electro-music.com   Dedicated to experimental electro-acoustic
and electronic music
 
    Front Page  |  Radio
 |  Media  |  Forum  |  Wiki  |  Links
Forum with support of Syndicator RSS
 FAQFAQ   CalendarCalendar   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   LinksLinks
 RegisterRegister   ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in  Chat RoomChat Room 
 Forum index » How-tos » Production - engineering/mixing
the Digital/Analogue time bog
Post new topic   Reply to topic
Page 3 of 6 [126 Posts]
View unread posts
View new posts in the last week
Mark the topic unread :: View previous topic :: View next topic
Goto page: Previous 1, 2, 3, 4, 5, 6 Next

Digital and Analogue together?
Totally
66%
 66%  [ 10 ]
Mostly
6%
 6%  [ 1 ]
A little bit
26%
 26%  [ 4 ]
No way
0%
 0%  [ 0 ]
Total Votes : 15

Author Message
Afro88



Joined: Jun 20, 2004
Posts: 701
Location: Brisbane, Australia
Audio files: 12
G2 patch files: 79

PostPosted: Thu Jun 08, 2006 11:01 pm    Post subject: Reply with quote  Mark this post and the followings unread

Thanks very much for replying Kassen, there's alot of great info in there.

Quote:
What's going on here? The errors added up. You can now see you could prevent this by using higher resolution files and only going down to cd quality at the end; in that case the rounding errors do get in the file but in the bits below the 16th (hopefully!) and still get discarded.


This inspired me to do some tests for myself. I did one in a 96k/24bit session with a 96k recorded saw wave and 5 chorus insert effects that are rated to 96k. I exported this to a 44.1k file and then did the exact same thing exclusively in 44.1k/16bit (including recording the same saw wave). Then I loaded each file into winamp multiple times, turned on shuffle and closed my eyes. Each time I was able to guess the file correctly, even though musically they are exactly the same - the only difference is one was rendered at 96k/24bit and downsampled to 44.1k/16bit and the other was done exclusively at 44.1k/16bit. Each time the 96k rendered one sounded clearer and better defined, and the 44.1k one sounded the opposite. Not once did I make an incorrect guess. This is not a subjective manner, it is maths - there are audibly less rounding/quantisation errors when it's rendered in 96k. Although I'm sure this will become a subjective matter in the 2010's when the 90's sound gets revived Wink

I think that's convinced me to start using higher sample rates whenever effects are going to applied. I typically run about 24+ tracks per song, with roughy 8 of these playing at once. If I could hear it on 1 track, there must be quite alot of noise going on when 8 are playing. And I understand what you mean when you say the noise isn't random but related to the audio itself (interaction between the sampling frequency and the frequencies it's trying to represent). It's probably for this exact reason that I never really picked up on it before, because I didn't know what I was listening for. I was listening for an analog noise floor or distortion from something clipping somewhere. If those two are missing, we've got a clean signal right? Rolling Eyes Sure I've heard the digitizer effect like everyone else, but it's so much more subtle at 44.1k (rather than 11k) and you can only really hear it when it adds up.

Quote:
When I first started using plug-in effects, I was lucky enough to learn about this phenomenon soon afterward.


Indeed you were one of the lucky ones. I'm sure there are many people out there who like me were taught that there's no point doing anything above 44.1k because it's all going to end up at 44.1k anyway when you burn to cd. I've never read or heard anything about too many plugins increasing quantisation noise. It's a bit too don't ask don't tell if you ask me.

By the way, if anyone wants I can post the wav files that I refered to above and post the exact method I used to create them. I'd be interested to see what other people think when they hear them.
Back to top
View user's profile Send private message Visit poster's website
elektro80
Site Admin


Joined: Mar 25, 2003
Posts: 21959
Location: Norway
Audio files: 14

PostPosted: Fri Jun 09, 2006 12:44 am    Post subject: Reply with quote  Mark this post and the followings unread

That is a great post, Kassen. I guess we have touched upon this several times before, but this post kinda summed up a lot of info.

As for high samplerates etc. , at least go for 24 bit no matter what. 24bit at 44.1khz is better than 16 bit. Another issue is that some DAWs and gear wont handle the 96khz to 44.1khz conversion that well either. In those cases 88.2khz might be a better choice.

Anyways, Kassen´s´post kinda explains why some prosumer digital mixers with effects have a really obnoxious sound quality. All the specs are "right" but it still sounds definitively "weird".

This also kinda explains again why I have been suggesting that certain types of signal processing should be done in the analog domain if possible.

As for samplerates etc. my old advice has been "use the high samplerates when you know you are going to process the signal a lot".

So, I reckon the next hot DIY projects will be analog preamps, linedrivers, summing arrays and compressors... and EQs? Very Happy

_________________
A Charity Pantomime in aid of Paranoid Schizophrenics descended into chaos yesterday when someone shouted, "He's behind you!"

MySpace
SoundCloud
Flickr
Back to top
View user's profile Send private message Visit poster's website
elektro80
Site Admin


Joined: Mar 25, 2003
Posts: 21959
Location: Norway
Audio files: 14

PostPosted: Fri Jun 09, 2006 12:47 am    Post subject: Reply with quote  Mark this post and the followings unread

Trivia: These facts are known to most plugin programmers and DSP designers. Some even consider this to be relevant. Not all plugins are designed the same sad way.
_________________
A Charity Pantomime in aid of Paranoid Schizophrenics descended into chaos yesterday when someone shouted, "He's behind you!"

MySpace
SoundCloud
Flickr
Back to top
View user's profile Send private message Visit poster's website
Afro88



Joined: Jun 20, 2004
Posts: 701
Location: Brisbane, Australia
Audio files: 12
G2 patch files: 79

PostPosted: Fri Jun 09, 2006 1:28 am    Post subject: Reply with quote  Mark this post and the followings unread

elektro80 wrote:
Trivia: These facts are known to most plugin programmers and DSP designers. Some even consider this to be relevant. Not all plugins are designed the same sad way.


I just did some more tests with EQ's. All I can say is Waves Q10 is exactly the same in both 96 and 44.1 when you export to 44.1. Even when running 5 of them as inserts on a track, all eqing their heart out. I guess that says something for the Waves coders! Unfortunately I don't have the Waves Metaflanger effect to test with though - that would have been interesting.
Back to top
View user's profile Send private message Visit poster's website
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Fri Jun 09, 2006 8:02 am    Post subject: Reply with quote  Mark this post and the followings unread

opg wrote:
[Therefore, it was easy to apply this same logic to the plug-ins, and I didn't even let the digital/analog issue enter my mind. It was simply sound degradation.


Yes, that's right. Chaining modules, analog or digital, will give sound degradation. The characteristics of the degradation are due to the design of the modules. Unless the modules are very poorly designed, these characteristics are independent of whether the underlying circuits are digital or analog.

I used to avoid bouncing tracks on my analog multitracks because of this degradation. Then one time I got to work on a really nice professional Studer machine. It was a different experience. You could bounce tracks several times and they'd still sound better than an unbounced track on my Tascam machines (all I could afford at the time). On Sonar at 24 bit and 192 KHz you can bounce quite a bit and not worry about picking up artifacts. Anyway, my point is that some of the really good old professional tape recorders (Studer, Nagra, etc), when well set up and aligned, sound more like digital recorders than commercial/consumer grade tape machines.

Quote:
With digital, if I can't hear these phenomenons which are obviously occuring, I would be left with so many options to try and fix it, including adding more/different plug-ins, but maybe those plug-ins weren't optimally designed, etc etc.


Interesting observation. It may be that if one had many very expensive and sophisticated low-noise, low-distortion analog modules in a chain you would possibly get this same experience. I think it's not so much the digital nature of the devices, but their functionality. This is difficult to demonstrate because some of the plugins if made with analog technolgy would be really expensive - many don't exist at any price.

_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
FLechdrop



Joined: Nov 11, 2005
Posts: 72
Location: Amstelveen, Holland
Audio files: 1

PostPosted: Fri Jun 09, 2006 8:46 am    Post subject: Reply with quote  Mark this post and the followings unread

Afro88 wrote:
By the way, if anyone wants I can post the wav files that I refered to above and post the exact method I used to create them. I'd be interested to see what other people think when they hear them.


That'd be interesting! Please do. Very Happy
Back to top
View user's profile Send private message Send e-mail MSN Messenger
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Fri Jun 09, 2006 8:55 am    Post subject: Reply with quote  Mark this post and the followings unread

elektro80 wrote:
Trivia: These facts are known to most plugin programmers and DSP designers. Some even consider this to be relevant. Not all plugins are designed the same sad way.


Almost all DSP chips have double-precision mode. If you are concerned about digital headroom artifacts and sound quality you can use it for intermediate results. On a general purpose microprocessor, you can use double precision or floating point. (There are even a few floating point DSPs too). Good digital circuits use an internal sample rates that are higher than the I/O rate.

Since the 1970s there are programs (free from the IEEE) that design IIR filters that will be very stable. Don't forget, you can design unstable filters using analog devices too. Very Happy

It's great to do tests, but it's important to be clear on how to interpret the results. It's tempting to jump to conclusions. If you test one particular sample rate converter you might make conclusions about that one, but there are many different algorithms.

How digital quantization noise interacts with the frequencies of the source material is also dependent on the DSP design. There are several techniques to minimize these problems.

About compressors. I've designed many analog compressors. There is one serious problem. If you are going to have an effective compression circuit, you need a delay. Analog envelop followers have a time lag. If you are using non-instantaneous attack time, then it is possible for transients to sneak through. Analog designs get around this with non-linear saturation circuits (tube pre amps are popular but you can do this with solid state devices also) - maybe the sound of these is attractive but it's an artifact. You can do much better compression in the digital domain because you can delay the signal - analyze it - and then process it. Essentially, you can see the peaks and transients before you adjust the gain. It's also much easier to make adaptive algorithms.

I would use an analog compressor on the input stage of my mixer if there was the possibility of unexpected loud transients. Here, non-linear saturation will save your butt. This was just what we did on the streaming at electro-music 2006. It worked great when someone dropped a mic or feedback started up.

All this talk about technology is interesting but it's a diversion. You can have all the greatest equipment in the world and still make audio that doesn't sound good and conversely you can make great productions using simple cassette decks and funky stomp boxes for signal processing.

_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
elektro80
Site Admin


Joined: Mar 25, 2003
Posts: 21959
Location: Norway
Audio files: 14

PostPosted: Fri Jun 09, 2006 9:41 am    Post subject: Reply with quote  Mark this post and the followings unread

Well, you are kinda making our collective point here. Ordinary musicians and hobbyists cannot afford all the latest state of the art digital gear. What people end up with in ther home studios will be some run of the mill DAW and a random collection of bundled plugins/free stuff and Behringer class "high end" effects. This means you gotta stay focused and understand the issues when producing your music.
_________________
A Charity Pantomime in aid of Paranoid Schizophrenics descended into chaos yesterday when someone shouted, "He's behind you!"

MySpace
SoundCloud
Flickr
Back to top
View user's profile Send private message Visit poster's website
Afro88



Joined: Jun 20, 2004
Posts: 701
Location: Brisbane, Australia
Audio files: 12
G2 patch files: 79

PostPosted: Fri Jun 09, 2006 7:35 pm    Post subject: Reply with quote  Mark this post and the followings unread

mosc wrote:
It's great to do tests, but it's important to be clear on how to interpret the results. It's tempting to jump to conclusions. If you test one particular sample rate converter you might make conclusions about that one, but there are many different algorithms.


This is true, and is why I'm going to post the exact method after this with the wav files, so people know that I used Cubase SL with the default downsampling algorithm, no dither, the Classic Chorus free plugin from Kjaerhus Audio etc. I didn't do this before because I couldn't be bothered unless people were actually interested.

I'd guess that in all cases, when doing multiple treatments digitally, working in 96k would be better than working in 44.1k. But if you know of situations where this wouldn't be advisable, please do tell. Perhaps 88.2k would be better than working in 96k though?

mosc wrote:
All this talk about technology is interesting but it's a diversion. You can have all the greatest equipment in the world and still make audio that doesn't sound good and conversely you can make great productions using simple cassette decks and funky stomp boxes for signal processing.


That last part is right, because it comes down to knowing your gear and using it well. The first part I disagree with. This thread has taught me what studying dsp theory and working with analog gear for a few months would have taught me - that both systems have their own noise that adds up and there are ways to avoid it in both realms.

Like I said before, all the literature I've read about higher sample rates has taken the recording engineer's point of view - that 44.1k is enough to represent our hearing range, so why bother going higher. I now know that this isn't true if you are processing your audio again and again. I now know how to use my gear more effectively, and hence make better productions. I hardly think this is a diversion from making great productions, but I can understand that it would seem that way to you seeing as you know all this stuff already.
Back to top
View user's profile Send private message Visit poster's website
Afro88



Joined: Jun 20, 2004
Posts: 701
Location: Brisbane, Australia
Audio files: 12
G2 patch files: 79

PostPosted: Fri Jun 09, 2006 8:35 pm    Post subject: Reply with quote  Mark this post and the followings unread

Here's the specifics of my test:

I started with a plain saw wave recorded at 96k/24bit from my G2 (with Rob's tilt filter) into Cubase SL. Then I inserted 5 chorus effects (Kjaerhus Classic Chorus which is apparently 96k capable) on the channel and exported to a cd quality 44.1k/16bit wav file. Each chorus had the same preset and I adjusted the gain on the channel fader to obtain the hottest signal without clipping.

I then did the same but in 44.1/16, starting with a 44.1/16 recording of the same saw wave from the G2. Then I inserted the same plugin 5 times with the same preset and made the exact same gain adjustment. I exported this to a 44.1/16 wav file. Also, when you export each lfo gets reset to it's initial value so each export musically speaking is exactly the same.

Then I loaded each file into Winamp multiple times, hit shuffle and loop and closed my eyes.

edit: This test was performed with the signal too high, so both files have been clipped/limited quite badly and I've subsequently removed them. Please see the attachments in my post below for the correct versions.

Last edited by Afro88 on Sun Jun 11, 2006 2:10 am; edited 2 times in total
Back to top
View user's profile Send private message Visit poster's website
Kassen
Janitor
Janitor


Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3

PostPosted: Sat Jun 10, 2006 10:28 am    Post subject: Reply with quote  Mark this post and the followings unread

mosc wrote:

All this talk about technology is interesting but it's a diversion. You can have all the greatest equipment in the world and still make audio that doesn't sound good and conversely you can make great productions using simple cassette decks and funky stomp boxes for signal processing.


Hmmm, how would you feel about this;

"
All this talk about orchestration is interesting but it's a diversion. You can have access to all the instruments in the world and still make compositions that don't sound good and conversely you can write great compositions using simple soons and funky rubber bands.
"


I mean, the last bit is true but it hardly makes a case for not knowing *when* to use a grand piano instead of a rubber band. That to me is not a diversion. Televisions are diversions, romantic relationships and ringing phones are diversions. Orchestration and knowing your gear and so on are part the skills you should have at least to some degree as a modern composer/musician/whatever. Alissing can sound good but if you want it to sound good it helps a lot to know where it comes from and how it works.

Also; I can't imagine how you can say it's a "diversion" in a threat on this very topic in a forum like this.

_________________
Kassen
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Kassen
Janitor
Janitor


Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3

PostPosted: Sat Jun 10, 2006 10:36 am    Post subject: Reply with quote  Mark this post and the followings unread

elektro80 wrote:

So, I reckon the next hot DIY projects will be analog preamps, linedrivers, summing arrays and compressors... and EQs? Very Happy


Could well be, yes.

I also think that as digital systems become more common-place musicians will slowly learn more about them. It's understood that if you are a serious electronic guitarist you should know a little about elements and how to pick the right setup for your "voice" as well as perhaps make some modifications. I think that if you rely on digital systems a lot it can't hurt to read up a little on DSP. Nobody expects you to implement your own algorithems but some knowledge of the phenomena in general helps a lot.

_________________
Kassen
Back to top
View user's profile Send private message Send e-mail Visit poster's website
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Sat Jun 10, 2006 12:08 pm    Post subject: Reply with quote  Mark this post and the followings unread

Kassen wrote:
All this talk about orchestration is interesting but it's a diversion. You can have access to all the instruments in the world and still make compositions that don't sound good and conversely you can write great compositions using simple soons and funky rubber bands.


Well, spoons and rubber bands can be cool, but I was thinking more of a single solo flute.

Quote:
Also; I can't imagine how you can say it's a "diversion" in a threat on this very topic in a forum like this.


Theat? Who did I threaten?

[EDIT - someone sent me a PM - apparently you meant to write "thread", not "threat". Well, that's a treat. Wink ]

_________________
--Howard
my music and other stuff

Last edited by mosc on Sat Jun 10, 2006 1:06 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Sat Jun 10, 2006 12:52 pm    Post subject: Reply with quote  Mark this post and the followings unread

Afro88 wrote:
Here's the specifics of my test:


I don't hear all that much difference, but I'm not sure what to listen for. Perhaps if you pointed out what to listen for, I would learn to recognize it. Maybe not. I just had my hearing tested at the audiologist last week - I do this every year near my birthday.

Both tracks show extensive clipping when I look at them in Audacity, so maybe these particular files are not the right ones. You said you recorded at a level where there was no clipping. Did you make sure the levels between the 5 stages of processing didn't clip as well? If so, how did you verify that? Phasors (both analog and digital) can be adjusted to have a lot of feedback so it's possible to get some very high signals levels. When you chain several devices, this potentiality might be enhanced.

Clipping is, of course, a non-linear process and should be avoided unless you are interested in that paricular effect.

If you are going to use 24 bits, and you are going to do a lot of processing, then use those extra bits for headroom. If you record your source material within a couple of dB of the maximum level, then there isn't too much advantage of using 24 bits.

Think of the tracks as glasses used in a bar. The drinks are 8 oz. If they use 8 or 10 oz glasses, then they'll spill when the drinks are stirred during mixing or carried by the waitress (or waiter) to the customer. Every drop spilled is like noise in the music. If the bar switched to 14 or 16 oz glasses, then the spillage problem would be significantly reduced. But this is the case only if the drinks are kept at 8 oz. If they start filling the glasses, then the spillage becomes a problem again. You could use 64 bit recording and run into these same problems if you use signals that are too hot.

_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
ocp



Joined: Sep 14, 2004
Posts: 533
Location: Porto, Portugal
Audio files: 9

PostPosted: Sat Jun 10, 2006 1:44 pm    Post subject: Reply with quote  Mark this post and the followings unread

mosc wrote:
Afro88 wrote:
Here's the specifics of my test:


I don't hear all that much difference, but I'm not sure what to listen for. Perhaps if you pointed out what to listen for, I would learn to recognize it. Maybe not. I just had my hearing tested at the audiologist last week - I do this every year near my birthday.



I found it hard to notice any difference at first, but then the 96k file seamed to have more "air".
In other words, it had more sparkle, the effect was more evident.
The 44.1k file sounded dryer, sort of "without the polish".
It would be intersting if others wrote about their impressions.

_________________
https://opcabpol.bandcamp.com | https://www.mixcloud.com/Boiar | http://editoradoporto.blogspot.com |
https://www.facebook.com/nofp15/
Back to top
View user's profile Send private message Visit poster's website
Kassen
Janitor
Janitor


Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3

PostPosted: Sat Jun 10, 2006 3:24 pm    Post subject: Reply with quote  Mark this post and the followings unread

mosc wrote:

Well, spoons and rubber bands can be cool, but I was thinking more of a single solo flute.


Yes, perfect!

Solo flute is a marvelous instrument, it's good for a wide range of styles and very portable; I like it and I still mean to work with the flute I have a around here in a more serious context.

Let's imagine that moment; If at that point I need some thrills in the mid-high followed by some breathy tones I'm in business because the flute is good at those. If there would be a need for some complex extended chord to swell over the cource of a long bridge section it might be wiser to look elsewhere because the flute can't be expected to do that, at least not without lots of trickery.

To get back to the 24 bits; how much you need all bits depends on a lot of factors like the dynamics of the piece, the resolution the end result needs to have and the total ammount of processing to be performed. For pure recording and playback I agree that it may be better to go for a concervative amplitude in order to avoid clipping. In extreme cases and where CPU's perit I'd considder upsampeling to 64 bit float. I tink Audacity supports 64bit floats since AMD released their 64 bit processors. Sure; that's overkill but if you can without aditional expenses, why not? In general I advocate keeping all signals as hot as they can be without clipping. In practice that's a tricky tradeoff because if you are both the recordingengineer and the musician it might be good to keep levels lower so your performance isn't degraded by subconcious wories about clipping.

And yeah; I mix up "thread" and "threat", it may be a hot topic, but not that hot! :¬)

_________________
Kassen
Back to top
View user's profile Send private message Send e-mail Visit poster's website
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Sat Jun 10, 2006 5:38 pm    Post subject: Reply with quote  Mark this post and the followings unread

Well, that's interesting - having 64 bits. It will be a while before you can get that from Bheringer or MOTU. Very Happy

No matter what the bit width, though, you still need to be careful to give yourself head room on both ends. If you set your levels just under the clipping point you will get the best possible recording of the low level material, but you have very little room for unexpected loud events. If you reduce the level too low, you may not overload, but the unexpected quiet passages could be noisy.

How one sets the input level control reveals a lot about ones philosophy of life and personality. It is good to have clairvoyant abilities, just as with buying stocks, when setting the levels. Some people have the knack...

I have had the priviledge to jam in a setup that Kassen engineered way back in 04. Kassen has the knack... thumb up

_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
Kassen
Janitor
Janitor


Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3

PostPosted: Sat Jun 10, 2006 6:22 pm    Post subject: Reply with quote  Mark this post and the followings unread

Yes, keeping some headroom has a lot of advantages, particularly if later processing is needed. It may well save several steps of re-scaling the data.

To get back for a second to that phaser; most modern plugins use floating point internally and actually clipping a good floating point implementation would require some very serious runaway feedback (we're talkingorders of magnitude here) after the phaser itself you'd have plenty of room for some tactfull non-aliassing limiting (this would be a good occasion to use a digital limiter, BTW, Mosc, you asked about that a while back), get it all back in some acceptable range and downsample to normal integer values again. Such a operation, even for extreme settings shouldn't need to clip the file by itself.

And yes; 64 bits is extreme but it makes sense. If you have a 64 bits processor you might be able to do 64bits operations at the same speed as 24 bit ones so if the memory and disk-space is there then why not? The reason MOTU and Behringer don't offer it is that a fast 64 bits ADC would be very expensive (might be impossible at the moment?). For a 16 bits ADC you need to make 16 either/or type desisions in series, one for each bit; those need to be done within one sample's time-span. 64 bits means four times as many of such calls, each getting progressively closer to the noise floor of the components and when you think about it; 196KHz (which people would demand, probably) realy is quite fast.

ADC speed v.s. resolution is another case where educated trade-offs need to be made and where theoretical perfection is well out of reach. I'd urge against losing sleep over this but I wouln't close my eyes and dream I had perfection either.

_________________
Kassen
Back to top
View user's profile Send private message Send e-mail Visit poster's website
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Sat Jun 10, 2006 6:46 pm    Post subject: Reply with quote  Mark this post and the followings unread

If used as a post processing tool, it (compression/limiting) can perform marvelously because of look ahead capability. It's sorta like rewriting history on the fly.
_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Sat Jun 10, 2006 6:54 pm    Post subject: Reply with quote  Mark this post and the followings unread

Kassen wrote:

ADC speed v.s. resolution is another case where educated trade-offs need to be made and where theoretical perfection is well out of reach. I'd urge against losing sleep over this but I wouln't close my eyes and dream I had perfection either.


OMG - if we don't close our eyes, we'll loose sleep.
But, if we sleep, we will dream of perfection. Shocked

I guess one might approach this conundrum just like one goes about adjusting the input level. Hmmm. I'll pick sleep and dream of perfection... Very Happy

Posted Image, might have been reduced in size. Click Image to view fullscreen.

_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
Afro88



Joined: Jun 20, 2004
Posts: 701
Location: Brisbane, Australia
Audio files: 12
G2 patch files: 79

PostPosted: Sun Jun 11, 2006 2:08 am    Post subject: Reply with quote  Mark this post and the followings unread

mosc wrote:
I don't hear all that much difference, but I'm not sure what to listen for. Perhaps if you pointed out what to listen for, I would learn to recognize it. Maybe not. I just had my hearing tested at the audiologist last week - I do this every year near my birthday.


That's fine, I don't doubt your hearing Smile Can I ask what kind of monitoring setup you're using though? It's all in the clarity of the top end - it seems more blurred and more dull in the 44.1k version.

mosc wrote:
Both tracks show extensive clipping when I look at them in Audacity, so maybe these particular files are not the right ones. You said you recorded at a level where there was no clipping. Did you make sure the levels between the 5 stages of processing didn't clip as well? If so, how did you verify that? Phasors (both analog and digital) can be adjusted to have a lot of feedback so it's possible to get some very high signals levels. When you chain several devices, this potentiality might be enhanced.


Yes, my apoligies, I accidentally the level too hot in Cubase. The problem was that when I was getting the gain structure right, the chorus' lfos were all out of sync. When I exported they were all in sync, so frequencies that were accentuated from the feedback of one chorus were accentuated much more with each successive chorus so the level was too hot.

Not that it makes any differerence to the test. They're both still exactly the same, just rendered at different bit depths and sample rates. However, I never stated that the signal was limited/clipped in the test, so I've added two new files to this post that have a couple of dB headroom, and definitely don't on the input or output of each chorus. Predictably, this new gain structure actually makes the difference more pronounced Smile I'd be interested to see if you can hear the difference with these two files mosc.

mosc wrote:
If you are going to use 24 bits, and you are going to do a lot of processing, then use those extra bits for headroom. If you record your source material within a couple of dB of the maximum level, then there isn't too much advantage of using 24 bits.


Yeah..... I know first hand that there is a justifiable advantage though. Back in Music Technology 3 we did some tests with bit depths and dither, specifically with reverb tails and track fades. If your recording has sections that fade to or from silence, or that reverb/delay out to silence then the extra 8 bits do make quite a difference. As a side note, choosing the right dithering algorithm when truncating to 16bits also makes quite a difference too. With storage space as cheap as it is these days, I think the hard drive cost vs. benefits of working in 24bit does work out.

Note: please see later in this thread for all audio files and a refinement of the test method.

Last edited by Afro88 on Tue Jun 13, 2006 6:04 pm; edited 1 time in total
Back to top
View user's profile Send private message Visit poster's website
Afro88



Joined: Jun 20, 2004
Posts: 701
Location: Brisbane, Australia
Audio files: 12
G2 patch files: 79

PostPosted: Sun Jun 11, 2006 2:15 am    Post subject: Reply with quote  Mark this post and the followings unread

ocp wrote:
I found it hard to notice any difference at first, but then the 96k file seamed to have more "air".
In other words, it had more sparkle, the effect was more evident.
The 44.1k file sounded dryer, sort of "without the polish".
It would be intersting if others wrote about their impressions.


Exactly. I'm glad someone else can hear it and I'm not going crazy Laughing
Back to top
View user's profile Send private message Visit poster's website
paul e.



Joined: Sep 22, 2003
Posts: 1567
Location: toronto, canada
Audio files: 2

PostPosted: Sun Jun 11, 2006 7:55 am    Post subject: Reply with quote  Mark this post and the followings unread

the 96khz does seem to have more 'sparkle' on top and also perhaps a slightly wider stereo field
_________________
Spiral Recordings
Back to top
View user's profile Send private message Send e-mail Visit poster's website
mosc
Site Admin


Joined: Jan 31, 2003
Posts: 18197
Location: Durham, NC
Audio files: 212
G2 patch files: 60

PostPosted: Sun Jun 11, 2006 9:23 am    Post subject: Reply with quote  Mark this post and the followings unread

Yes, I hear this subtile difference. To me, your second non-clipping version (both sample rates) differs from the clipped version more than the change in sample rates, But if you just listen to the last two files, there is a slight improvement in the definition at 96 K.

I know a guy who runs a recording studio. He does very little non-linerar post processing. He professes to use 24 bit 44.1 KHz for his sessions because of the artifacts generated when down sampling. He feels, if it is intended to be used at 44.1, then use 44.1 throughtout. His recordings are superb.

I think there are cases that call for higher sampling rates and cases where it isn't the best choice. That's why there is a knob.

_________________
--Howard
my music and other stuff
Back to top
View user's profile Send private message Send e-mail Visit poster's website AIM Address
Kassen
Janitor
Janitor


Joined: Jul 06, 2004
Posts: 7678
Location: The Hague, NL
G2 patch files: 3

PostPosted: Sun Jun 11, 2006 12:51 pm    Post subject: Reply with quote  Mark this post and the followings unread

mosc wrote:

I know a guy who runs a recording studio. He does very little non-linerar post processing. He professes to use 24 bit 44.1 KHz for his sessions because of the artifacts generated when down sampling. He feels, if it is intended to be used at 44.1, then use 44.1 throughtout. His recordings are superb.


I can imagine this working out for a *recording* studio. Most of my comments are aimed mainly at at *generative* studios. I still feel that mixing those concepts up is one of the main factors that currently hold back electronic music.

I do agree that if you know for certain in advance what format you'll release to this does create a big advantage. If you know you'll be using 44.1 I'd advocate using integer multiples of that everywhere. Even some of the more reputable packages use atrociously bad downsampling algorithems.

Also; I imagine your friend does most of the "work" before the material hits the ADC's and he probably uses high quality outboard gear that he knows well. In that case this might be a very sensible choice. However; high quality outboard geart ends to be expensive and hard to pick for artists on a budget or that are starting out and so many use freeware plugins extensively. That creates a entirely different situation with different questions to answer. this is basically what I said above; Nequist talked about recording and playback and it holds true there. He does *NOT* talk about good choices when programing a good sounding drill&bass track using 50 VST plugins.

_________________
Kassen
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Display posts from previous:   
Post new topic   Reply to topic
Page 3 of 6 [126 Posts]
View unread posts
View new posts in the last week
Goto page: Previous 1, 2, 3, 4, 5, 6 Next
Mark the topic unread :: View previous topic :: View next topic
 Forum index » How-tos » Production - engineering/mixing
Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum


Forum with support of Syndicator RSS
Powered by phpBB © 2001, 2005 phpBB Group
Copyright © 2003 through 2009 by electro-music.com - Conditions Of Use