Jump to content
TerryBritton

192khz - Could Someone Explain Why You Would Record At This Rate?

Recommended Posts

As the title asks, can anyone (of sound mind and body) explain to me why anyone would waste resources to record at 192khz (and higher)?

I know the short answer -- clients. But...

(Please tell me nobody "bounces up" to 192khz. If so, why would they? That seems crazy to me.)

Perhaps this article (on playback at 24/192) has tainted my understanding. After all, that article is about playback and music delivery. I'm willing to give the benefit of the doubt that recording may be a different matter, but I need convincing.

Anyone? Or has this become such an ideological debate that the question should be banned from the forum? :-p

Terry

 

Share this post


Link to post
Share on other sites

.. it's definetly a bit overrated.

Go 24 Bit / 48KHz. The "upgrade" from 44.1 to 48KHz is the one you will definetly hear ...

 

Share this post


Link to post
Share on other sites

At 96k where i work at, thee is a big difference, but NOT in the way most think.    Between 44.1 and 48k you hear some differences in the highs and high mids. but going to 96k, there is no freq response difference really, but there is a big difference in depth of sound, in resolution (meaning things sound more natural and as they are in the room). A 3D and smooth-ness to the stereo field or sound stage. An acoustic gt sounds more there in front of you, especially when recording with room mics or stereo micing.  I am totally sold on 96k but not 192k, the trade off for resources is too much.  96k seems the best middel ground for my work,

Upsample to...  I do convert files from 44.1 or 48k up to 96k simply because plugs sound better for the most Part. Example, the UAD 88RS, my favorite channel strip, sound bad on the top end (EQ) at 44.1 or 48. (the orignal 88RS, it does not upsample to prcoess).  but the same plug on 96K file has air and a beautiful top end. 

Just my 2cents.  I did post examples in my blog abouit my own testing, but you should test yourself at 96K and 192k. 

http://timdolbear.blogspot.com/2013/12/96k-verse-441k-sample-rates-my-real.html

Share this post


Link to post
Share on other sites
1 hour ago, Tim Dolbear said:

Just my 2cents.  I did post examples in my blog abouit my own testing, but you should test yourself at 96K and 192k. 

http://timdolbear.blogspot.com/2013/12/96k-verse-441k-sample-rates-my-real.html

Very interesting read, Tim. I've waffled on whether to go to 96k a few times with very light experimentation, but I think I'll give it more opportunities to impress me after that article. I keep going between 44.1 and 48k, but your experience definitely has me curious now!

Terry

Share this post


Link to post
Share on other sites
16 hours ago, TerryBritton said:

Very interesting read, Tim. I've waffled on whether to go to 96k a few times with very light experimentation, but I think I'll give it more opportunities to impress me after that article. I keep going between 44.1 and 48k, but your experience definitely has me curious now!

Terry

Thanks for checking it out.   I also suggest something like Saracon to do your SRC, its worth it!

here is my blog about it if you're interested:

http://timdolbear.blogspot.com/2014/12/96k-world-year-plus-into-my-adventure.html

Share this post


Link to post
Share on other sites

All modern A to D converters operate at high sample rates and then down sample to a stream or file format. It's really all about when, how and by how much to down sample.

Share this post


Link to post
Share on other sites

I work at 88.2k since 2007. Why? Because Double Speed allows using anti-alias filters of much better behaviour than 44 or 48k. Some sources have a lot of energy above 20kHz. Cymbals in particular, have so much HF energy that even with the microphones low-pass filtering, there is enough content above the Shannon frequency (1/2 SR) to "disturb" the steep digital filter response, resulting in the presence of non-negligible aliases.

At Double Speed, none of these problems exist, because the amplitude of the remaining 40kHz components is so small that the artefacts are inaudible. Well, that may not be the case if you used a B&K ultrasonic mic and dedicated preamp, but using "standard" (designed for 20kHz) equipment, these 40k+ components are enough attenuated.

QS (192k) would make a difference if using microphones  (or electronic signals) with a response up to 80 kHz. Most of the times, the recording chain takes care of these potentially offensive frequencies.

Share this post


Link to post
Share on other sites

I agree.   I do not like upampling plugs for the most part.   Remember, it does not matter what the science is, or if its the SRC in the A/D or the was an A/D works are different rates or if the moon is aligned with Mars... simply go with your ears, doing blind test over many days. In the end, I don't care if its 8bit Radio shack setup that sound the best, ya know? I certainly do not do 96k cause its cool.  if so I would be bragging I work at 192...

Share this post


Link to post
Share on other sites

It's unfortunate that some people's manhood seems threatened by the possibility that 44.1x16 might not be enough. Back when the CD was introduced, none of the top DSP engineers thought it was enough which was why the Society of Motion Picture and Television Engineers insisted that a higher sample rate be made standard. We wound up with 48k x 20 bits as a minimum standard for professional production.

Share this post


Link to post
Share on other sites
1 hour ago, Bob Olhsson said:

It's unfortunate that some people's manhood seems threatened by the possibility that 44.1x16 might not be enough. Back when the CD was introduced, none of the top DSP engineers thought it was enough which was why the Society of Motion Picture and Television Engineers insisted that a higher sample rate be made standard. We wound up with 48k x 20 bits as a minimum standard for professional production.

I cannot help but wonder why they stopped at 48K with 60K seeming to be the agreed upon optimum, according to that article referenced earlier. Perhaps designing bandwidth-limiting filters for that rate of 60k had not yet attained the required level of refinement.

Terry

Share this post


Link to post
Share on other sites

One aspect of that article mathias shared above that has been nagging at me (along with Tim, Jean Luc and Bob's testimonials) is that of how different converters might perform at 96k vs 44.1k, especially the part of 96k being more "forgiving" of weaknesses in an A/D or D/A converter's filter designs.

I only have three devices here to test besides the ones built into the computers (like the RealTek ones) -- a MOTU Ultralite mk3 hybrid, a MOTU 828 mk3 hybrid (which I assume shares the same converters and filters as the Ultralite) and an original version Focusrite 2i2. The MOTU boxes often get high marks in reviews. I never expected much from the $149 Focusrite 2i2.

So, last night I switched about 30 times between 44.1k and 96k playback on the MOTU Ultralite to see if the D/A converters exhibited any difference in sound, which I did not expect them to since I was just feeding them a 44.1k digital signal to process. Much to my surprise (and chagrin), the 96k playback sounded clearer, with better space between instruments and spatial definition (clearer boundaries and placement), less "klonk" in the upper bass/lower mids, and better depth! I was listening using Shure SE846 in-ears for consistency so that no comb-filtering could affect my listening from head position movements in a room with speakers. This really boggles my mind, as it just seems "wrong"! I am going back to try it some more with different program material tonight.

Part of me wants the 96k to sound worse... being the cheap sort who doesn't want to waste resources! ;-)

I am 64 years old and should not be able to hear any difference, if you ask me (good up to 14kHz in tests clearly), so I am wary of myself falling prey to that famous error of expectation-fulfillment. But it seemed quite obvious last night. Oh my! This is a quandary! :-|

Terry

[afterthought] Do the bandwidth filters even apply to the D/A direction? Or do they only come into play during the sampling stage of A/D?

[Edit] Wait - they must -- audiophiles spend hefty amounts to get high-end D/A converters, right?

Share this post


Link to post
Share on other sites

In case you were wondering...

A decent hearing test that does not fall prey to the issues the ones on YouTube face. (YouTube compression is limiting to 16-17k)

http://onlinetonegenerator.com/hearingtest.html

(Nobody needs to report their results!)

The Focusrite 2i2 is definitely producing artifacts for me doing this test, putting out odd subharmonic sweeps over 14k. Annoying.

Terry

 

Share this post


Link to post
Share on other sites
1 hour ago, Bob Olhsson said:

Headphones aren't nearly as demanding of low distortion as speakers in a live room! This is why most club DJs won't use mp3s.

No kidding?! I absolutely did not know that. 

Terry

Share this post


Link to post
Share on other sites
19 hours ago, TerryBritton said:

Part of me wants the 96k to sound worse... being the cheap sort who doesn't want to waste resources! ;-)

I am 64 years old and should not be able to hear any difference, if you ask me (good up to 14kHz in tests clearly), so I am wary of myself falling prey to that famous error of expectation-fulfillment. But it seemed quite obvious last night. Oh my! This is a quandary! :-|

Terry

[afterthought] Do the bandwidth filters even apply to the D/A direction? Or do they only come into play during the sampling stage of A/D?

[Edit] Wait - they must -- audiophiles spend hefty amounts to get high-end D/A converters, right?

now try an RME UFX+ and you will be reaching for your credit card

both AD & DA are critical in creating the sought after airy realism 

Share this post


Link to post
Share on other sites

There is a funny story about the EBU spending a fortune developing and blind testing a lossy audio codec only to have some guy in Los Angeles spot an artifact on loudspeakers that was blatantly obvious once people had been told what to listen for. My own jaw-dropping experience of this was a codec demo in a painfully live convention hall. The presenter wasn't sure we'd be able to hear any differences due to the horrendous acoustics. To our amazement what had been subtle in his mastering room was blatantly obvious over the PA system.

Share this post


Link to post
Share on other sites

By the way, hearing loss begins at birth starting in the midrange and progresses over time with more exposure. It is not a loss of high frequency bandwidth and is not a matter of age or sex which was believed decades ago.

Share this post


Link to post
Share on other sites
Guest
You are commenting as a guest. If you have an account, please sign in.
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×