the red and for some tunes especially the more organic ones (Rhodes/guitars/vocals etc) the clipping is more noticeable so I work quiet for those tunes at least until the mix is balanced enough to cope with more volume. People should be scared of 'the red' though, lots of people argue NEVER to go in the red not even properly understanding why.

Macc wrote:I understand what Rob's doing - and as he says, he doesn't do it all the time - but you know how internet boards go... people start running their tunes hot 'cos that what Rob Sparx does and his tunes are fat'. That's why I'm hesitant to recommend it as a general way to mix. You can get the best of both worlds and run no risk of ruining things if you pop a plug on the out and clobber it (at unity gain of course) to check, then turn it off and return to a good gain structure as per that other thread.

I'm not a conservative bloke generally, I drink like a motherfucker and I drum like a nutcase. I love running tubes hotter than I should and pushing transformers further than they're meant to go (when appropriate anyway). But digital IS NOT ANALOGUE and it definitely pays, in the final analysis, to keep it conservative.

The floating point thing is kind of relevant by the way, but not entirely - it fits under the general 'digital works best under 0dB' thing.

Sorry for the waffle...

There's always ppl who will misunderstand ppls advice - just likes there's those who misunderstand your advice working in the red there will probably be other ppl who do clipped mixdowns in their projects (even if a recording is peaking at <0db it can still be clipping as digital meters can miss some short peaks so worth reducing the levels a bit and watching your gain structure etc) and do their premaster recordings in the red then say thats something I recommended.

BTW Macc would like to hear what you have to say about floating point format as I do all my premasters in 32bit and would like to know what the 'clipping data' the extra digit relates to actually is!

Nerdgasm

Rob Sparx wrote: There's always ppl who will misunderstand ppls advice - just likes there's those who misunderstand your advice working in the red there will probably be other ppl who do clipped mixdowns in their projects (even if a recording is peaking at <0db it can still be clipping as digital meters can miss some short peaks so worth reducing the levels a bit and watching your gain structure etc) and do their premaster recordings in the red then say thats something I recommended.

It's funny you mention intersample peaks - that's one huge disadvantage with using clipping on your output. I had one a week or two ago that was clipping at +4.2dB!! It sounded horrific, crackling badly all over the shop, brittle sounding… a mare.

BTW Macc would like to hear what you have to say about floating point format as I do all my premasters in 32bit and would like to know what the 'clipping data' the extra digit relates to actually is!

Nothing wrong with 32-bit FP, although it is kind of missing the point of running things hot (all dynamic info is retained, no clipping etc). But then looking at it the other way, it’s a handy ‘shortcut’ saving you turning it down and exporting at 24bit with headroom (all dynamic info is retained, no clipping etc). But further to that, 24 bit is generally a safer way to deliver to an ME or some such as it takes that whole sliding scale out of the equation. What I mean here is that if it is 32-bit, peaking over 0, and the ME wants to send it to analogue, he’ll have to turn it down under zero to avoid clipping the DAC on the way out (practically all PCM convertors on the market are 24 bit).

Anyway, so on the one hand it’s sort of useful, on the other it’s sort of pointless. 24 bit is more than adequate, actually slightly superior if you want to go really deep into it.

As for the technical aspect, floating point almost literally refers to the decimal point being able to ‘float’. Consider the numbers 10000.000 and 0.0001000 – the ‘only’ difference is the placing of the decimal point. So you can describe the ‘content’ of the number, called the significand, and then the exponent makes the decimal point move about.

So essentially what you have in a 32bit float file is 24 bits of ‘content’ (23 for content, one for the sign if I remember rightly) and then 8 bits to describe the exponent, ie scaling factor. When you think about this in terms of signals in your mixer, it becomes easy to understand why things won’t actually clip when operating in a floating point environment. The floating decimal point gives rise to hundreds of dB of headroom over 0dB. *

BUT

When all is said and done you always have to come down under 0dB to put things back into the ‘real world’. Convertors are 24 bit and there’s a ‘bit bottleneck’ at the output, which is why you still get clipping at your output, despite it being 32bit FP up to that point.

With all this in mind and with regard to this conversation: it makes more sense to me to mix the way I usually bang on about, and occasionally punch in a clipping/limiting plugin on your output set to, say +9in, -9 out. So if you’re mixing to leave about 3dB of headroom this will give you 6dB of clipping at unity gain. Then you can hear how well your mix holds up under those conditions without being bound by it, without having it there all the time, and without the extra volume fooling you. You can switch it in and out and hear how your mix is doing, but you aren’t married to the clipping or dealing with varying volume levels. Best of both worlds – just don’t forget to turn it off

Lastly, to say it again… submitting clipped mixdowns for mastering is a number one sin in my book. Even if they seem to sound great it always always ends up sounding not as good, with that horrid ‘fizz’ coming out as corrections are made and the level gets brought back up. Often (very often) it ends up in quieter masters, not louder. There’s already so much distortion hidden away. Using clipping as a tool as described above, cool. Just turn it off when you’re done.

Man that post got out of hand, sorry!


* It does cost slightly in terms of precision though as I understand it – it gets quite heavily into maths and it all gets a bit much for me tbh. I have to be in the mood to get into that stuff!

lol thanx for the explanation!

i hate getting 32 bit mixdowns besides the fact that i need waveburner to convert 'em to 24, i don't need 192db worth of transient info in order to do my job.

i don't want to move this into a mastering discussion, because it really, really was about drums, but i've been way into various bits of analog clipping on some drum busses lately. Macc-- i've tried clipping my a/d at about +2 at various points recently,w/ different amounts of usefulness. thoughts?

you definitely get more volume, sure, but that's not without some tradeoff.

@Rob - my pleasure. Obviously it's all way more complicated than that in truth, but that's about the size of it.

TeReKeTe wrote:i hate getting 32 bit mixdowns besides the fact that i need waveburner to convert 'em to 24, i don't need 192db worth of transient info in order to do my job.

It's quite a lot more than that I think... But yes.

i don't want to move this into a mastering discussion, because it really, really was about drums, but i've been way into various bits of analog clipping on some drum busses lately. Macc-- i've tried clipping my a/d at about +2 at various points recently,w/ different amounts of usefulness. thoughts?

Depends very greatly on the AD. I just moved up to the Mytek 8x192 ADDA * and clipping it sounds much better than my old Mytek Stereo96. Even from the same company they sound pretty different and take different amounts of abuse.

Handle with care, basically. Most of it is very 'hard knee' diode clipping. You might clip your signal and check it out and it sounds fine and then when you're recording ONE little peak will be that little bit too much and will bust up horrendously.


* the best convertor I have ever heard btw. I am in love!!

Macc wrote:* the best convertor I have ever heard btw. I am in love!!

What's the second best?

Future One wrote:

Macc wrote:* the best convertor I have ever heard btw. I am in love!!

What's the second best?

Second best I have heard? Well, I liked the Mytek Stereo96 AD a lot. Benchmark is very good... Lots of options all the way up really. You're after a DA I assume, as you have the A2D...? Are you clocking your interface from the API clock? That might well help, depending on the interface (sorry if you're already doing that).

Macc wrote:^ Think disco baby!

I'd take running all the percussion and drums and all that to a bus over sidechaining them every time, FWIW.

Amounts to the same thing in many ways, but gives much richer sound when they're all working together that way, under a nice compressor of course. Get the buss treatment right and the levels right going into that last group buss and you have lovely thick, dense rhythm tracks.

yeah do this, give a kind of side-chainey effect, but more subtle and nicer!

Macc wrote:

Future One wrote:

Macc wrote:* the best convertor I have ever heard btw. I am in love!!

What's the second best?

Second best I have heard? Well, I liked the Mytek Stereo96 AD a lot. Benchmark is very good... Lots of options all the way up really. You're after a DA I assume, as you have the A2D...? Are you clocking your interface from the API clock? That might well help, depending on the interface (sorry if you're already doing that).

I'm probably not going to upgrade my D/A at the moment. The ultralite suits me ok for now. Just interested.

I am clocking the ultralite from the A2D but to be hounest I didn't notice any difference in the D/A playback quality.

Cheers!

Future One wrote: I am clocking the ultralite from the A2D but to be hounest I didn't notice any difference in the D/A playback quality.

Test test test test

Wonder if Macc still has that Mytek 96 kicking around. Macc if you're reading this, mail me your leftovers please!