The No.1 Website for Pro Audio
Is 96khz really worth the cpu ?
Old 9th July 2010
  #1
Registered User
 
🎧 10 years
Is 96khz really worth the cpu ?

I have been recording in 96 khz, just because i can lol. But i really don't hear that much of difference after mixing and mastering the track. Any input people's.
Old 9th July 2010
  #2
Lives for gear
 
ISedlacek's Avatar
 
1 Review written
🎧 15 years
If you don't hear the difference, then don't do it, very easy ... Why to ask others ? If they hear the difference, it will not help you ...
Old 9th July 2010
  #3
Lives for gear
 
Liquid Shadow's Avatar
 
🎧 10 years
For me it really depends on the music itself. For a heavily distorted guitar I doubt you'll hear any difference between 44KHz and 96KHz. I, for example, records almost exclusively classical music, so I use 96KHz.
Old 9th July 2010
  #4
Lives for gear
 
🎧 10 years
I used to work at 96K when I first started out for about a week. But then I realised the extra space it was eating on my drives for no audible gain.

So I quickly reverted to working at either 44.1K or 48K.
Old 9th July 2010 | Show parent
  #5
Registered User
 
🎧 10 years
The benefit may be twofold:

1) Your converters may or may not perform better run at 96K.
2) Nonlinear DSP processes may avoid aliasing distortion when run at 96K, if they don't already have another means of avoiding it at 44.1.

You may run low on disc space, DSP power, drive/interface bandwidth and memory however. Also some VI's may have weak internal SRC algorithms that make them sound worse when not run at their native speed.

If there was one answer for everybody at all times we'd tell you.
Old 9th July 2010
  #6
Registered User
 
🎧 10 years
Quote:
Originally Posted by Scotty Traxx ➑️
I have been recording in 96 khz, just because i can lol. But i really don't hear that much of difference after mixing and mastering the track. Any input people's.
I did it in the early 2000's but after a few years I realized it's not worth it. It brings in more problems than it solves. Plus you can't control what you can't hear.

24-bit / 44.1 kHz is enough imo, they'll listen to it as some shi*ty mp3 anyway
Old 9th July 2010 | Show parent
  #7
Lives for gear
 
ISedlacek's Avatar
 
1 Review written
🎧 15 years
Quote:
Originally Posted by StereoPari ➑️

24-bit / 44.1 kHz is enough imo, they'll listen to it as some shi*ty mp3 anyway
So why not to work directly in mp3 ?
Old 9th July 2010 | Show parent
  #8
Registered User
 
9 Reviews written
🎧 10 years
I used to work exclusively at 88.2kHz but anymore I just use 44.1khz. It's only Rock n Roll
Old 9th July 2010 | Show parent
  #9
Registered User
 
🎧 10 years
Quote:
Originally Posted by ISedlacek ➑️
So why not to work directly in mp3 ?
Bit-depth is not the same as "frequencies which are not there". The vast majority of listeners do not care about frequencies over 16-18 kHz, as evidenced by their consumer behavior.

Plus, the Nyquist theorem states that you need to sample only twice / waveform needed to be recreated; no use in sampling a 20 kHz wave 5 times, 2.0000001 times is enough in order to recreate it perfectly.
Old 10th July 2010 | Show parent
  #10
Lives for gear
 
Matti's Avatar
 
2 Reviews written
🎧 15 years
Quote:
Originally Posted by spicemix ➑️
The benefit may be twofold:

1) Your converters may or may not perform better run at 96K.
2) Nonlinear DSP processes may avoid aliasing distortion when run at 96K, if they don't already have another means of avoiding it at 44.1.

You may run low on disc space, DSP power, drive/interface bandwidth and memory however. Also some VI's may have weak internal SRC algorithms that make them sound worse when not run at their native speed.

If there was one answer for everybody at all times we'd tell you.
You have a point there at 1) and overall

Matti
Old 10th July 2010
  #11
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by Scotty Traxx ➑️
I have been recording in 96 khz, just because i can lol. But i really don't hear that much of difference after mixing and mastering the track. Any input people's.
If you use, or are looking to use, a lot of plugs..... its worth having a quick look at this thread. Essentially working at higher sample rates improves the aliasing of plugs measurably.

https://gearspace.com/board/so-much-...is-thread.html

Last edited by dr_Jezz; 10th July 2010 at 07:07 PM.. Reason: typo
Old 10th July 2010 | Show parent
  #12
Registered User
 
🎧 10 years
Smile Very Good Replies....

I agree with most, esp. on 44.1 recording. Also as one post stated, not only does the CPU take a bit more of a hit. Most importantly you need a fast dedicated Hard Drive with lots of space. And a good interface, for its DSP etc..

I always recommend using a dedicated external, or internal if you have a tower, Hard Drive. This makes multi-track high rate recording much much smoother and in some cases possible...
Old 10th July 2010 | Show parent
  #13
Registered User
 
🎧 10 years
Work at 88.2 or higher if you want your plug-ins to perform better and to preserve your audio.
It is better to track and work in 88.2 or higher and then convert to 44.1 than it is to start in 44.1.

I won't go in depth, look it up if you want to know the theory behind my advice.

Cheers!
Old 10th July 2010 | Show parent
  #14
Lives for gear
 
Matti's Avatar
 
2 Reviews written
🎧 15 years
Your Daw program is working floating point maths.

Matti
Old 10th July 2010 | Show parent
  #15
Gear Head
 
🎧 10 years
Not worth it.
Old 10th July 2010 | Show parent
  #16
Registered User
 
🎧 10 years
Quote:
Originally Posted by dr_Jezz ➑️
If you use, or are looking to use, a lot of plugs..... its worth having a quick look at this thread. Essentially working at higher sample rates improves the aliasing of plugs measurably.

https://gearspace.com/board/so-much-...is-thread.html

Many plugins do sound better at 96khz, this is one of the reasons i work at 96khz. But there are several scientifically provable reasons that 96khz is great to work at and superior to 44khz for some. Latency is also greatly reduced at 96khz. With my Core i7 920 i'm able to have a 2.6 ms 'Roundtrip' Asio latency at 96khz to compliment my zero-latency SCOPE dsp enviroment. When doing complex REALtime-Routing in-and-out of Native DAWs and hardware the less latency you have with Asio buffer-sizes, Wave and converters the better. Also with complex chains of effects, etc less latency is best.

Another reason for 96khz is if you wish to put your music on higher quality mediums in the future such as dvd-audio, etc. My current project is being made for dvd-audio and i believe it is superior soundwise to cd's and mp3's BUT even if some people can't hear the difference between dvd-audio and cd's there still is superiority in dvd-audio in terms of storage space for data and other bells and whistles.

Alot of people talk as if music is only being made for mp3's and that's just not true, to my knowledge physical media still out sale virtual. But even if mp3's where the only option, the low quality sound of mp3's suggest there is even more importance to produce music with as high as quality as possible to better handle the down conversion to mp3's. If you produce at higher quality the mp3 should sound better.

Actually the Music industry is moving more away from any storage Mediums for main products and the bulk of the money is now made with Live performances. Madonna was the highest paid Music Star earning any where from 240-280 million dollars in 2009 and the overwhelming majority of that money came from touring. Here's a link to many artists statistics for touring, it'll show where all the money in the Music business is now being made - Madonna Tops 2009 Music Money Makers List | Billboard.com

The top money making groups and solo artists make most their money touring and some don't have new hits on the latest charts and some don't even have new albums out, but these are often older groups and individuals like Barbra Streisand (2006), Neil diamond, etc. But newer acts with recent releases that are top money getters are listed aswell as getting most their money threw touring. live performance is most important now in the Music industry (low-latency, and great sound quality for most venues).




EDITED
Old 10th July 2010 | Show parent
  #17
Lives for gear
 
Matti's Avatar
 
2 Reviews written
🎧 15 years
Why this old depute again, my mistake as well

Matti
Old 11th July 2010 | Show parent
  #18
Registered User
 
🎧 10 years
Just thought I'd say my opinion once more...

Whatever benefit you might get from 96 kHz when using lots of plug-ins does in no way compensate for the shortcomings of a 96kHz system as a whole.

"Using" 96kHz means that EVERY aspect of your signal chain should be able to cope with it, including your microphones, preamps, etc...and most importantly, your ears. You can't control what you can't hear --why bother recording some super-fast transients in the dog area, if you really have no idea about its sonical qualities?

Making music is all about making intentional sonical arrangements which you can control. I understand the use of 96kHz or 192kHz for scientific purposes, but to me they both seem a bit unnecessary for the stuff most of us are doing.

And if someone wants to capture the "true" sound of source X --including its harmonics which extend well to the ultrasonic area-- why stop at 96?

Why not use 384 kHz, just to make sure that Sabian crash cymbal really gets "properly documented"?

thumbsup
Old 11th July 2010 | Show parent
  #19
Registered User
 
🎧 10 years
i am sticking it out with 48k/24bit on a MIX system.
having made several records at 96k, i will say that it is audible on high end material like cymbals, and on the occasional long reverb tail.
For the type of music i work with (rock/punk) i don't think its an issue.
Old 11th July 2010
  #20
Registered User
 
🎧 15 years
Lightbulb

192 is the new 96.

Anyone Record At 192k Via Firewire?
Old 11th July 2010 | Show parent
  #21
Gear Maniac
 
Bretster's Avatar
 
🎧 10 years
I guess I see it as the soft sound vs. hard sound.

96k is great if you hear a difference. I hear a very recognizable difference in plug-ins.

You got the rockers that don't give a sh!t and proclaim that its a waste of cpu and you have the soft melancholics that proclaim their are differences.

Choose your poison, either way you can use both to your advantage when doing a project. Why always do the same thing when you have invested in a versatile recording setup?

I hear a difference with 88.2/96k and use it when I want less latency and a harmonically rich mix.
Old 11th July 2010 | Show parent
  #22
Registered User
 
🎧 15 years
Quote:
Originally Posted by StereoPari ➑️
Bit-depth is not the same as "frequencies which are not there". The vast majority of listeners do not care about frequencies over 16-18 kHz, as evidenced by their consumer behavior.

Plus, the Nyquist theorem states that you need to sample only twice / waveform needed to be recreated; no use in sampling a 20 kHz wave 5 times, 2.0000001 times is enough in order to recreate it perfectly.

I dont believe the motivating factor with MP3 success is people not caring about quality.. I think its been more of the trade off between quality versus bandwidth limitations..

We have ended up with a landscape of crap quality recordings out there due to crap bandwidth and storage options dictating what gets delivered.. MP3 started growing in 95-96 and back then MP3 was easily the best option for bandwidth limited applications. Nowdays that is diminishing and i think as bandwidth continues to get better around the globe and formats like FLAC or other lossless types gain more popularity i am expecting things will slowly return to higher quality formats..

Your average person may not be as tuned in to the quality differences but they will take the better option if its easy and its becoming easier for them to do so.. I dont think we are stuck with MP3 forever.. But we may be stuck with overcompressed mastering for a while.
Old 11th July 2010 | Show parent
  #23
Gear Maniac
 
Bretster's Avatar
 
🎧 10 years
Quote:
Originally Posted by author ➑️
Awesome thread, thanks for posting!
Old 11th July 2010 | Show parent
  #24
Registered User
 
🎧 10 years
Quote:
Originally Posted by ebulb ➑️
I dont believe the motivating factor with MP3 success is people not caring about quality.. I think its been more of the trade off between quality versus bandwidth limitations..

We have ended up with a landscape of crap quality recordings out there due to crap bandwidth and storage options dictating what gets delivered.. MP3 started growing in 95-96 and back then MP3 was easily the best option for bandwidth limited applications. Nowdays that is diminishing and i think as bandwidth continues to get better around the globe and formats like FLAC or other lossless types gain more popularity i am expecting things will slowly return to higher quality formats..

Your average person may not be as tuned in to the quality differences but they will take the better option if its easy and its becoming easier for them to do so.. I dont think we are stuck with MP3 forever.. But we may be stuck with overcompressed mastering for a while.
I understand what you're saying, and I agree the whole mp3 thing is slowly moving on. However, I think the last ten years have - to some extent - proved that most listeners choose price and portability over quality.

(One interesting question is this: now we are trying to emulate the tape/vinyl character in a medium which nature is not meant for it (ie. we want to hear nonlinear responsiveness in a linear system). As the lossy digital formats have now been very popular for over a decade, I think many people've become accustomed to that squashy "mp3 sound". Maybe, 15 years from now on, we are trying to emulate something else than tape... )
Old 11th July 2010 | Show parent
  #25
Lives for gear
 
doorknocker's Avatar
 
1 Review written
🎧 20 years
Quote:
Originally Posted by Liquid Shadow ➑️
For me it really depends on the music itself. For a heavily distorted guitar I doubt you'll hear any difference between 44KHz and 96KHz. I, for example, records almost exclusively classical music, so I use 96KHz.
I never understood this. Why should sound quality be genre-dependent? It kind of assumes that 'high art' needs high-resolution why popular music can do with lower resolution. But wait, isn't that why we have mp3s in the first place?

True, a distorted guitar might not fill out the whole frequency spectrum in the same way a full orcehstar does but I rather have high-resolution print of a picture even if the picture itself isn't super-sharp or even blurred (on intent).

I'm perfectly fine with 44.1Khz so I'm not saying that you necessarily need higher resolution but I argue against the genre bias in the same way that I would argue against 'classical' vs. 'rock' preamps, mics, etc

It's all music!
Old 11th July 2010 | Show parent
  #26
Registered User
 
jwleeman's Avatar
 
🎧 10 years
Quote:
Originally Posted by StereoPari ➑️
24-bit / 44.1 kHz is enough imo, they'll listen to it as some shi*ty mp3 anyway
Heh, well he's got a point! But it's always a good idea to record at as high a sample rate as possible, because professional mastering equipment and engineers can more effectively process your track. And the difference IS audible, provided you're listening in a lossless format, of course.

~Jeff
Old 11th July 2010 | Show parent
  #27
Lives for gear
 
marchhare's Avatar
 
🎧 15 years
Not worth it at all. 44.1
Old 11th July 2010 | Show parent
  #28
Lives for gear
 
Taurean's Avatar
 
🎧 15 years
Starting at a quality setting like 44.1Khz/24-bit to record, into mixing, and dithering down or rendering to MP3 is not the same as doing this entire process starting at the setting of an MP3. The MP3 argument is just silly.

The genre argument is silly. If anything rock n' roll and many "modern" styles of music could potentially have more grit and distortion to them in which by the faulty logic leading this genre perspective should need the higher sample rates as opposed to classical.

Overall, 44.1 Khz, with quality converters, is perfect for recording. Making converters work faster will only introduce other problems and artifacts. Even for processing while mixing, it's very very situation dependent and I will go as far as to say most cases 44.1 again is just fine. While higher sampling rates can help with aliasing, it seems you would need a much higher rate to really help any, impractical local rates. Oversampling is the best bet here, and done properly within the plugin itself. The differences some talk about with higher rates such as 96 Khz, whether in recording or processing, is not necessarily better, just different. The differences are probally more due to conversion processes than anything else.
Old 11th July 2010 | Show parent
  #29
Lives for gear
 
Lee Cardan's Avatar
 
🎧 10 years
Quote:
Originally Posted by StereoPari ➑️
Bit-depth is not the same as "frequencies which are not there". The vast majority of listeners do not care about frequencies over 16-18 kHz, as evidenced by their consumer behavior.

Plus, the Nyquist theorem states that you need to sample only twice / waveform needed to be recreated; no use in sampling a 20 kHz wave 5 times, 2.0000001 times is enough in order to recreate it perfectly.
whooaaa fella, hold your horses tutt

nyquist, yep sample twice the frequency to capture it, but if you sample 22KHz at 44.1KHz you will get nothing more than a messed up triangle wave.. that equates to a whole lot of missed information as of course the music we record is rarely a pure sine wave even
Old 11th July 2010 | Show parent
  #30
Lives for gear
 
dub3000's Avatar
 
🎧 15 years
Quote:
Originally Posted by Lee Cardan ➑️
whooaaa fella, hold your horses tutt

nyquist, yep sample twice the frequency to capture it, but if you sample 22KHz at 44.1KHz you will get nothing more than a messed up triangle wave.. that equates to a whole lot of missed information as of course the music we record is rarely a pure sine wave even
the reconstruction filter will turn that triangle back into a sine wave.

also, (and this was something that took me a while), if you were sampling a 22khz triangle or square wave, there are components of that signal above 22khz that will be lost. that's fine, because your ear can't pick up on those anyway. a 20khz sine wave and a 20khz square wave sound the same to a human ear when played back on a good playback system, and they'll be represented by the same thing in 44khz digital.
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 57 views: 25695
Avatar for gpiccolini
gpiccolini 12th May 2016
replies: 15929 views: 1610923
Avatar for Ragan
Ragan 11th January 2019
replies: 1513 views: 285095
Avatar for stillthinking
stillthinking 3 weeks ago
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump