Quantcast
The difference between 44.1k and 96k is INSANE - Gearspace.com
The No.1 Website for Pro Audio
The difference between 44.1k and 96k is INSANE
Old 6th April 2011
  #1
Lives for gear
 
heisleyamor's Avatar
 
🎧 10 years
The difference between 44.1k and 96k is INSANE

Why did I not pay attention before? The plug-ins have a night and day difference when sampling up to 96k from a project recorded at 44.1k, especially with Altiverb. There was so much more clarity in the verb and it improved the mix greatly. I know there is a big debate about this, but damn...
Old 6th April 2011
  #2
Lives for gear
 
Lipps's Avatar
 
🎧 10 years
Quote:
Originally Posted by heisleyamor ➑️
Why did I not pay attention before? The plug-ins have a night and day difference when sampling up to 96k from a project recorded at 44.1k, especially with Altiverb. There was so much more clarity in the verb and it improved the mix greatly. I know there is a big debate about this, but damn...
What did it sound like after you printed it to cd at 16/44.1? Thats the true test. Record at all sample rates, dither and burn a disc at 16/44.1, THEN tell me what it sounds like.

Last edited by Lipps; 6th April 2011 at 03:21 AM..
Old 6th April 2011
  #3
Lives for gear
 
5 Reviews written
🎧 15 years
Quote:
Originally Posted by heisleyamor ➑️
Why did I not pay attention before? The plug-ins have a night and day difference when sampling up to 96k from a project recorded at 44.1k, especially with Altiverb. There was so much more clarity in the verb and it improved the mix greatly. I know there is a big debate about this, but damn...

96 is wider, has more depth.......
whatever the terminology, something is different about it
Old 6th April 2011 | Show parent
  #4
Gear Maniac
 
cianthreetimes's Avatar
 
🎧 10 years
I helped Eric Valentine last year with a comparative listening presentation he gave at AES. We did a myriad of tests – different converters, sample rates, ITB/OTB summing – and the results were really surprising.

The most important thing to learn is when comparing A vs. B, is to set them up properly so you really are just hearing the difference between the two.

Eric prepped a little document on how to really compare two sound sources: UnderToneAudio

This is definitely going to save me alot of money down the road by not buying into the hype and claims of many of the products out there.
Old 6th April 2011 | Show parent
  #5
Gear Guru
 
2 Reviews written
🎧 10 years
What were the results of the tests?
Old 6th April 2011 | Show parent
  #6
Gear Maniac
 
affe110's Avatar
 
🎧 10 years
Quote:
Originally Posted by Dean Roddey ➑️
What were the results of the tests?
Good question! :P

Anyway, what I want to know from OP is. Did it really work with upsampling the recordings or do you "have" to record it at a higher sample rate?
When I did my last session I wanted to record in 88.2 KhZ just for the sake of it but one of my interfaces wasn't compatible so I ended up with 44.1 anyway.

However, I use alot of the Stillwell plugins which sometimes has an "Oversample" button and it often makes the sound a little better.

This is interesting! Everytime I try recording at a higher sample rate something has stopped me, I'm still eager to find out the difference.
Old 6th April 2011 | Show parent
  #7
Lives for gear
 
foamboy's Avatar
 
🎧 10 years
Quote:
Originally Posted by affe110 ➑️
Good question! :P

Anyway, what I want to know from OP is. Did it really work with upsampling the recordings or do you "have" to record it at a higher sample rate?

You might find this interesting.Seems to work just in general.

fb


Running your software synths at higher sample rates
Old 6th April 2011 | Show parent
  #8
Lives for gear
 
symmetricalSound's Avatar
 
🎧 10 years
foobar has a really useful plugin to setup double blind listening tests foobar2000: Components Repository - ABX Comparator

it also has a plugin to let it use ASIO drivers
foobar2000: Components Repository - ASIO support
Old 6th April 2011 | Show parent
  #9
Lives for gear
 
filipv's Avatar
 
🎧 10 years
Good plugin should upsample the signal anyway. If you notice a difference when feeding it with 44 vs 96 then it's a lousy plugin.
Old 6th April 2011 | Show parent
  #10
Lives for gear
 
filipv's Avatar
 
🎧 10 years
Quote:
Originally Posted by robertshaw ➑️
96 is wider, has more depth.......
whatever the terminology, something is different about it
the exact term is "confirmation bias"
Old 8th April 2011 | Show parent
  #11
Lives for gear
 
5 Reviews written
🎧 15 years
Quote:
Originally Posted by filipv ➑️
Good plugin should upsample the signal anyway. If you notice a difference when feeding it with 44 vs 96 then it's a lousy plugin.
there wouldn't be any difference within the plugin itself, the word length is wider that's all. That's the difference, it's at 'machine' level
Old 8th April 2011 | Show parent
  #12
Moderator
 
psycho_monkey's Avatar
 
🎧 15 years
I'd rather have a good interface/converter at 44.1 than a prosumer one that can go to 192kHz.

But yeah, why not? if you've got the processing power to spare.
Old 13th April 2011 | Show parent
  #13
Gear Maniac
 
cianthreetimes's Avatar
 
🎧 10 years
Quote:
Originally Posted by Dean Roddey ➑️
What were the results of the tests?
Here are the results from the booth at AES in which we had a blind test between the conversion quality of a Lavry D/A unit vs. an M-Box 2 pro.

There were two days of testing. The first day, everybody knew which was was which. Note the comments on the left.

The second day, we removed the names of the units and replaced with A or B. A was the Lavry, B was the M-Box.

Prior to this, Eric and I did multiple tests trying different sampling rates, converters, clocks, external summing, etc, etc. I'll try to dig up these results and post them.

Again, not bashing anyone for being able to hear (or not hear) the differences between sample rates/converter units. We just were very surprised when we were doing these tests at much how much the marketing speak influences what we think sounds 'better'.
Attached Files
File Type: doc AES TEST RESULTS.doc (30.5 KB, 3957 views)
Old 13th April 2011 | Show parent
  #14
Jai guru deva om
 
warhead's Avatar
 
🎧 15 years
There is more space / depth / air with higher sampling rates for sure, 88.2 and 96k definitely deliver the resolution for that difference to occur.

We have outlined a converter test here that is fair and repeatable, and are starting to build a special system for it now that we can constantly add new converters to. Not a Clipalator test (those are 320k mp3) but we will eventually have a dedicated page with full wav files available.

I am very excited about this, and overall I agree that there can be a lot of hype surrounding perceived quality of conversion.

War
Old 13th April 2011 | Show parent
  #15
Lives for gear
 
Taurean's Avatar
 
🎧 15 years
The difference is 51.9K, that's not that insane
Old 13th April 2011 | Show parent
  #16
Lives for gear
 
emrr's Avatar
 
24 Reviews written
🎧 10 years
I think the most telling points arise when you are mixing in the analog realm, with multiple passes through the converter chain. I'll take higher versus lower, please.
Old 13th April 2011 | Show parent
  #17
Lives for gear
 
Ain't Nobody's Avatar
 
🎧 10 years
For converters on the way in, sure... but I'm still not getting how this rate thing works once you're in a mix session.

If everything is already recordedat 44.1, is there any point running the session at 96? (I'm assuming eventual bounce back to 44.1 for cd master.)

If you're using lots f plugs with high end processing like nebula, cpu intensive verbs, etc... don't they do it internally anyway? Is there a benefit for them as well?

About to start a new mixdown tomorrow, and using nebula now which has many 96k programs, so wondering if I should try something new and run the session at higher rate, or if it doesn't really matter since source material is all 44.1

What does it mean exactly anyway to be running at the higher rate? Is that a global setting within the daw? (logic) Does it need to be set per plug? Once and done?

Does the unevenness of the 96 to 44.1 (as opposed to thoretically simple halving of 88.2 to 44.1) cause unnecessary artifacts?
Old 13th April 2011 | Show parent
  #18
Lives for gear
 
🎧 10 years
cianthreetimes,

Thanks for posting your results. I was a little confused by how the columns are labeled. Did the bottom percentages mean that when people new which was which 80% preferred the Lavry, and when they didn't 78% preferred the Mbox?

Many Thanks!thumbsup
Old 13th April 2011 | Show parent
  #19
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by robertshaw ➑️
there wouldn't be any difference within the plugin itself, the word length is wider that's all. That's the difference, it's at 'machine' level
Huh?
Old 13th April 2011 | Show parent
  #20
Lives for gear
 
5 Reviews written
🎧 15 years
Quote:
Originally Posted by cianthreetimes ➑️
Here are the results from the booth at AES in which we had a blind test between the conversion quality of a Lavry D/A unit vs. an M-Box 2 pro.

There were two days of testing. The first day, everybody knew which was was which. Note the comments on the left.

The second day, we removed the names of the units and replaced with A or B. A was the Lavry, B was the M-Box.

Prior to this, Eric and I did multiple tests trying different sampling rates, converters, clocks, external summing, etc, etc. I'll try to dig up these results and post them.

Again, not bashing anyone for being able to hear (or not hear) the differences between sample rates/converter units. We just were very surprised when we were doing these tests at much how much the marketing speak influences what we think sounds 'better'.
what were the listening conditions? I'm sure you could hear a pin drop.....
Old 13th April 2011 | Show parent
  #21
Gear Nut
 
Filmusic's Avatar
 
🎧 10 years
i'm enjoying this

back only a few years ago (2008) i actually released a blues album recorded with an mbox pro...it got radio play...no one asked me how i recorded it...ha ha...i guess (i hope) it was the music that mattered...(fingers crossed)

Sound quality is secondary to nailing it musically...**** the numbers...we've had above cd quality in our grasp for a long time now (my three year-old records himself singing with an HD camera at 24/48)...so we've won the sound quality battle (within our home studios as far as recording machines go...HOW YOUR ROOM SOUNDS IS A DIFFERENT POINT OF CONCERN...but the battle of machines that offer above cd quality recording...well we got'em...this is great!!)

but does that really matter

(ducking here...cause i know it does to a ton-of-you)...but not to me

what does matter (time to tear up and get misty)

a song's emotion is the only thing that attarcts an audience to a song...nothing else...last i heard emotions wern't being ranked on numbers...bring on that gut feeling...like it or not...stop spending coin on gear hoping that's the sole solution...instead buy gear, BUT ensure you find/create great songs to record...cause that's what matters

ok ....i'll go away now...ha ha...what number was that...
Old 13th April 2011 | Show parent
  #22
Lives for gear
 
5 Reviews written
🎧 15 years
Quote:
Originally Posted by rcb4t2 ➑️
Huh?
yeah, what he said


Old 13th April 2011 | Show parent
  #23
Lives for gear
 
djanthonyw's Avatar
 
🎧 10 years
The thing that you have to understand is that if you record anything that is an analog source or acoustic such as an analog synth, voice, drums, ect... you're already recording at it's highest resolution. So when recording these sources 44.1 is fine.

When recording something such as a digital synth, if it's running at 44.1 it's not at it's highest resolution yet. The idea is to capture it's source at the highest resolution possible even if you are down sampling and working in 44.1. It's the same idea as recording that acoustic source. You want to emulate what the analog medium is doing and the best way to do that is have high resolution sources.

This is not about hearing frequencies that are un-audible to our ears, it's about resolution. The higher the frequency / resolution that the digital synth is running at, the more "naturally" the sound will be produced.
Old 13th April 2011 | Show parent
  #24
Lives for gear
 
Taurean's Avatar
 
🎧 15 years
Quote:
Originally Posted by djanthonyw ➑️
The thing that you have to understand is that if you record anything that is an analog source or acoustic such as an analog synth, voice, drums, ect... you're already recording at it's highest resolution. So when recording these sources 44.1 is fine.

When recording something such as a digital synth, if it's running at 44.1 it's not at it's highest resolution yet. The idea is to capture it's source at the highest resolution possible even if you are down sampling and working in 44.1. It's the same idea as recording that acoustic source. You want to emulate what the analog medium is doing and the best way to do that is have high resolution sources.

This is not about hearing frequencies that are un-audible to our ears, it's about resolution. The higher the frequency / resolution that the digital synth is running at, the more "naturally" the sound will be produced.
Becareful with words like "resolution" when addressing samping rate. With frequency response, to have full representation in capturing a signal and assuming a properly implemented A/D converter, 44.1 is all you need. When processing, it's a different story and a controversial one at that. Some say aliasing is a major problem and insist on using higher rates to filter out foldback aliasing. Resolution equates to bit depth and when it comes to bit depth the more you have the better when processing. For recording, 16-bit is where you actually have sufficient resolution. For processing however, 24-bit and higher yields better results. The math involved in DSP benefits from higher and higher bit depth as the "Word Length" builds to represent the product.
Old 13th April 2011 | Show parent
  #25
Lives for gear
 
ryanojohn's Avatar
 
🎧 10 years
Whatever the initial source sample rate may be, PLENTY of plugins sound better running at 88.2/96 than 44.1/48, particularly reverbs. Take D-Verb as the most extreme example, throw that on a snare in a 44.1 session, it (to me) sounds like digital garbage with weird flamming and such, up-sample the session to 88.2, hit play again, the D-Verb actually sounds like a reverb.

The source hasn't changed, but the processing has....

I honestly feel there's a significant difference in the sound that plugins have at different sample rates, but there's pretty much no way I could tell a 44.1 source apart from a 96k.

Would I be able to tell a 96k mix from a 44.1k mix, more likely, but each mix is different so there are far too many factors in that to be able to tell anything, outside of which mix you like better...
Old 13th April 2011 | Show parent
  #26
Lives for gear
 
giraffe's Avatar
 
🎧 15 years
i officially don't believe anything could make d-verb sound worth a f*^k.
Old 13th April 2011 | Show parent
  #27
Here for the gear
 
rObObnOxiOus's Avatar
 
🎧 10 years
i could not tell a difference in sample rates with my apogee ad/16x. but i could tell a huge difference between sample rates with the akai dps 24.
Old 13th April 2011 | Show parent
  #28
Gear Guru
 
🎧 15 years
Quote:
Originally Posted by robertshaw ➑️
what were the listening conditions? I'm sure you could hear a pin drop.....
the salient point is the listening conditions were the same in both tests
the difference being one test was labeled and the other test was blind.

it's not evidence of how good an mBox sounds, its evidence of how useless unblinded testing is
Old 13th April 2011
  #29
Gear Maniac
 
kcmoonshine's Avatar
 
🎧 10 years
I didnt read this whole thread, but jesus, 88.1 and 96khz sound so much better than 44.1. Right now im working on a song at 88.1 and the clarity is amazing, and so smooth.
Old 13th April 2011 | Show parent
  #30
Lives for gear
 
s34nsm411's Avatar
 
🎧 10 years
I hate you for making this thread...

I just spent some good time A/Bing exports of various layers of a project and it has opened my eyes quite a bit

the synths do open up a bit more but the biggest issue I have is with amp sim guitars... they sound VERY different

http://dl.dropbox.com/u/2093596/guitartest96.mp3
http://dl.dropbox.com/u/2093596/guitartest441.mp3

in these files i have line 6 podfarm with the cab sim turned off into an impulse, nothing more (guitars are doubled L and R)

as you can tell the sound is quite a bit different, the 96 k has more top end and sounds a little more open, but the 44.1 almost sounds like it has more body... cant tell which i like more for sure but the 96 does have a cool sound im not used to hearing from podfarm

edit: normalized both files to -3db
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 4616 views: 619522
Avatar for smoke
smoke 7th May 2021
replies: 55 views: 34332
Avatar for IM WHO YOU THINK
IM WHO YOU THINK 13th October 2020
replies: 98 views: 40730
Avatar for dfghdhr
dfghdhr 5th June 2021
replies: 1296 views: 186133
Avatar for heraldo_jones
heraldo_jones 1st February 2016
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