Login / Register
 
going from 44kHz to 96kHz! what to think about?
New Reply
Subscribe
Mr Keys
Thread Starter
#1
22nd February 2008
Old 22nd February 2008
  #1
Gear Head
 
Joined: Aug 2006
Posts: 47

Thread Starter
Mr Keys is offline
going from 44kHz to 96kHz! what to think about? I want oppinions!!!!!

We’ve been looking for a new computer to the studio (all we do basically is mix other peoples tracks, and record vocals). I started thinking about switching from 44.1 kHz to 96 kHz instead, at the same time…but a few questions hit me:

1: can you combine sample rates in logic? If the project is running 96 kHz, can we still throw in a sample at 44 kHz, without the smurfsound? Or is it pitched down? Whatever….

2: what parts in the computer will “drop out” first? Is it the HD? CPU? Or the memory? Will it help to RAID the hard drives (two hard drives that work as one, but a lot faster, right???)
We often work with a lot of softsynths combined with a lot of audio tracks.

3: will our UAD-card peak faster if we go for 96 kHz?

4: will there be any improvements in quality when it comes to the softsynths? Or do they work in 44.1 kHz no matter what, and you gain nothing from switching to 96 kHz?

5: our computer today is: 2 x 2 GHz G5, 2,5gb memory (2 x 1 GB plus 2 x 256mb) with two internal HD’s with NO raid. With this setup we do need some extra power sometimes…that’s the original reason for upgrading….
I know it’s hard to say, but if we switch to Intel Xeon 2 x 3 GHz Quad core, 8 gb memory (4 x 2 gb) and two 500gb Raided HD’s, AND go for 96kHz, will our system be more powerful compared with our setup today, where we work in 44.1kHz. Or will it be “slower”? Or?


Maybe I forgot to ask something, so, please enlighten me in what I need to think about before we go for 96 kHz?



Will we hear the difference?
For example: we used to record vocals with this setups: tlm49-vintagedesign CA73-summitaudio tla50-motu 2408mk3…now we use “Waves MaxxBCL”(everything bypassed though) instead of the motu, because of the converters, and I must say the difference is HUGE!!!
Will it be that kind of difference, or will we barely hear it?
#2
22nd February 2008
Old 22nd February 2008
  #2
Lives for gear
 
anguswoodhead's Avatar
 
Joined: Feb 2008
Location: On the East Coast of Australia
Posts: 3,861
My Recordings/Credits

Send a message via Skype™ to anguswoodhead
anguswoodhead is offline
I don't have all the answers but I have heard that Raid is NOT the go.
I have 2 x UAD cards so would be keen to know how they go at 96K - I have been thinking of making the same move.
I would say stay Mac (but I only own PC's - 4 of 'em)
I would suggest that going to 96K will def. sound better but have never tried.
I know some Mastering guys upsample everything then do all there processing and dither back to 44.1 right at the end.
Oh BTW when you import a piece of audio the software should give you the option to change it to the rate you're working at - so no smurfs etc.
__________________

Mastering all styles and genres using an hybrid approach.
Special rates for unattended online mastering.
Add 24 Ch SSL analog summing for an extra $50 per song.
Prefer 10 stereo stems.

Sunshine Coast Mastering, Queensland, Australia
#3
22nd February 2008
Old 22nd February 2008
  #3
Lives for gear
 
peeder's Avatar
 
Joined: Aug 2006
Location: No longer participating here.
Posts: 6,705

peeder is offline
The thing I'd be thinking about is why I didn't use 48K instead.
#4
22nd February 2008
Old 22nd February 2008
  #4
Audio Alchemist
 
Lagerfeldt's Avatar
 
Joined: Sep 2004
Location: Copenhagen, Denmark
Posts: 5,155

Lagerfeldt is offline
Quote:
Originally Posted by Mr Keys View Post
4: will there be any improvements in quality when it comes to the softsynths? Or do they work in 44.1 kHz no matter what, and you gain nothing from switching to 96 kHz?
It can never be an improvement in itself going from a 44.1 kHz file to a 96 kHz file (SRC). You may or may not experience a very minor improvement while recording new tracks (real instruments).

If you experience a radical improvement going from 44.1 kHz til 96 kHz in your A/D conversion - you need to get a better A/D because the filter at 44.1kHz is seriously messed up then.

With softsynths and many plug-ins you will actually get WORSE sound as they are not geared to this SR and will starting sounding very funny.

If you're not satisfied with your sound, sample rate is one of the very last places to starting looking. Look at your recording technique, your acoustics, your monitoring, your outboard gear, your plug-ins - and your mixing skills.
__________________


Online Mastering.dk

I have come here to chew bubblegum and kick ass... and I'm all out of bubblegum
Mr Keys
Thread Starter
#5
22nd February 2008
Old 22nd February 2008
  #5
Gear Head
 
Joined: Aug 2006
Posts: 47

Thread Starter
Mr Keys is offline
Quote:
Originally Posted by anguswoodhead View Post
I don't have all the answers but I have heard that Raid is NOT the go.
I have 2 x UAD cards so would be keen to know how they go at 96K - I have been thinking of making the same move.
I would say stay Mac (but I only own PC's - 4 of 'em)
I would suggest that going to 96K will def. sound better but have never tried.
I know some Mastering guys upsample everything then do all there processing and dither back to 44.1 right at the end.
Oh BTW when you import a piece of audio the software should give you the option to change it to the rate you're working at - so no smurfs etc.
yeah, we will stay mac....but the now Mac Pro, with Intel processors...!

Damn, it's hard to get a grip on the change....some say it's not worth it, and some say "go for it, you'll hear the difference"....maybe the only way to find out is to try....

but I still think there are a lot of guys/girls out there working with 96 (or higher), and there MUST be a reason for that, right? I wanna hear about it!!!!!


so, basicaly, when we get a project made on a laptop with a lot of softsynths and a few audiotracks with some samples, logic will ask if we wanna convert the samples to 96kHz when we open it in our studio....
cause that's how we work mostly, we get the complete projectfiles on a USB-memory, and then we open it up in the studio. (so the project will probably be in 44.1kHz, but we change it to 96 instead (will logic convert right away??).

what about EXS24? all the samplebanks there?will there be a conversion automaticly or will it just work?
#6
22nd February 2008
Old 22nd February 2008
  #6
Lives for gear
 
anguswoodhead's Avatar
 
Joined: Feb 2008
Location: On the East Coast of Australia
Posts: 3,861
My Recordings/Credits

Send a message via Skype™ to anguswoodhead
anguswoodhead is offline
If you are tracking new material (especially vocals) and your converters can handle it then 96K should be better.
As far as the samplers and software synths I don't know - just try it and see what happens - then let us know.
I would print all the software synths to an audio track (each part to a seperate track if you know what I mean) then switch off the synths - that will save you a lot of DSP, and then apply plugs to the audio tracks. etc.
Most samples are at 16/44.1 or 24/44.1 in my experience so no benefit to going up for there sake.
The other issue is - will your plug-ins work at 96K - some won't.
Good luck.
#7
22nd February 2008
Old 22nd February 2008
  #7
Lives for gear
 
jeremycox's Avatar
 
Joined: Dec 2007
Location: Brooklyn, NY
Posts: 1,648
My Recordings/Credits

jeremycox is offline
1) You cant combine sample rates in logic. In logic 8 you can import a 44.1 file into your audio bin and then convert it to 96k there. Ideally all content should be recorded at 96k.
2) The hard drive is going to be the weakest link in the configuration. Before purchasing 8gb of memory for a mac pro try 4gb. I have 5gb in mine and the memory is never fully utilized unless im doing video editing in FCP. Don't use raid. Upgrading to faster hard drives your best bet. I've been using 80gb raptor drives in my Mac Pro--the computer works way better under high loads especially with fast drives as the system disk and data disk. The factory drive in Mac Pros is not very fast--this upgrade will be very noticeable.
3)Way faster
4)The factory soft synths in logic sound better at 96k. I notice this the most with sculpture and ES2. 3rd party soft synths don't seem to sound any better at a higher sample rate. Not sure how to explain this.
5)Your new mac pro should be able to handle tons of tracks at 96k with no problem. I can do 40 or 50 in my dual core 2.66 mac pro.

With good conversion you will be able to tell the difference. Not sure how the waves converters sound so i cant advise to that. I can certainly tell the difference with my set up.


Jeremy Cox
#8
22nd February 2008
Old 22nd February 2008
  #8
Lives for gear
 
laser's Avatar
 
Joined: May 2003
Location: On the Road to Escondido
Posts: 682

laser is offline
Quote:
Originally Posted by Mr Keys View Post
Damn, it's hard to get a grip on the change....some say it's not worth it, and some say "go for it, you'll hear the difference"....maybe the only way to find out is to try....

but I still think there are a lot of guys/girls out there working with 96 (or higher), and there MUST be a reason for that, right? I wanna hear about it!!!!!
It depends on many factors whether it's worth it or not: your converters, the type of music you're recording/mixing, how picky you are.......

With my converters (RADAR), I find that delicate instruments in the higher frequency range (cymbals, higher registers on a piano, jangly accoustic guitars, female Minnie Ripperton-type singers) benefit tremendously from recording and mixing at 96kHz. Mid-rangey and punchy instruments (trumpets, distorted power chords from a Marshall Amp, male vocals ala Molly Hatchett) with my converters actually sound better to me at 48KHz.

What helped me decide was recording several instruments in the same enviroment at both 48kHz and 96kHz and putting the A/B samples on a CD. I played the CDs back on several players (my expensive home stereo, my cheap car stereo, mid-quality boom box as well as other various quality stereos owned by friends).

With YOUR converters in the style of music YOU record and mix, it may make a worthwhile difference----or, not. Only you can decide this, because your circumstances are unique. One rule of thumb--only apply SRC and dither once.

EDIT FOR CLARIFICATION: The SRC/dither I'm referring to is that used when burning to a CD format (16bit, 44.1 kHz), not SRC applied to soft synth files to enable mixing at a higher sample rate. I don't use softsynths, but I do subscribe to Lagerfelt's belief that SRC of any track before or while mixing has no sonic advantage and often sounds worse. If you're going to mix at 96 kHz, record or import files at 96kHz.


Good luck. Please share with us your results/decision.

Laser

Last edited by laser; 23rd February 2008 at 02:53 AM.. Reason: Clarification of my use of SRC
#9
23rd February 2008
Old 23rd February 2008
  #9
Lives for gear
 
Joined: Dec 2006
Posts: 1,015

MarsBot is offline
Go to 88.2 kHz. The math is much better when you have to convert back down to 44.1 for your CD or MP3 or whatever.
JPZ
#10
23rd February 2008
Old 23rd February 2008
  #10
JPZ
Lives for gear
 
Joined: Apr 2007
Posts: 574

JPZ is offline
Quote:
Originally Posted by MarsBot View Post
Go to 88.2 kHz. The math is much better when you have to convert back down to 44.1 for your CD or MP3 or whatever.
What he said and then of course put it on your ipod as an mp3 so you get the full benefit of that high sample rate(tic).
#11
23rd February 2008
Old 23rd February 2008
  #11
Lives for gear
 
Joined: Apr 2007
Posts: 2,913

Joe Porto is offline
It should be mentioned that the undeniable advantage of going from 44.1k to 96k is that your latency will be cut by more than half given the same buffer size. The biggest disadvantage is that your files will be twice as large, and your UAD-1 will only be able to process half the amount of plugins.

I run an 8-core Mac with 4MB ram and 2 UAD-1 cards. I can run my Symphony interface at 96k with a 32k buffer, giving me 1.5ms of latency from input to output. At 44.1k, it was over 3ms.

Any soft-samplers you use most likely have 44.1k samples. Logic will SRC them up to 96k. In my opinion, the quality loss is negligible, but it is an additional hit in CPU performance. Probably not an issue with a quad or 8-core.

I'm not sure if I hear a difference in quality between 44.1 and 96k, but the lower latency alone is worth the switch if you have the resources to do it. If you are concerned about loosing UAD-1 resources, it's not that big of a deal. Just freeze your UAD-1 tracks.
#12
23rd February 2008
Old 23rd February 2008
  #12
Lives for gear
 
Joined: Apr 2007
Posts: 2,913

Joe Porto is offline
Quote:
Originally Posted by MarsBot View Post
Go to 88.2 kHz. The math is much better when you have to convert back down to 44.1 for your CD or MP3 or whatever.
The math may be simpler, as the LCD is lower, but todays computers, and todays SRC algorithms can handle converting 96k to 44.1k just as well as going from 88.2k to 44.1k.
#13
23rd February 2008
Old 23rd February 2008
  #13
Lives for gear
 
peeder's Avatar
 
Joined: Aug 2006
Location: No longer participating here.
Posts: 6,705

peeder is offline
Quote:
Originally Posted by MarsBot View Post
Go to 88.2 kHz. The math is much better when you have to convert back down to 44.1 for your CD or MP3 or whatever.
Common, frequently repeated falsehood.
#14
23rd February 2008
Old 23rd February 2008
  #14
Lives for gear
 
jslevin's Avatar
 
Joined: Sep 2004
Location: Philadelphia
Posts: 2,819

jslevin is online now
Quote:
Originally Posted by Lagerfeldt View Post
It can never be an improvement in itself going from a 44.1 kHz file to a 96 kHz file (SRC). You may or may not experience a very minor improvement while recording new tracks (real instruments).

If you experience a radical improvement going from 44.1 kHz til 96 kHz in your A/D conversion - you need to get a better A/D because the filter at 44.1kHz is seriously messed up then.

With softsynths and many plug-ins you will actually get WORSE sound as they are not geared to this SR and will starting sounding very funny.

If you're not satisfied with your sound, sample rate is one of the very last places to starting looking. Look at your recording technique, your acoustics, your monitoring, your outboard gear, your plug-ins - and your mixing skills.
Goddam, that is just the best post, packed with common sense and not one iota of bullshit or hype, that I've seen here in a long time. And I quoted all four paragraphs, because each of the four bears repeating. Nice effing work, man.

JSL
Tui
#15
23rd February 2008
Old 23rd February 2008
  #15
Tui
Lives for gear
 
Joined: Nov 2006
Posts: 5,676

Tui is offline
96K only makes sense if you expect your music to get released on a format that can accommodate high sample rates, such as DVD.
#16
23rd February 2008
Old 23rd February 2008
  #16
Lives for gear
 
jeremycox's Avatar
 
Joined: Dec 2007
Location: Brooklyn, NY
Posts: 1,648
My Recordings/Credits

jeremycox is offline
Quote:
Originally Posted by Tui View Post
96K only makes sense if you expect your music to get released on a format that can accommodate high sample rates, such as DVD.
Not true....
#17
23rd February 2008
Old 23rd February 2008
  #17
Lives for gear
 
anguswoodhead's Avatar
 
Joined: Feb 2008
Location: On the East Coast of Australia
Posts: 3,861
My Recordings/Credits

Send a message via Skype™ to anguswoodhead
anguswoodhead is offline
Quote:
Originally Posted by jeremycox View Post
Not true....
Yeah - I must agree - everything I've read says all the number crunching etc is better at 96K (It's really all just maths) - work as long as possible at 96K - even give your mastering guy 96k files - let him do the dither as his last process - will sound better.
Tui
#18
23rd February 2008
Old 23rd February 2008
  #18
Tui
Lives for gear
 
Joined: Nov 2006
Posts: 5,676

Tui is offline
Quote:
Originally Posted by jeremycox View Post
Not true....
If he works mostly ITB, it is true. If he does a lot of mixing OTB, and needs to convert back and forth, higher sample rates might be beneficial. However, as has been remarked earlier, there are other factors that are more important, such as the quality of converters and mixing skills.
#19
23rd February 2008
Old 23rd February 2008
  #19
Lives for gear
 
jeremycox's Avatar
 
Joined: Dec 2007
Location: Brooklyn, NY
Posts: 1,648
My Recordings/Credits

jeremycox is offline
Quote:
Originally Posted by Tui View Post
If he works mostly ITB, it is true. If he does a lot of mixing OTB, and needs to convert back and forth, higher sample rates might be beneficial. However, as has been remarked earlier, there are other factors that are more important, such as the quality of converters and mixing skills.
Monitoring higher sample rates should result in better mix. With decent conversion it is EASY to notice the difference between 44.1 and 96.
#20
23rd February 2008
Old 23rd February 2008
  #20
Lives for gear
 
peeder's Avatar
 
Joined: Aug 2006
Location: No longer participating here.
Posts: 6,705

peeder is offline
Quote:
Originally Posted by anguswoodhead View Post
Yeah - I must agree - everything I've read says all the number crunching etc is better at 96K (It's really all just maths) - work as long as possible at 96K - even give your mastering guy 96k files - let him do the dither as his last process - will sound better.
Not true again. Plus a common confusion of dither and SRC.
JPZ
#21
23rd February 2008
Old 23rd February 2008
  #21
JPZ
Lives for gear
 
Joined: Apr 2007
Posts: 574

JPZ is offline
SRC is destructive. I try and avoid it.
#22
23rd February 2008
Old 23rd February 2008
  #22
Lives for gear
 
peeder's Avatar
 
Joined: Aug 2006
Location: No longer participating here.
Posts: 6,705

peeder is offline
Quote:
Originally Posted by JPZ View Post
SRC is destructive.
Also false!

I'm not even sure the latest Bob Katz book covers everything you need to know to steer clear of these tarpits of common and outdated misinformation. I've been thinking of adding a Mythbusters section to the tips and techniques thing....
#23
23rd February 2008
Old 23rd February 2008
  #23
Lives for gear
 
Joined: Apr 2007
Posts: 2,913

Joe Porto is offline
Quote:
Originally Posted by Tui View Post
96K only makes sense if you expect your music to get released on a format that can accommodate high sample rates, such as DVD.
Or if you want to track with virtual instruments and plugins with the lowest possible latency.

By the way, I hate the "if it's just going to end up on .mp3" argument. My personal findings, which are derived from real world recording, is that the .mp3 format is the biggest reason to use the highest quality a/d conversion and equipment when tracking and mixing, because the .mp3 compression process will inevitably degrade the original mix. Compress an already compromised audio file and you'll get a crap .mp3. Compress a high quality audio file, and at least you'll have a listenable .mp3.
JPZ
#24
23rd February 2008
Old 23rd February 2008
  #24
JPZ
Lives for gear
 
Joined: Apr 2007
Posts: 574

JPZ is offline
Quote:
Originally Posted by peeder View Post
Also false!

I'm not even sure the latest Bob Katz book covers everything you need to know to steer clear of these tarpits of common and outdated misinformation. I've been thinking of adding a Mythbusters section to the tips and techniques thing....
I would suppose it would depend on what sample rate converter you're using. Most people would probably use whatever their main DAW uses out of simplicity. Some DAWs do it better than others. Either way you're trusting a process that takes place after you've mixed thus you're not mixing what you're actually going to hear in the end...if you can hear it. I recently did a project that started off being archived off 2". We brought it in at 88.2 and at 44.1 using a Protools HD/Apogee combo. We also SRC'd the 88.2 to compare which was better because the project needed to bounce around to several systems one of which was PTLE and we didn't feel that guy's system would handle the higher sample rate well. We used Pro Tools tweakhead as the process. The consensus was that nobody could for sure tell the difference between the straight 88.2 and 44.1 transfers but everybody could tell the SRC'd audio. Sorry to go against what you say and I know this could all be material dependent as well but that is my recent real world experience with it.
#25
23rd February 2008
Old 23rd February 2008
  #25
Lives for gear
 
Joined: Apr 2007
Location: Sydney, Australia
Posts: 965

musikwerks is offline
"I don't have all the answers but I have heard that Raid is NOT the go."

Who said RAID is no good???? Raid is the best thing you could do IMO.

a) Raid 0 = way better read/write times.

b) Raid 5 = Better read/write times (but not as fast as Raid 0) AND complete redundancy for your data.

At 96khz, striping a Raid array will definitely help your data flow. To the guys who say no Raid.... care to explain why?
#26
23rd February 2008
Old 23rd February 2008
  #26
Lives for gear
 
peeder's Avatar
 
Joined: Aug 2006
Location: No longer participating here.
Posts: 6,705

peeder is offline
Quote:
Originally Posted by JPZ View Post
I would suppose it would depend on what sample rate converter you're using. Most people would probably use whatever their main DAW uses out of simplicity. Some DAWs do it better than others. Either way you're trusting a process that takes place after you've mixed thus you're not mixing what you're actually going to hear in the end...if you can hear it. I recently did a project that started off being archived off 2". We brought it in at 88.2 and at 44.1 using a Protools HD/Apogee combo. We also SRC'd the 88.2 to compare which was better because the project needed to bounce around to several systems one of which was PTLE and we didn't feel that guy's system would handle the higher sample rate well. We used Pro Tools tweakhead as the process. The consensus was that nobody could for sure tell the difference between the straight 88.2 and 44.1 transfers but everybody could tell the SRC'd audio. Sorry to go against what you say and I know this could all be material dependent as well but that is my recent real world experience with it.
Try Weiss Saracon, the Izotope SRC in AudioFile Sample Manager/Wave Editor or Izotope RX Advanced, or R8Brain Pro for modern sample rate converters that are only destructive wrt signal in the inaudible band (more or less...I challenge anyone to a blind test with a DAC that works well at both rates).
#27
23rd February 2008
Old 23rd February 2008
  #27
3 + infractions, forum membership suspended.
 
Joined: Jun 2007
Location: Salt Lake City, Utah
Posts: 2,378

dylansmale is offline
always remember that bit crushing isn't necessarily a bad effect, and enjoy the bigger sound i guess!
#28
23rd February 2008
Old 23rd February 2008
  #28
Lives for gear
 
anguswoodhead's Avatar
 
Joined: Feb 2008
Location: On the East Coast of Australia
Posts: 3,861
My Recordings/Credits

Send a message via Skype™ to anguswoodhead
anguswoodhead is offline
Quote:
Originally Posted by musikwerks View Post
"I don't have all the answers but I have heard that Raid is NOT the go."

Who said RAID is no good???? Raid is the best thing you could do IMO.

a) Raid 0 = way better read/write times.

b) Raid 5 = Better read/write times (but not as fast as Raid 0) AND complete redundancy for your data.

At 96khz, striping a Raid array will definitely help your data flow. To the guys who say no Raid.... care to explain why?
Just something I heard somewhere. I tried using Raid for a while but it didn't make that big of a difference. (And I'm only on a P4 3.2) Plus I guess it's just something else that can go wrong - if you know what I mean - KISS it.
soulstudios
#29
23rd February 2008
Old 23rd February 2008
  #29
soulstudios
Guest
 
Posts: n/a

Quote:
Originally Posted by Mr Keys View Post
Maybe I forgot to ask something, so, please enlighten me in what I need to think about before we go for 96 kHz?



Will we hear the difference?
For example: we used to record vocals with this setups: tlm49-vintagedesign CA73-summitaudio tla50-motu 2408mk3…now we use “Waves MaxxBCL”(everything bypassed though) instead of the motu, because of the converters, and I must say the difference is HUGE!!!
Will it be that kind of difference, or will we barely hear it?
Try recording a song in a session in 96k, then one (same band, setup etc) in 44k. Export at 192khz and resample down using a decent SRC algo. You won't regret it, and you won't go back, trust me. People are very good at talking as if they understand the maths- here and elsewhere.
Go with your ears, and get someone else's ears to do a blind test.
Good luck-
m@
#30
23rd February 2008
Old 23rd February 2008
  #30
Gear interested
 
Joined: Feb 2008
Posts: 1

antiacid is offline
I have noticed that you get better results upsampling 44.1 to 96. Then apply FX plugins and after that downsampling again to 44.1. I've read that's because lots of plugins don't upsample internally before doing their math. by upsampling manually you give the plugin more data to work with which results in a more accurate and smoother results (with a phaser for example). not sure though..
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
cleantone / Music Computers
20
Dunebuggy / Post Production forum!
15
AdamJay / Low End Theory
7
ISedlacek / High end
60

Forum Jump

SEO by vBSEO ©2011, Crawlability, Inc.