Why the debate over 44.1 or 192 .
phatbeatstudio
Thread Starter
#1
21st May 2012
Old 21st May 2012
  #1
Lives for gear
 

Thread Starter
Why the debate over 44.1 or 192 .

Why the debate over 44.1 or 192 . Do people really believe there is absolutely no sonic benefit from 192 over 44 or 48.
I know there are many reasons not to because of CPU usage and such but if computers could process 192 as good as 44,1 or 48 would this debate be over.
I wanted to record at 192 I tried and found out that with regular $1500 monitors I could not here any improvement , BUT , at 32bit 192k in HDX with HD converters and Barefoot MM27s I could .
So does this debate of 192 is no better really just a matter of resources . CPU power and other equipment not actually being good enough to hear it. In 15 years when a Mac Pro is 40 cores will people still say 192 is no benefit . Im just thinking out loud here but is this debate really just a matter of resource VS reward. When did a audio engineer decide the very best audio quality possible wasn't wasn't a concern.
#2
21st May 2012
Old 21st May 2012
  #2
Lives for gear
 

It is not a question of computer power, but the optimal working of the ADC. Higher sample rates bring problems with the converters, that is why higher sample rate is not necessarily better. Optimal quality seems to be around 88.2 or 96 kHz sampling right now, and it is dubious we really need more, as only the upper frequency limit gets raised, nothing more, and it is already well past our hearing limit.

Have you not read any of this: The Optimal Sample Rate for Quality Audio just a few threads down?
phatbeatstudio
Thread Starter
#3
21st May 2012
Old 21st May 2012
  #3
Lives for gear
 

Thread Starter
What do you think of my opinion of people not really hearing the quality difference due to there own equipment not actually being good enough to hear its benefits
#4
21st May 2012
Old 21st May 2012
  #4
Lives for gear
 
SWAN808's Avatar
Quote:
Originally Posted by phatbeatstudio View Post
What do you think of my opinion of people not really hearing the quality difference due to there own equipment not actually being good enough to hear its benefits
its a reasonable point but there is more to it than that. The main one being that its possible (and reasonably logical) that 192 results in poorer quality audio record/playback than 96 due to issues lavry has pointed out. Also faster sampling does not necessarily = higher quality due to reasons outlined in that article.

There is another aspect here however that Lavry etc dont discuss as its a separate issue - but if we are talking 'processing' audio ITB in a mix session once it has been sampled - then you could argue 192 or higher could be advantageous. Its not unusual for plugins to upsample 2x for example...however - it sounds like recording at 96 and processing at 96 or up-sampling if necessary could be best.
#5
21st May 2012
Old 21st May 2012
  #5
Lives for gear
 

The majority at this site still record at 44.1k or 48k (someone starts a poll every 6 months on that it seems). Consensus as I understand per many threads here is that with great modern converters 44.1k or 48k is all you will need (there are great old converters at 44.1k / 48k). Some converters will sound better at 88.1k or 96k. 192k is for marketing purposes (my amp goes to 11). I think Lavry basically said 60k was the optimal sampling rate with computers which several others agree with, but given the standards in place I doubt we will ever see that come into play. I remember in the very first days of this stuff Sony thought that 50k was the optimal number so 48k was not much of a compromise. Given the Alge brothers are still mixing with Sony 3348's which only do 44.1k / 48k there is no shame with those sample rates or good old units despite what the marketer's want you to feel to push you into buying something new (the Alge's also feel that digital tape is the best LONG TERM storage medium to recall tracks decades in the future). If you already have converters then do some testing (hopefully with a friend so you can eliminate some expectation bias while listening) and see what works best for you. If this is about what converter to buy, look at what interface you want first as that will limit your converter choices, then read up here on your workable choices with that interface and other's experiences with them.
#6
21st May 2012
Old 21st May 2012
  #6
Gear interested
 

Quote:
Originally Posted by SWAN808 View Post
its a reasonable point but there is more to it than that. The main one being that its possible (and reasonably logical) that 192 results in poorer quality audio record/playback than 96 due to issues lavry has pointed out.
Except that the "issues" lavry has "pointed out" are not really issues. He's pointing out ghosts and using influence to cause unwitting, less educated engineers to follow his erroneous statements. That "paper" is smoke and mirrors.

Quote:
Also faster sampling does not necessarily = higher quality due to reasons outlined in that article.
Well, faster "sampling" [actually "system rate", because the same megahertz oversampling rate is used for all target rates] = lower latency, which means better music, lower quantization noise in the audio band, and better sounding processing ........... So.........


Maybe lavry's own converters don't sound right at higher rates [so he leaves them out], but that's a design issue, and it doesn't necessary apply to other manufacturers' products.

Really, when was the last time you heard a client say, "Oh man, that sounds bad, like you used too high a sampling rate, dude". "Please redo it at a lower rate, like Dan Lavry likes".? Never is when you heard that, and never is when you will even hear it, because stuff recorded at 192kHz sounds great [assuming its well recorded stuff and you're using good gear].

I don't know how many times it has to be stated, the only "issue" with real high sample rates is reduced track count. Now that is a legitimate issue, and its why most people settling on using 96kHz instead of 192kHz. But, for smaller projects [or if you have money to be using multiple Pro Tools rigs simultaneously], 192kHz is great. Or even 384kHz if you have Pyramix or one of the other systems that do 384kHz.

Quote:

There is another aspect here however that Lavry etc dont discuss as its a separate issue - but if we are talking 'processing' audio ITB in a mix session once it has been sampled - then you could argue 192 or higher could be advantageous. Its not unusual for plugins to upsample 2x for example...however - it sounds like recording at 96 and processing at 96 or up-sampling if necessary could be best.

The "optimum" way, where practicable, is to record at high rates, and keep the audio at high rates throughout the entire project. That way the plugs don't have to upsample, they can just process natively at the higher rate.
#8
21st May 2012
Old 21st May 2012
  #8
Gear interested
 

fyi, for anyone who does not already know this, "the sampling theorum" is an idealized statement. It assumes any number of things that are not true, and cannot be true, in implementation.

Basically, that is not really even the way sampling is done, or the way converters work, although it is the basis for those things.

It assumes perfect sample values [i.e. no quantization], perfectly periodic waveforms, and perfectly narrow, infinite dirac impulses, perfect filters [which do not exist], and on and on.

Once you start making allowances for all the inaccuracies and errors in real-life sampling [and there are dozens of different sources of inaccuracy in sampling] you see that it is necessary to go far beyond the "requirements" of the theory to get it "right".
phatbeatstudio
Thread Starter
#9
25th May 2012
Old 25th May 2012
  #9
Lives for gear
 

Thread Starter
#10
25th May 2012
Old 25th May 2012
  #10
Lives for gear
 
Magucci's Avatar
 

Quote:
Originally Posted by phatbeatstudio View Post
earth is flat isn't it everybody else thinks so.
Yeah and the elephants are standing on turtles
#11
25th May 2012
Old 25th May 2012
  #11
Lives for gear
 
elambo's Avatar
 

When an argument like "which sounds better" is as tough to prove as this I realize that there's no pressure to decide between them. Instead, I look at other factors, like CPU usage, and that's why I stick with 48K. It sounds as good as it needs to and mid-sized sessions don't consume my entire CPU, as happens with 96K.

As another wrinkle (one I find more relevant): someone made a point that different sample rates render different results. "Colors," I believe they called it. I agree, which means that one SR would be better for Genre A and another SR would be better for Genre B. When going with a natural, organic sound, I sense a benefit in 96K. For rock, I feel a slight advantage in lower SRs.
#12
25th May 2012
Old 25th May 2012
  #12
Lives for gear
 
popmann's Avatar
I don't need Barefoots or HDX to hear the benefit of 88/96. I can not, however hear one lick of difference in 96 and 192.

Look at how many people profess to hearing how much better Slate's VCC sounds at 2 and 4 times over sampling...lots of the same people who claim to not hear a difference between 44 and 96. What do you think the Slate plugs are doing? Up and downsampling, as many modern plug ins DO--using more CPU than they would to run natively at the higher sample rates.

I think it comes down to people not knowing what to listen for. Maybe on the lower end, their systems can't reproduce it--but, I can plug my Sony V6s into the Benchmark DAC1 and hear the same qualities, to a lesser degree that I hear at 96 on my Paradigms.

I don't believe I have super human hearing...nor do I believe I'm just biased--hell, I record mostly at 44.1, because that's what my (non computer based) recorder works at...I would LOVE to hear no difference! But, the issue is that it's not like BluRay versus DVD on a nice HD panel--you'd have to be BLIND not to see that difference. I think people in general never bought HD audio because it's NOT that different. It manifests at higher volumes...on nicer systems...better imaging and "third dimension"...better reverbs ITB...no need to boost highs to make up for the poor LPF getting in the way at 44. It's better-it's just not smack your mama upside the head better--which is what it takes to get people to buy.

Engineers here and elsewhere on the net lack the experience to know what to listen for...some are listening for some ray of God from above grand sonics that make their shitty Chinese U47 knock off and RME preamp sound like a classic record. They won't hear that with higher sample rates.

Hell-maybe I lack whatever it takes to hear that 192 is better...or worse, as Larvy seems to think it is--I feel like that's just taking up resources for no reason.
#13
25th May 2012
Old 25th May 2012
  #13
Lives for gear
 

Quote:
Originally Posted by phatbeatstudio View Post
The problem with that is . Its exactly what it says it is, A theory .
which is not anymore relevant than a opinion with hypothetical graph charts .
but it is the leading answer.

earth is flat isn't it everybody else thinks so.
It's a theorem, a mathematical proof, not a theory, they are very different things.

The theory is that 20kHz is the bandwidth we need to capture, that a 40kHz sample stream can contain all the information in a 20kHz bandwidth signal is not theory, it is fact.

(It's also fact that due to limitations in filters and other factors you'll need more than 40kHz, but thats because you'll have a signal with greater than 20kHz bandwidth that you need to capture, Nyquist-Shannon still applies).

There are reasons why in many circumstances there are advantages to working with even higher sample rates, but that's because there can be advantages to working with higher bandwidth signals (aliasing is the usual problem), Nyquist-Shannon still applies in those cases.
#14
25th May 2012
Old 25th May 2012
  #14
Quote:
Originally Posted by phatbeatstudio View Post
The problem with that is . Its exactly what it says it is, A theory .
which is not anymore relevant than a opinion with hypothetical graph charts .
but it is the leading answer.

earth is flat isn't it everybody else thinks so.
Being a theorem (which it is) doesn't make it of equal relevance to an opinion. The problem with laypersons criticizing established scientific theory is that they often reason poorly and are largely ignorant of the facts involved in the matter. Such is the case with the Nyquist theorem and typical conversation in this place surrounding it's validity.

In credible testing, with credible sources the Nyquist theorem has been supported. It has been peer reviewed. It is NOT opinion.

Lavry pointed out that the only matter of opinion is that of how high in frequency one would like to capture in audio recording for human beings to ultimately listen too. Given that no human hearing is capable of responding to ultra sonic frequencies or even frequencies in excess of (VERY generously) 30kHz (and in all likely hood none extends even that far, certainly none tested have come close) then there is simply no need to use sampling rates of greater than around 60kHz, which he mentioned in his paper. He also mentioned that it is therefore overkill to sample at 88.2kHz and above, but that if a compromise was to be made to satisfy those with the desire to sample at ultra high frequencies to accommodate the most sensitive human hearing, while still leaving comfortable margins 88.2 or 96kHz are fine (assuming an excellent converter in the first place). Sampling faster than that causes errors in the process that negatively affect the conversion. This is proven. It's not an opinion.

What ARE opinions (and hearsay, and rumor) are the many anecdotal, subjective, unverifiable claims that Joe Goldenears can hear improvement using 192kHz sampling rate. THOSE claims are "tough to prove". Sampling theory isn't. It IS proven. That so many are ignorant of how it works and insist on being considered worthy of consideration, despite the overwhelming ACTUAL evidence against them, is a testament to human self delusion.

All of our sciences are rooted in theory. That is not because one can simply create one's own truth. That's because no self respecting scientist is willing to foreclose on the possibility of future breakthroughs in understanding and perspective. However, just because a POSSIBILITY exists that something MIGHT change one day doesn't mean that what we know and can prove time and again is somehow just opinion.
#15
25th May 2012
Old 25th May 2012
  #15
Gear maniac
 
Headstack's Avatar
 

Quote:
Originally Posted by Enlightened Hand View Post

What ARE opinions (and hearsay, and rumor) are the many anecdotal, subjective, unverifiable claims that Joe Goldenears can hear improvement using 192kHz sampling rate. .
Is this your theory? or do you have scientific proof that Joe actually can't hear the difference?
I can personally hear the difference.
#16
25th May 2012
Old 25th May 2012
  #16
Quote:
Originally Posted by Headstack View Post
Is this your theory? or do you have scientific proof that Joe actually can't hear the difference?
I can personally hear the difference.
I will ACTUALLY bet you $1000 that you can't prove that you can hear the improvement. Let me be clear. I mean IMPROVEMENT. Not difference. Difference you should hear, difference for the worse.
#17
25th May 2012
Old 25th May 2012
  #17
Gear maniac
 
Headstack's Avatar
 

Quote:
Originally Posted by Enlightened Hand View Post
I will ACTUALLY bet you $1000 that you can't prove that you can hear the improvement. Let me be clear. I mean IMPROVEMENT. Not difference. Difference you should hear, difference for the worse.
sorry I think betting is against forum rules
#18
25th May 2012
Old 25th May 2012
  #18
Lives for gear
 
projektk's Avatar
 

I can hear a difference between conversions but I couldn't tell you which is which. I don't think higher conversions sound better, but there is a difference. Digital signal has an effect on audio that's for sure. But again who cares. I use 48khz, gives me much better performance than say 96 or 192...

Sent from my LG-P925 using Gearslutz App
#19
25th May 2012
Old 25th May 2012
  #19
Gear maniac
 
Headstack's Avatar
 

Quote:
Originally Posted by projektk View Post
I can hear a difference between conversions but I couldn't tell you which is which. I don't think higher conversions sound better, but there is a difference. Digital signal has an effect on audio that's for sure. But again who cares. I use 48khz, gives me much better performance than say 96 or 192...

Sent from my LG-P925 using Gearslutz App
I agree, which is better is subjective.
#20
25th May 2012
Old 25th May 2012
  #20
Which is exactly what I said in the first place.
#21
25th May 2012
Old 25th May 2012
  #21
Gear addict
 

Quote:
Originally Posted by phatbeatstudio View Post
The problem with that is . Its exactly what it says it is, A theory .
which is not anymore relevant than a opinion with hypothetical graph charts .
but it is the leading answer.

earth is flat isn't it everybody else thinks so.
...and with that, the final nail in the coffin of whatever previously easily understood distinction between fact and opinion existed has been resoundingly hammered into place.

Score: Humanity 0, Internet 2. (The previous point being the notion of privacy, which may find a resurgence after enough people are denied employment over the content of their Facebook pages)
#22
25th May 2012
Old 25th May 2012
  #22
Lives for gear
Quote:
Originally Posted by phatbeatstudio View Post
What do you think of my opinion of people not really hearing the quality difference due to there own equipment not actually being good enough to hear its benefits
Most only think they can hear the benefits when listening on low or high end systems. Since you're human, and are subsceptible to many outside biases through alternate non-aural senses, most of which you, as a human, are unaware of, what you are "hearing" is not what initially entered your ears before being mangled by outside forces before reaching its destination known as the human brain. And I am not making this up. It's called perception.

To combat this phenomenon and to find out what you're truly hearing, you need to perform a proper and valid scientific A/B test with at least a confidence level of 95% or greater.
#23
25th May 2012
Old 25th May 2012
  #23
Lives for gear
Quote:
Originally Posted by THE MAN IS HERE View Post
I think being Greedy is a part of human nature and this really has nothing to do with bits....

15-20 years ago most home recordist used 4 tracks, and did the best they could with them. Most folks were happy to have that. You were damn lucky if you had a R2R or a Dat.

Then, digital comes along....

I'm not a huge fan of digital, but hell... is it cleaner sounding and you can do alot when it comes to editing, plugs, etc...

For some, Digital is basically a dream compared to the old way of doing things. SO... with that said, you'd think people would be happy with even the "basics", say 16bit 44.1 compared to what was available in the old days, as mentioned.

No, they're not content, they want more... and then debate over it, ONLY to abandon the whole thing as crap or obsolete when the next latest greatest comes along....
Interesting perspective!
#24
25th May 2012
Old 25th May 2012
  #24
Gear maniac
 
phillipc421's Avatar
 

Most consumers will not be able to differentiate between the sample rates on their gear or with their ears.
#25
25th May 2012
Old 25th May 2012
  #25
Lives for gear
 
popmann's Avatar
That's true...but, to what relevance? They listen to mp3s from Amazon and think they sound fine. So, we should now record directly to mp3? Did Fleetwood Mac's engineer say "well, we just have to make Rumours sound good enough for 8 track cassette...?"

That's kind of a straw man argument. You should go ahead and throw in "well, a good song at 44 is better than a bad one at 192"...
#26
25th May 2012
Old 25th May 2012
  #26
Gear interested
 
ricordu's Avatar
 

I'd prefer 96k on Lavry than 192 on Motu

Ricordu
#27
25th May 2012
Old 25th May 2012
  #27
Lives for gear
 
SWAN808's Avatar
Quote:
Originally Posted by popmann View Post
Look at how many people profess to hearing how much better Slate's VCC sounds at 2 and 4 times over sampling...lots of the same people who claim to not hear a difference between 44 and 96. What do you think the Slate plugs are doing? Up and downsampling, as many modern plug ins DO--using more CPU than they would to run natively at the higher sample rates.
but you are confusing 2 separate processes here. One is Ad/DA conversion - the other is signal processing at a given sample rate.

AD/DA conversion is about effectively sampling - then reconstructing an audio signal. Lavry states that because of the limitations of human hearing and the potential for inaccuracies at high rates - higher sample rates does not necessarily mean higher quality audio reproduction in digital systems.

However - in a computer - digital signal processing - IE changing that sampled analogue waveform which came in thru the A/D converter and now exists as 1s and 0s, is a different topic....

My own opinion is that I dont think there is a huge problem with 44.1 recording and playback provided you use good converters. If you are simply using the computer as a tape machine - doing mixing and processing OTB I think its probably fine.

The issue is that I dont think 44.1 provides enough headroom for the significant amounts of digital signal processing done ITB that is common now days. Once you start stacking on the plugins I think the digital audio information is changed in ways that digital artifacts mount up.

One potential solution is to record at 44.1 and use plugins that upsample. However here you have to rely on real time sample rate conversion - which has its own limitations.

I think if you are running a system that uses a lot of ITB - 96k is probably the way to go and keep plugins processing at project sample rate-no realtime SRC-just a clean overall rate for recording and processing.

However at the moment I dont think thats ideal for most because of CPU power issues - not sure we are there yet. Also - a lot of sample libraries are at 44.1 so they need to be upsampled. I am hoping the next generation of computers will make 96k more common and possible on a wider scale due to faster CPU and cheap storage, and because of the rise of ITB - production tools such as sample libraries and processors will be optimised for 96k for producers to work with at high quality to maintain high audio integrity - then downsampled for consumption.

Id be curious what others think - Im just freewheeling some thoughts...
#28
25th May 2012
Old 25th May 2012
  #28
70% Coffee, 30% Beer
 
Doc Mixwell's Avatar
44.1 sounds perfect in my studio, with the level of ADA converters I am using....its a no brainer...Working on Music, that will end up on a CD and Digital Distribution Medium.

I stopped drinking the high sample rate KOOL AID many years ago, after using 96K for years, the results led me to believe my audio was not aesthetically pleasing.

It was brighter, with looser bass. Thinner sounding. 44.1 to me sounds more even with less problematic brightness, [exacerbating the treble range] and the bass is more even and firm to my ear. Its more pleasing to my ear.

It is not "loose" sounding. It is not "bright"...

I can easily discern differences with sample rates. And my conclusion is that 44.1 sounds the best in my studio. I base my conjecture on many years of Digital Audio experience.
#29
25th May 2012
Old 25th May 2012
  #29
Lives for gear
 
Nick Morris's Avatar
 

What were you using for years at 96k that sounded this way? When did you decided to change and what made you discover this difference? I worked for many years at 48k and was sent a couple 44.1khz sessions and those instantly sounded lofi compared to what I was use to hearing. In my experience, the higher the sample rate the better my converters sound. 384khz is king for PCM in my studio. The only thing that sounded better was DSD. This is based off a merging technologies sphynx2 converter. I currently work large PCM projects at 96k. 384khz for small PCM projects like classical recordings. Capture in DSD. This may have to do with most of the mixing taking place on a euphonix CS2000.

I spent a looooong time with converters and believe that there is no rule to which sample rate is best. To me it really depends on the converter you have. Every converter will have a sample rate which it performs it's best at. Mine happen to sound better at higher rates, (sphynx2, Burl B2, Euphonix AM713,MA703).
#30
25th May 2012
Old 25th May 2012
  #30
70% Coffee, 30% Beer
 
Doc Mixwell's Avatar
Quote:
Originally Posted by Nick Morris View Post
What were you using for years at 96k that sounded this way? When did you decided to change and what made you discover this difference? I worked for many years at 48k and was sent a couple 44.1khz sessions and those instantly sounded lofi compared to what I was use to hearing. In my experience, the higher the sample rate the better my converters sound. 384khz is king for PCM in my studio. The only thing that sounded better was DSD. This is based off a merging technologies sphynx2 converter. I currently work large PCM projects at 96k. 384khz for small PCM projects like classical recordings. Capture in DSD. This may have to do with most of the mixing taking place on a euphonix CS2000.

I spent a looooong time with converters and believe that there is no rule to which sample rate is best. To me it really depends on the converter you have. Every converter will have a sample rate which it performs it's best at. Mine happen to sound better at higher rates, (sphynx2, Burl B2, Euphonix AM713,MA703).
More importantly than what hardware I've used and what mixes I've fought....to your point, all ADA converters sound different at different sample rates. Because of their design, filter networks, clocking, circuit topology,PSU etc. They all sound different and the subjectivity of it is mind boggling. The simple answer is to find out for yourself. I do not have a nuclear booster pack attached to my computer hardware to record, process and store more samples, for this difference. There may be a time in my career, where sampling high makes more sense to the application, but I don't think I'll see that day any time soon.

I speak in highly general, preferential terms. I do not preach, or dictate what the other gearmonkey women or man, should be using in their applications. I only mean to share, my viewpoint, which has been derived by years of working at different rates, with different genre's, and people compeled to test science against music. It is derived from completing projects from start to finish, pretty much doing everything myself. Because I have found, that it doesn't matter what you use. Sure. The good ones sound great at all sampling rates. They have Different responses, but all great. I think it does't matter, what you use, because no matter what - we are recording[encoding and decoding] more High Freq. content with higher sample rates,

They might sound different from one another, but we are capturing up to 48kHz, with 96K, and then 96kHz with 192k sampling rates. Electronically, I think this can be a problem for the entire system, instead of breeding more fidelity. It will depend on many things. I have been tricked many times by 96K being "better", and then when I start processing the audio, things are not "better" at all. My viewpoint, is that 44.1 with good converters sounds absolutely fine to me, [if you like those ADA converters] and that you will have less conversion in the end, to the medium its planned for. Whatever sampling rate you use, you should plan for as little [or none] sample rate conversion as possible. I hate sample rate conversion. It sucks.

If the project is going to mastering, I will let the ME do that...because he has better hardware and software than I...but even still, I have been very happy with the projects I mix, and send to master @ 44.1. With 96K, I have found, the sound overall, has more HF artifacts, but I must stress - I can only speak to my own worn ass ear's. I have found, I struggle more with mixing at Higher Rates. I find, the bass does not sound that great to me, and I struggle to find a good sound. All of this could be good or could be bad. All of it, could be good or bad. Context is king here. Application key. Aesthetic is subjective. The Objective, is a cement wall. Your work? Is done with a sledge hammer. It will get messy.
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
BLueROom / The Moan Zone
260
AlexLakis / Music Business
306
Matt Grondin / High end
30

Forum Jump
 
Register FAQ Search Today's Posts Mark Forums Read

SEO by vBSEO ©2011, Crawlability, Inc.