Gearslutz

Gearslutz (https://www.gearslutz.com/board/)
-   Gear Shoot-Outs / Sound File Comparisons / Audio Tests (https://www.gearslutz.com/board/gear-shoot-outs-sound-file-comparisons-audio-tests/)
-   -   What file was mixed @ 96k? (https://www.gearslutz.com/board/gear-shoot-outs-sound-file-comparisons-audio-tests/363269-what-file-mixed-96k.html)

Deleted 37a7726 5th February 2009 10:36 AM

What file was mixed @ 96k?
 
I'm curious to see what file is preferred. I'm once again weighing out the benefits of mixing @ 96k.
One file was mixed @ 48k exported to 44.1
The other was mixed @ 96k exported to 44.1
This is a rough drum mix with heavy processing for testing purposes. I've tried to make this test be as true as possible. No verbs, no modulation FX, & a identical setup.

Mix info.
Tracked @ 24/48
Processing = SSL Duende, URS 1980,1975, Saturation, & native sonar 8 channel eq.

test1.wav
test2.wav

What file sounds better?

MarkRB 5th February 2009 07:33 PM

Despite 1 file being nearly a dB louder. I think there is a clear difference. Good test. Cheers.

space2012 5th February 2009 08:31 PM

some plugins change too much at double or quad rate, specially EQs
if the test only had 1 plugin, could be easy to tell.

also if the source files were Sample.Rate.Converted for the test,
some softwares create distortion. see the src shootouts.

i would like to know more about how the source files comes from.
or how they were used for the tests.

original 96k downsampled with sonar pre-plugins?
twice recorded one at 96k other at 48k from tape machine?

anyway...
i like the highs in test1. they sound like 44/48k to me.
but the test2. mids & lows sounds more agressive "analog".

Deleted 37a7726 5th February 2009 08:45 PM

Both mixes used the same original 48k files. They were up-sampled to 96k on import. Up-sampling 48 to 96 is near lossless unlike down-sampling. Think of it as the same process that oversampling plugins use. Having the added bandwidth for processing lets distortions be spread out further into the inaudible spectrum. This is done often in mastering sessions as well.
This test is only to test mixing benefits, it dosen't deal with any tracking benefits.

MarkRB 5th February 2009 10:23 PM

I'm gonna bite the bullet and guess test1 is 96kHz, just because of the high end. It seems more open to my cloth ears. It's also the one I prefer by a shade.

I agree with Space re: the lows. They sound more upfront and aggressive (in a good way) in file 2.

I love this stuff. Thanks again RM.

Deleted 37a7726 5th February 2009 10:27 PM

No prob. I'm really curious to see what file ends up being more favored.

Deleted 37a7726 5th February 2009 10:28 PM

I'd love to read some of the reasons why the other voters picked what they did as well.

NiKaMusic 5th February 2009 10:59 PM

I have to be honest: I really didn't hear that much difference. But after listening with my eyes closed one after the other, I picked test 1. It seems to have more brightness...

Some engineers are saying 44.1 Hz is closer to analog than 96... Anyway, in the end it all ends up as an mp3 anyway, so why even bother... hittt

greetz,

NiKa

Ricey 6th February 2009 07:00 AM

i was thinking this was a trick - they both sound equally good. actually, sample rate issue aside, it sounds fekking great! first i thought either test 2 was louder or it was 96kHz. but i decided i preferred the sound of test 1 anyway! after listening again probably too many times i voted test 1 to be 96kHZ. i pull the trig and look, it's an even tie at this point!

Deleted 37a7726 6th February 2009 06:00 PM

Any last listeners before I give it away?

Zuewi 6th February 2009 06:25 PM

I vote against the others what spoken here and say test 2 is 96kHz. It sounds somehow more defined and has better hights and lows and is tigher.
test 2 leads now 6:5 stike
If test 2 isn't the 96kHz then the plugins may work better with 44,1 or the up and downsampling made some strange things.

I can't find 1 db more on a track??

Deleted 37a7726 6th February 2009 06:32 PM

Quote:

Originally Posted by Zuewi (Post 3888758)
I vote against the others what spoken here and say test 2 is 96kHz. It sounds somehow more defined and has better hights and lows and is tigher.
test 2 leads now 6:5 stike
If test 2 isn't the 96kHz then the plugins may work better with 44,1 or the up and downsampling made some strange things.

I can't find 1 db more on a track??

Yeah there shouldn't be any drastic db differences at all. Exact same session/plugin settings. Only the variables of the processing running 96k might cause a boost or dip in level.

NotVeryLoud 6th February 2009 08:23 PM

I have no idea which one was the 96k, but test2 sounded much better to me.

Not sure what sort of dynamic effect you were going for, but the pumping on test1 was way more aggressive, in a not-so-good, cheap sounding way.

The compression felt much more controlled and smooth in test2. The ambient (room tone?) at the start of the file sounds seemed much more pronounced as well.

MarkRB 6th February 2009 09:24 PM

Quote:

Originally Posted by Zuewi (Post 3888758)
I can't find 1 db more on a track??

Test 1 peaks at -2.1, test 2 at -1.3. But interestingly both have exactly the same RMS values. I guess this was matched by RM intentionally?

If you adjust both files to peak at the same level (or normalize them) then test 1 is perceptually louder.

Deleted 37a7726 6th February 2009 09:39 PM

I didn't change any settings.. It's the saturation, compressors, & eq responding differently with the different rates.

MarkRB 6th February 2009 09:46 PM

Quote:

Originally Posted by rhythminmind (Post 3889373)
I didn't change any settings.. It's the saturation, compressors, & eq responding differently with the different rates.

ah OK cool. It's odd that their RMS is identical down to 0.1 of a dB but their peaks are different by nearly a dB.
The question is, is the higher sample rate better representing the transients or the compression? Hmmmmm
The differences are far harder to hear at home on my RP5s than on my PMCs but I'm still a file 1 fan. (probably due to it being "louder"mezed. LOL but there you go).

Deleted 37a7726 7th February 2009 12:50 AM

Well here it is

test1.wav = 48k
test2.wav = 96k

I'm torn to pick one I prefer. I like the highs of test1, most likely do to the greater saturation. But the lows of test2.wav. I'm thinking with some re-adjustment of the compression i could recreate the feel of the 96k mix @ 48k.
So i'm thinking having the full 32 channels worth of my SSL duende (only 16 @ 96) will have a greater impact on my sound then mixing @ 96k.
Interesting experiment.

Let me also take note of the poll results before # test2 explodes do to me giving out the answer.
Pre Answer poll results.
test1.wav = 6
test2.wav = 8

MarkRB 7th February 2009 12:56 AM

Quote:

Originally Posted by rhythminmind (Post 3889928)
Well here it is

test1.wav = 48k
test2.wav = 96k

I'm torn to pick one I prefer. I like the highs of test1, most likely do to the greater saturation. But the lows of test2.wav. I'm thinking with some re-adjustment of the compression i could recreate the feel of the 96k mix @ 48k.
So i'm thinking having the full 32 channels worth of my SSL duende (only 16 @ 96) will have a greater impact on my sound then mixing @ 96k.
Interesting experiment.

Definitely interesting and cool test! Bit gutted I called the sample rate wrong but I stand by my favourite.
More when you're ready! ;) Maybe a recorded 96 vs 48? although that is tricky to set up.

thumbsup

Quint 7th February 2009 01:28 AM

Quote:

Originally Posted by NiKaMusic (Post 3886373)
Some engineers are saying 44.1 Hz is closer to analog than 96... Anyway, in the end it all ends up as an mp3 anyway, so why even bother... hittt

greetz,

NiKa

Really? Who? And I'm so sick of the whole "it all ends up on mp3 anyway" argument. That argument is as flawed an argument as I've ever had the displeasure of hearing (endlessly).

SFTPH 7th February 2009 03:06 AM

Your test was flawed from the beginning. The benefit of mixing at 96 or 88.2 is the fact you can RECORD at a higher sampling rate. Recording at 44.1 or 48 and then up- sampling to mix is just plain stupid. Why in the hell would someone purposely do that? Record and mix @ 96k and then down-sample to 48k and mix and there will be a huge-ass-no-contest result.

Deleted 37a7726 7th February 2009 03:13 AM

Quote:

Originally Posted by SFTPH (Post 3890264)
Your test was flawed from the beginning. The benefit of mixing at 96 or 88.2 is the fact you can RECORD at a higher sampling rate. Recording at 44.1 or 48 and then up- sampling to mix is just plain stupid. Why in the hell would someone purposely do that? Record and mix @ 96k and then down-sample to 48k and mix and there will be a huge-ass-no-contest result.

Learn how dsp processing works & you will learn why one would want to. Or talk to Bob katz he has more patience then I.

MarkRB 7th February 2009 03:19 AM

Quote:

Originally Posted by SFTPH (Post 3890264)
Your test was flawed from the beginning. The benefit of mixing at 96 or 88.2 is the fact you can RECORD at a higher sampling rate. Recording at 44.1 or 48 and then up- sampling to mix is just plain stupid. Why in the hell would someone purposely do that? Record and mix @ 96k and then down-sample to 48k and mix and there will be a huge-ass-no-contest result.

It was pretty well explained. I guess you missed it.
The tracks were recorded at a lower sample rate then upsampled (arguably a lossless conversion) for mixing.
Many people think there are benefits to giving plug-ins more samples to work with, and the results prove there are definitely differences.

No need to get upset, just listen to the files and take away what you can. *shrug

SFTPH 7th February 2009 04:29 AM

Quote:

Originally Posted by MarkRB (Post 3890297)
It was pretty well explained. I guess you missed it.
The tracks were recorded at a lower sample rate then upsampled (arguably a lossless conversion) for mixing.
Many people think there are benefits to giving plug-ins more samples to work with, and the results prove there are definitely differences.

No need to get upset, just listen to the files and take away what you can. *shrug

Yeah, I know how the test was poorly conducted. You need to read what I wrote. Who in the f*ck up-samples to mix. It is a stupid test. More samples are better so actually *record* more samples and don't up-sample. I understand what the OP was doing but it is useless in the real world. Just use up-sampling plugs if you insist on recording @ 48k, as it is the same thing. Plugins sound better at 96k but guess what? audio files recorded at 96k sound even better and have a much higher impact on mixes. Tests that have no place or merit in the real world are just silly, IMO of coursepeachh

MarkRB 7th February 2009 04:35 AM

Quote:

Originally Posted by SFTPH (Post 3890469)
Yeah, I know how the test was poorly conducted. You need to read what I wrote. Who in the f*ck up-samples to mix. It is a stupid test. More samples are better so actually *record* more samples and don't up-sample. I understand what the OP was doing but it is useless in the real world. Just use up-sampling plugs if you insist on recording @ 48k, as it is the same thing. Plugins sound better at 96k but guess what? audio files recorded at 96k sound even better and have a much higher impact on mixes. Tests that have no place or merit in the real world are just silly, IMO of coursepeachh

It's fairly commonly talked about. Get out more.

What exactly is an "upsampling plug"?

This whole thing was very well explained a couple of days ago, so you bitching about it now after showing no previous interest is irrelevant at best. I'm pretty sure no-one who got involved was as confused about it as you are.

It's only a "stupid test" if you are not interested in the results. It's pretty valid for me, particularly as it was an honest well conducted test here for a change.

If you don't care for the test method then don't take part and don't butt in spouting bollocks after the event.

SFTPH 7th February 2009 04:43 AM

Quote:

Originally Posted by MarkRB (Post 3890486)
It's fairly common. Get out more. It was very well explained. I'm pretty sure no-one involved was as confused about it as you are.

It's only a "stupid test" if you are not interested in the results. It seemed pretty valid to me.
Don't like it? Well.. you know the rest.

Are you a complete f*cking ****** or something, my brother, because you sure are acting the part. I GET THE TEST BUT IT IS A POINTLESS TEST! THE MAIN REASON YOU MIX AT HIGH SAMPLING RATES IS SO YOU CAN RECORD THE AUDIO AT HIGH SAMPLING RATES. PLUGINS ARE ONLY A SMALL PART OF MIXING.
WE ALL ALREADY KNOW PLUGS SOUND BETTER AT HIGHER SAMPLING RATES- IT IS COMMON KNOWLEDGE.

MarkRB 7th February 2009 04:53 AM

Quote:

Originally Posted by SFTPH (Post 3890505)
Are you a complete f*cking ****** or something, my brother, because you sure are acting the part. I GET THE TEST BUT IT IS A POINTLESS TEST! THE MAIN REASON YOU MIX AT HIGH SAMPLING RATES IS SO YOU CAN RECORD THE AUDIO AT HIGH SAMPLING RATES. PLUGINS ARE ONLY A SMALL PART OF MIXING.
WE ALL ALREADY KNOW PLUGS SOUND BETTER AT HIGHER SAMPLING RATES- IT IS COMMON KNOWLEDGE.

Haha. Why do I attract these people. mezed

Yes It's common "knowledge" but check the results of this test. You could even listen and decide for yourself instead of getting all your information from the internet.
Maybe read the thread, learn something, stop being a dickhead?

Deleted 37a7726 7th February 2009 05:10 AM

Some of us work in commercial studios. In a mixing/mastering house you receive elements from tracking studios or archives. Mixing out of house elements for example blue-ray your sometimes lucky to have DA88 stems nevermind a 24/96 protools archive. There are many worlds of audio other then the personal project studio. You would be amazed how many mixes start off with crap audio & end up as a 7.1 mix on latest & greatest format of the day..

SFTPH 7th February 2009 05:18 AM

Quote:

Originally Posted by MarkRB (Post 3890525)
Haha. Why do I attract these people. mezed

Yes It's common "knowledge" but check the results of this test. You could even listen and decide for yourself instead of getting all your information from the internet.
Maybe read the thread, learn something, stop being a dickhead?

hey *sshole- I listened to the both files and guess what? There was only a small difference compared to real world benefits of 88.2 or 96. I'm just trying to point out flaws that could potentially mislead the less informed.

MarkRB 7th February 2009 05:22 AM

Quote:

Originally Posted by SFTPH (Post 3890596)
hey *sshole- I listened to the both files and guess what? There was only a small difference compared to real world benefits of 88.2 or 96. I'm just trying to point out flaws that could potentially mislead the less informed.

Noted. But as pointed out several times before, this test is not about that. Even the "less well informed" can read, listen and make up their own minds.

SFTPH 7th February 2009 05:26 AM

Quote:

Originally Posted by rhythminmind (Post 3890574)
Some of us work in commercial studios. In a mixing/mastering house you receive elements from tracking studios or archives. Mixing out of house elements for example blue-ray your sometimes lucky to have DA88 stems nevermind a 24/96 protools archive. There are many worlds of audio other then the personal project studio. You would be amazed how many mixes start off with crap audio & end up as a 7.1 mix on latest & greatest format of the day..

I have NEVER been in or worked in a commercial studio that up-samples to mix. NEVER. The only time I've seen anything of the kind done is going from 44.2 or 48 to 2" tape AND THEN 88.2 or 96, but that is not the same thing at all. A mixed tune up-sampled for mastering is a different beast altogether. Thanks for the effort of your test though, my remarks where not meant as a personal attack.