The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
96K really sound better? Studio Monitors
Old 7th January 2009
  #91
Motown legend
 
Bob Olhsson's Avatar
 

It's all about keeping processing artifacts OUT of the audible range below 20kHz.

Where 2x sample rate conversion has a potential edge is in real time converters because both sample rates can use the same clock which eliminates a source of jitter and may allow the use of shorter, less computationally challenging filters. On the other hand some real time converters have fewer artifacts at certain sample rates which is a greater advantage.

All engineering is about managing trade offs. Which trade offs are best comes right back around to using your ears as badly as we all simply want to know what's best. I know that a certain combination of settings and gear gets what I think are better sounding results than another combination but no global statement of "X" being better than "Y" has any real meaning outside of some specific combination of gear because there can be no perfect gear.
Old 11th January 2009
  #92
Gear Head
 
Rev. JimBo's Avatar
 

Quote:
Originally Posted by peterwild View Post
Yeah -- we don't need to see graphs showing potential for better sound .
This would seriously threaten our ability to use Pro-tool plug ins in our pretend virtual world while imagining that we are using a real compressor
and are kicking serious Ass. Take your Apostasy down the road pal ----
Old 11th January 2009
  #93
Lives for gear
 

I hope you all realize that in the above graph the "analog" sample does not represent analog recording (tape, LP whatever) but just a short voltage burst containing only frequecies starting from around 333000 Hz and higher. NOTHIG to do with audible signals.
Old 12th January 2009
  #94
Gear Maniac
 

Hey guys and gals. I have an API A2D preamp+ A/D converter. I run protools 7.4 (can't wait for PT 8!) but when I changed the sample rate on my A2D, I got a lot of latency. Seeing is my converters are pretty damn good, I want to benefit recording at a higher sample rate if it's worth it, but how do I make the latency go away in my Protools session? So I've just been recording at 44.1. I track everything at my home/project studio, but I will be sending my stuff out for a pro to mix. Anyone know how to get around my problem? Thanks! Hope I didn't hi-jack the thread.
Old 12th January 2009
  #95
Lives for gear
 

Quote:
Originally Posted by peterwild View Post

Quote:
Originally Posted by petrus
I hope you all realize that in the above graph the "analog" sample does not represent analog recording (tape, LP whatever) but just a short voltage burst containing only frequecies starting from around 333000 Hz and higher. NOTHIG to do with audible signals.

hi,

are you confusing periodic signals with impulses? they are used for different purposes i think.


The Delta Function

right.
Old 12th January 2009
  #96
Lives for gear
 

The pic is still confusing for the one with limited understanding.

There will never be a signal with that fast risetime into a converter.

Also I think the DSD example is without a LP filter. A real world 2.8MHz DSD system will look more like the 192kS/s PCM impulse if memory serves me.

If you lowpassfilter the impulse so that the spectra is more representative of music before it enters the digital system the picture would be very different.

Yes, an impulse tell us lot about the bandwith and time behaviour of a system but it's easy to misinterpret the results.


/Peter
Old 12th January 2009
  #97
Gear Nut
 

My personal experience is that I noticed a difference in the top end when comparing 24 bit 44100 to 48000. I found a noticable improvement in the top end of the 48000 sample rate. When I then upsampled to 96, I battled to distinguish between the 2 and hence settled on 48000 being the more storage friendly of the 2. To each his own tho...
Old 12th January 2009
  #98
Lives for gear
 

Quote:
Originally Posted by Audiop View Post
The pic is still confusing for the one with limited understanding.

There will never be a signal with that fast risetime into a converter.

Also I think the DSD example is without a LP filter. A real world 2.8MHz DSD system will look more like the 192kS/s PCM impulse if memory serves me.

If you lowpassfilter the impulse so that the spectra is more representative of music before it enters the digital system the picture would be very different.

Yes, an impulse tell us lot about the bandwith and time behaviour of a system but it's easy to misinterpret the results.


/Peter

hi,

yeah pcm is probably better [less noisy?]. i think the pyramix device is 5.6MHz / 384 kHz, though. i don't know anyone who has ever seen one in real life. maybe they do not really exist.

i think the point they are trying to make is that impulse response is better within the typical audio band with higher sample rates. the amplitude of the impulse response should be consistent / similar throughout the bandwidth, no? i don't see how the rise time is relevant to the amplitude measurement. how you are going to lowpass filter a 10 microsecond pulse?

I don't know how else they are supposed to graph it [i'm not a math guy]. isn't a 10 microsecond pulse what they usually use? I thought the impulse had to be faster than the degree of accuracy of the system to qualify as an impulse in a practical sense.


right.
Old 12th January 2009
  #99
Lives for gear
 
taturana's Avatar
i track at 24/48 and bounce it at 32/96, that way i get the best of both worlds, low cpu and use the plugins at max resolution.

in 44.1 i think the lpf is too damn close to the hearing limit... at 48 i think things sound a bit better.

then i do SRC with a good program like SoX or R8..

but regardless of the SR, in 16/44.1 the result sounds quite similar.

and 192k is simply useless.. for me at least. i really can't hear any diference from 96 to 192. and the cpu use/file sizes are impractical.
Old 12th January 2009
  #100
Lives for gear
 
plexisys's Avatar
 

With all the above said, why would one buy a 192k i/o unit vs a 96k i/o unit?
Old 12th January 2009
  #101
Lives for gear
 
Tube World's Avatar
Lets say you record an acoustic guitar and vocal at 96, and then add some EQ , compression, and reverb. Then do the same thing at 44, and compare the files; are some of you guys truly say you don't hear the nicer top end at the 96 recording? If so then your converters and your room don't work and sound the way they do in my studio. To me this is not even a grey area, 96 sounds better (when you leave it at 96 on the comparrison). But after saying that, there are issues that come into play that affect the outcome of a comparrision once you bring the 96 recording down to 44.

The first is the quality of your SRC, as this will effect your 96 recording when you bring it down to 44. If your only using a ok SRC like the one in Wavelab, it won't sound as good as when you use something like the Weiss or even the 8Rbrain Pro.

The second problem I see is in a full blown song recorded at 44 and 96, it would be impossible to get the same recording. First because you won't get the same performance from the band members. Secondly the volume of each track won't be exactly the same, which would then change how your effect levels affect the song. So by the time you bring down a 96 song to 44, even by 1 db of volume difference on the drums, bass, or vocal would affect what your hearing. So it's hard to get an accurate test. With only two tracks on my guitar, and vocal test, it is easier to get them to be closer since your only working with two tracks, not a full band.

3. I believe Bob Katz said in his mastering book (and other's here) that the filters in our hearing range on our converters are different at 96 compared to 44 which is what why we hear a difference. Then depending on the filters on each different converter brand will also affect what your hearing.

To me, 96 gives a better top end, so I will use with the best SRC I can afford to keep that better quality as best I can.
Old 12th January 2009
  #102
Lives for gear
 

Quote:
Originally Posted by plexisys View Post
With all the above said, why would one buy a 192k i/o unit vs a 96k i/o unit?

hi,

i think you have the question backwards, right?


right.
Old 12th January 2009
  #103
Lives for gear
 
plexisys's Avatar
 

No, if you are really doing everything at 96k and either staying at 96k or 44k why would you buy a 192k?
Old 12th January 2009
  #104
Gear Maniac
 

Quote:
Originally Posted by newlyformedmind View Post
Hey guys and gals. I have an API A2D preamp+ A/D converter. I run protools 7.4 (can't wait for PT 8!) but when I changed the sample rate on my A2D, I got a lot of latency. Seeing is my converters are pretty damn good, I want to benefit recording at a higher sample rate if it's worth it, but how do I make the latency go away in my Protools session? So I've just been recording at 44.1. I track everything at my home/project studio, but I will be sending my stuff out for a pro to mix. Anyone know how to get around my problem? Thanks! Hope I didn't hi-jack the thread.

Anyone?
Old 12th January 2009
  #105
Lives for gear
 

Quote:
Originally Posted by plexisys View Post
No, if you are really doing everything at 96k and either staying at 96k or 44k why would you buy a 192k?

hi,

oh, i see. well then you should record everything at 192kHz, right?


right.
Old 12th January 2009
  #106
Lives for gear
 

Quote:
Originally Posted by newlyformedmind View Post
Hey guys and gals. I have an API A2D preamp+ A/D converter. I run protools 7.4 (can't wait for PT 8!) but when I changed the sample rate on my A2D, I got a lot of latency. Seeing is my converters are pretty damn good, I want to benefit recording at a higher sample rate if it's worth it, but how do I make the latency go away in my Protools session? So I've just been recording at 44.1. I track everything at my home/project studio, but I will be sending my stuff out for a pro to mix. Anyone know how to get around my problem? Thanks! Hope I didn't hi-jack the thread.
hi,

you changed from what sample rate to what other sample rate? you should get less latency the higher your system's sample rate, right?


right.
Old 12th January 2009
  #107
Lives for gear
 

Quote:
Originally Posted by newlyformedmind View Post
Hey guys and gals. I have an API A2D preamp+ A/D converter. I run protools 7.4 (can't wait for PT 8!) but when I changed the sample rate on my A2D, I got a lot of latency. Seeing is my converters are pretty damn good, I want to benefit recording at a higher sample rate if it's worth it, but how do I make the latency go away in my Protools session? So I've just been recording at 44.1. I track everything at my home/project studio, but I will be sending my stuff out for a pro to mix. Anyone know how to get around my problem? Thanks! Hope I didn't hi-jack the thread.
hi,

maybe you forgot to change the sample rate in your session as well as on your converters?


right.
Old 12th January 2009
  #108
Lives for gear
 
dangoudie's Avatar
 

As far as I can recall the main reason higher sample rates may sound better is because of high frequency smear caused by less-than-perfect LPF-ing and anti-aliasing. which manifests itself below Nyquist.

So the point being that at 44.1k (where Nyquist is 22.05k) there ain't far to go until the artifacts are in the audible spectrum. Therefore raising the sample rate even just to 48k will lift most of this problem out of audible range and clean up the high frequencies.

In my experience, I can hear a difference and smoother is a good descriptor.

As is said previously, rock bands can lose some grit at higher sample rates so whatever suits.

At the end of the day I think this is a REALLY minor issue and there are a huge number of factors that will impact more greatly on a recording.
Old 12th January 2009
  #109
Gear Addict
 

Quote:
Originally Posted by surflounge View Post
EMcasts has an interview with producer/engineer/musician Joel Hamilton (Soulive, Ludiacris, Tom Waits) discussing your point about when he uses 44.1 vs 96k
EM Podcast Interviews | Audio interviews with professional musicians, producers, audio engineers about mixing, recording, producing | Hear audio engineering, mixing, production techniques podcasts
I remember listening to this, and Joel was basically downplaying 96k. Then I listened to the Chuck Ainley interview from the week prior, and he was talking about how he was amazed that people didn't hear the difference. And he always tracks at 96k.
Old 16th January 2009
  #110
Lives for gear
 
DSD_Mastering's Avatar
Quote:
Originally Posted by oky**** View Post
hi,
oh, i see. well then you should record everything at 192kHz, right?
right.
Actually you should record at the highest sampling rate with a base freq. of 44.1 if your going to CD (ie... 88.2 , 176.4 , 352.8 )
If you're going for DVD, then use the base freq. of 48k


Regards,
Old 16th January 2009
  #111
Gear Guru
 
AllAboutTone's Avatar
 

Quote:
Originally Posted by dangoudie View Post
At the end of the day I think this is a REALLY minor issue and there are a huge number of factors that will impact more greatly on a recording.
DITTO !!!
Old 16th January 2009
  #112
Lives for gear
 
jamwerks's Avatar
 

Quote:
Originally Posted by DSD_Mastering View Post
Actually you should record at the highest sampling rate with a base freq. of 44.1 if your going to CD (ie... 88.2 , 176.4 , 352.8 )
If you're going for DVD, then use the base freq. of 48k,
I've read alot here on GS saying that this is not true. What is your reasoning?
Old 16th January 2009
  #113
Lives for gear
 

Quote:
Originally Posted by jamwerks View Post
I've read alot here on GS saying that this is not true. What is your reasoning?
The idea being that computing down from 88.2 to 44.1 is exactly half and therefore far easier (or accurate) for a digital system to handle.

When you have things like floating point audio engines and the like, the theory is probably selling the poor computer short these days. I'd wager no-one could tell the difference between an 88.2kHz file or a 96kHz file downsampled down to 44.1.
Old 17th January 2009
  #114
Lives for gear
 
DSD_Mastering's Avatar
Quote:
Originally Posted by jamwerks View Post
I've read alot here on GS saying that this is not true. What is your reasoning?
We've been doing DSD transfers for Chesky's HDtracks. The base freq. of DSD is 44.1 Chesky wanted 24/96 files for download. They just didn't sound right. We tried every conceivable conversion method and ended up with 44.1 base freq files sounding the best. We are now delivering 88.2 and 176.4 files.
Whenever you have an odd-integer down or upsampling, it just doesn't sound "right".
Old 18th January 2009
  #115
Gear Nut
 

I may have to open a new thread for this but Ill try here first.
The last few months Ive been recording my live shows. Its just solo acoustic and vocals with harmonica. I usually sing right on the mic. Im using cubase 4.5 on a newer macbook pro with a prism Orpheus recording at 96k/24 clocked to the orpheus. Also using the prisms mic pre's. I noticed a weird smearing or washy sound to the sibilance on my vocals. I wasnt hearing it in my in-ears or the FOH. I tried three different live vocal mics and the outcome was the same. The next shows I backed away from the mic and this helped 60% of the issue. the other night i didnt have space on the hard drive for 96k so I recorded at 44k/24 but I used a AEA trp for the vocal mic and the guitar di and the washy sibilance problem was gone!!! I thought the preamp did the trick so last night I used exactly the same setup but went back to 96k and guess what. Washy sibilance is back. if I use a LPF anywhere between 15-20khz in Cubase after the fact it does help but doesnt sound as good as the 44k recordings. Also a de esser doesnt help as the smearing is still audible. Am I hearing aliasing? And if 96k helps with aliasing then why is it worse then 44k in my situation?
Old 19th January 2009
  #116
Lives for gear
 
DSD_Mastering's Avatar
This sounds more like a clocking/jitter issue. When I do a digital - digital transfer where one is a 44.1 base freq. and the other is a 48k base freq. then there is a ringing because of this issue. My guess is that you would be fine at 88.2 or 176.4
What was master/slave? Do you have a mismatch?


Regards,
Old 19th January 2009
  #117
outofphase
Guest
At least CLA can't hear the difference.

He mix from a Sony Dash 16 bits open reel machine, even if he gets multitrack PT session at 96/24, he will reduce it to 44/16. And hey, his mixes sound outstanding.

Those purist using recording at 96 are only loosing their precious processor for a barely audible difference.



oOp.
Old 19th January 2009
  #118
Gear Maniac
 
adamlloyd's Avatar
 

I have to say...after reading several of these threads about the 44/48/88/96 debate...

I think it's kind of goofy, the argument about "Well CDs are 44.1, so it all ends up 44.1, so it's dumb to do anything higher."

I'm sorry, that's just kinda goofy!!

That's like saying, "Well this cover photo of Britney Spears is only going to be 8x10 on the magazine, so why shoot it and photoshop it in high resolution???"

And SRC isn't a matter of just "throwing out" half the data. It's interpolation. Everything gets averaged in. Higher sample rates make a difference, in my opinion.
Old 19th January 2009
  #119
Gear Nut
 

Quote:
Originally Posted by DSD_Mastering View Post
This sounds more like a clocking/jitter issue. When I do a digital - digital transfer where one is a 44.1 base freq. and the other is a 48k base freq. then there is a ringing because of this issue. My guess is that you would be fine at 88.2 or 176.4
What was master/slave? Do you have a mismatch?


Regards,
the Orpheus is the master (set to local) as its just the orpheus going into the macbook via firewire. nothing else involved. pretty sure I had the same results at 88k. but Im gonna do some tests at home and see where the smearing begins sample wise. I get a feeling that 44k is just rolling off the very top end eliminating the innacurate response of the live mics. I may start a new thread thats more specific and see if anyone else has experienced this.
Old 19th January 2009
  #120
Lives for gear
 

Quote:
Originally Posted by outofphase View Post
At least CLA can't hear the difference.

He mix from a Sony Dash 16 bits open reel machine, even if he gets multitrack PT session at 96/24, he will reduce it to 44/16. And hey, his mixes sound outstanding.
It's actually 48/16..(without dithering apparently) but yeah.

Then he prints the mixes at 96kHz/24bit, so.......
Post Reply

Welcome to the Gearslutz 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…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
nexxoussone / So much gear, so little time
3
jax72 / Low End Theory
8
not_so_new / Music Computers
8

Forum Jump
Forum Jump