MOTUnation.com

Sure, I could be wrong but I doubt it. I answered the question as best I could, it's absolutist in nature, and that's not how the world works. With one converter set it may not matter, with another it may make a huge difference. Which is to say it's not necessarily the rate itself, but the hardware. Way too many variables to make across the board statements like 'there's no proof of benefit'.

Let's add another question. How many people hear no difference in different converters? I know people who say this.

How many people here use analog hardware at mix? I would guess very few. That makes it less your problem.

Try another one: you could argue some additional benefit with some converters could be had capturing at high rates and then down converting to 48, versus capturing at 48, use a better decimator in software than exists in the converter.

EMRR wrote:It's situation specific. You are asking an absolutist question.

How is my question situation-specific or "absolutist", exactly?

It's quite a simple question, actually. Either, you accept you may be wrong or you don't. If that's what you mean by "absolutist", I see no other way. Either you accept it (and cherish it) or you don't.

But IF you think you're beyond human bias, and IF you think there's not some degree of chance that you may be dead flat out wrong, then that takes you a step closer to arrogance, and being dogmatic and stubborn, especially if you can't demonstrate your claims (remember the burden of proof????)

You ARE aware of the things people claim without evidence, correct?
Are you familiar with something so typical like pereidolia? People with that condition are as sure of what they hear and see, as you are sure that you exist in the here and now. But what they see and hear is not real, except for the faulty chemistry inside their heads when presented with certain stimuli.

How do you know your head isn't playing tricks on you?

Let me ask you something else... How can you REALLY know or be sure of what you think you know or perceive?
Being "sure" about it because of self-impsosed artificial confidence, or because "you feel it in your gut" and nice anecdotes does not count. That is NOT even evidence.

You do know there's a whole branch of philosophy that deals precisely with those kinds of questions, yes? And it's not pretty... It's actually very humbling...

[I just saw your last post where you admit you may be wrong, but you "doubt it".
On what basis do you doubt you may be wrong?]

As for SR conversion artefacts when you lower it form say, 192 KHz to 44.1, I've read some people take out of the box to a good hardware mixer (and take advantage of is possible colouring) and then simply recapture or rerecord to the new lower sample rate.

Is that more effective than digital SR algorithms? I've never done it, but it sounds to me like the best option if you must SR convert something... Is this the case?

Would you call that resampling?

Getting back to the initial question from the o/p: What is the preferred recording format, WAV or AIFF?

Is anyone still using aiff these days?

That said, I'll just add that we're all digital nerds here to some extent. We've bought into the digital v. analog and for good reasons, but it leaves me wondering if analog isn't capturing stuff that even digital can't, and we just don't know it yet? One could make the same argument about a sample rate.

For me, it's a matter of realistic function for my purposes and in today's world, recording at 48k is a pretty good bet you'll remain current for the foreseeable future - barring any unusual requirements for specific projects and movie formats, etc.

I and others here fall into a simple category: preserve my music in a reasonably high quality format that stands a chance of surviving at least a few years after our demises. Frankly, I find written and printed music (on paper, that is) to be the most reliable. And yet, I also get great joy out of the Robert Johnson mono recordings, a piano role by Gershwin, and the Deutsche Gramophone 1-bit recordings of the Berlin Phil. with von Karjian conducting Tchaikovsky's 6th. I mention that because when it was released, an excellent audio engineer poo-poo'd it as "unnatural" and not a realistic representation of the way an orchestra is balanced. I agreed, but as a composer, I appreciated hearing every note.

That's kind of what is being argued here at the moment. People want to hear every part of every note, and you can't blame them for that. I'm happy if people just play my offing music in the first place. Beyond that (and my PRO) I don't care what sample rate or format you others are using.

No offense intended!

That sounds like a terrible idea if conversion is the only point. Worst of all worlds, and require two systems running at two different rates. If I read you correctly. I may not. I'd call it re-recording.

FMiguelez wrote:As for SR conversion artefacts when you lower it form say, 192 KHz to 44.1, I've read some people take out of the box to a good hardware mixer (and take advantage of is possible colouring) and then simply recapture or rerecord to the new lower sample rate.

Is that more effective than digital SR algorithms? I've never done it, but it sounds to me like the best option if you must SR convert something... Is this the case?

Would you call that resampling?

One type of sound that really highlights the shortcomings of MP3 loss, low bit depths and low sample rates is cymbals. Hi-hats, rides, and crashes really challenge the upper ranges of a digital system. The other thing is the stereo spread. That was something I noticed when going from 24 bits down to 16 is that the stereo field seemed less evenly spread. Could I have been inventing that in my mind? I suppose, but I'm USUALLY not prone to this type of invention. If I don't hear a difference, I'm not ashamed to admit it.

I've been using nothing but AIFs since the SDII days ended. I've never encountered a complaint or problem. But after reading this thread, I'll switch to Broadcast Wave (is that right?)

EMRR wrote:That sounds like a terrible idea if conversion is the only point. Worst of all worlds, and require two systems running at two different rates. If I read you correctly. I may not. I'd call it re-recording.

Perhaps.
To tell you the truth, I wouldn't know, since I've never done it.
They argue that, as long as your hardware is awesome (they probably meant Neve desks and similar), this prevents digital SR conversion artefacts, while benefiting from the characteristics of such desk, and a very clean recapture of the audio at the new SR. (I'm glad I don't have to deal with this, BTW )

I'm under the impression that DP was among the top contenders regarding the quality of its SR conversion algorithm. I remember seeing some charts comparing every DAW, and DP came out in the top 3.

Is this still the case?

I think DP conversion is still very good. If I were going DA for color/processing purposes, I'd still capture at whatever rate I'm already at, and down convert later. Probably post mix. All my mastering guys still prefer to receive at least 88K2 SR and 32 bit float if I can send it. If we're not paying them for their good educated ears and experience (ignore the posers/hacks/amateurs), what are we paying them for?

FMiguelez wrote: How is my question situation-specific or "absolutist", exactly?

It's quite a simple question, actually. Either, you accept you may be wrong or you don't. If that's what you mean by "absolutist", I see no other way. Either you accept it (and cherish it) or you don't.

FMiguelez wrote: Will you, or any user of high SRs, AT LEAST be willing to accept that you may be wrong and you are not really hearing what you think you are?
Emotional investments can become quite expensive at MANY LEVELS...

Do you at least acknowledge that possibility, even if you think it's quite unlikely???

It's both open-ended and across the board, while focusing solely on the mechanism of sample rate. I said, sure, I could be wrong, and I also detailed a lot of ways it's NOT ENTIRELY SAMPLE RATE, but how sample rate potentially eliminates lots of other converter and/or coding shortcomings. There are certainly converter types that don't improve with an increase in rate, usually real top shelf expensive stuff well beyond the means of most.

Converter quality nuance: the AVB line sounds better than the older PCIe line. PCIe 48 to AVB 48 is a bigger step than PCIe 48 to PCIe 88K2. AVB 88K2 is an even bigger step up. Use the converter at max potential if you can afford the real estate and processing.

This is the audio industry, bastard child of science and technology. There's little interest and investment in finding the smoking gun, if any. It's much easier to work at higher rates to obviate audible problems than buy and try every converter out there, much less A/B everything on the market.

Next question: do you accept there could be audible phase/intermodulation problems you either don't experience on your system, that might still be real on others systems, which higher sample rates fix? Yeah, the smoke and mirrors are endless, and this is really tough stuff to measure reliably. It requires test gear most people don't have, with a steep learning curve. Unless someone funds such tests, they probably won't ever happen, there's no motivation to do so.

Sample rate example, how it's not about the extended response: I can track a session with nothing but relatively band-limited 1940's preamps at high rates and capture a level of detail and nuance in the upper frequencies that I can't capture at lower rates. Preamps which start rolling off at 12K-15K, well down at 25K. I've compared sessions at both rates over and over for a decade now, and I hear a repeatable result. Want a corroborating public test? We'd need to track a session on multiple parallel systems with identical converters.....can't do it here, I don't have duplicate systems.

Another question: how many people capture a lot of their own sounds with microphones, versus using sample libraries predominately/exclusively? I have to think if you only deal with samples for the most part, there's less chance you have to hear the differences between a live feed. I could be wrong.....

EMRR wrote: Next question: do you accept there could be audible phase/intermodulation problems you either don't experience on your system, that might still be real on others systems, which higher sample rates fix?

Oh, absolutely! It could be that I don't experience them or I simply don't notice them. It could definitely be the case. I certainly hope it isn't the latter!

EMRR wrote:Another question: how many people capture a lot of their own sounds with microphones, versus using sample libraries predominately/exclusively? I have to think if you only deal with samples for the most part, there's less chance you have to hear the differences between a live feed. I could be wrong.....

[/quote]
Yup. That's me. 95% of my work consists of composing using sampled instrument libraries. Sometimes I record singers or choirs or guitars, but that's about it.

And frankly, if I wanted to improve my mic-to-DAW chain, there would be lots of places where I could improve before even thinking of SRs.

So yes. A lot of this does not apply to me (but at least some of my points still stand ).

OK, so we have fairly different work flow experiences. Everything I deal with is something I capture with a microphone. I wouldn't begin to know how to describe how sample libraries step through sample rates, and if they provide same samples at multiple rates, they probably vary per provider in how they arrive at the varying rates. Apples to apples is probably impossible to comprehend.

Sample rate would seem much more esoteric to me if I dealt with samples only and never dealt with AD, or DA other than for playback reasons. Not to belittle anyone's experience, just to point out how different it can be amongst people called 'audio engineers'.

Not trying to take your points away, conversation is good.

One type of sound that really highlights the shortcomings of MP3 loss, low bit depths and low sample rates is cymbals. Hi-hats, rides, and crashes really challenge the upper ranges of a digital system. The other thing is the stereo spread. That was something I noticed when going from 24 bits down to 16 is that the stereo field seemed less evenly spread. Could I have been inventing that in my mind?

A friend of mine calls it the "only hihat in LA". On pop CDs, it sounds as if they are all using the same one due to the limitations of the compact disk. Easy to hear the difference if you have a good pressing in vinyl to compare.

I've already posted about the stereo spread but I was talking about true stereo: a live performance captured by two or three microphones. Panned mono sounds better at 24bits, too but it's not as apparent when listening to a CD in its 44.1k 16bit glory.

Sony and Phillips wanted to get Fürtwangler's recording of Beethoven Symphony 9 onto a small digital disk. They succeeded and we have been living with the compromise of 1981 ever since.

That same friend says the CD "elevated the level of mediocrity to pretty good". I agree.

EMRR wrote: Sample rate would seem much more esoteric to me if I dealt with samples only and never dealt with AD, or DA other than for playback reasons. Not to belittle anyone's experience, just to point out how different it can be amongst people called 'audio engineers'.

Point totally well taken.
I'm the first to admit I don't have much experience with AD/DA or analog equipment, or even advanced engineering stuff.

But I still enjoy asking questions and challenging people (who probably know more than I do) to explore their own ideas or biases. And I almost always learn something along the way...
Like I said before, I have this weird fascination with certain aspects of human behaviour, especially the cognitive ones.

I've participated in a few informal listening tests between friends. I don't remember being blown away by SR comparisons. At least nothing that was obvious and clear like a hammer blowing your head. I remember not being sure of actually being able to tell if I heard something sounding "better", or I simply focused on different aspects of the sound on each play... Or maybe I moved an inch away from the sweet spot... Or maybe I had cecina for lunch (salted meat) and my blood pressure got temporarily higher.
The point is, nobody on those tests could be consistent, and our answers were indistinguishable from a random answer. The tests were definitely not set rigorously, BTW (just friends being nerds).

I trust we all will eventually discover the (definitive?) answers to these questions eventually, as science and technology develop more.

It's been an interesting thread, for sure

Amusing conversation. Debate, if you will, and seemingly one we've been conducting with regularity for decades. I've been on various sides of this over the years, sometimes arguing for 88.2K or 96K, and sometimes for 48K. (I use 48K fairly regularly now)

I think there is one thing we ALL can agree on: 44.1K sounds worse than any of the higher sample rate options, limiting the minimum rates to which most of us would consent at 48K/24bits.

As for the rest of the arguments, just be glad that we CAN debate this, and that we are not limited to what Philips chose for us back in the 1970s, when they were determining what would become the standard for digital audio.

Most of us agree that the difference between 44.1K and 48K is profound, especially if you factor in 16 bit "CD Quality" vs. 24 bit or 32 bit FP. So, what the debate seems to center on, now, are the differences between 48K and 88.2K or 96K. I think people present valid reasons for all of the above. It comes down to a choice. There is no need for scientific proof. In fact, science kind of gets in the way of understanding this stuff at this level — caught between the technical realities of mixing, sharing, and archiving vs. the reality of human hearing and people's choices of phonographic media for replaying what we create. There are simply too many right ways and wrong ways, depending on too many factors to ever hope to test in any consistent scientific way.

So, it's been a nice debate, but not a one of us has been changed by it; in fact, we've used it as yet another platform to clarify to ourselves and others why we do what we do. That's ok. I expect the identical debate to be continuing a decade from now, although the numbers may have changed by then. Still, it will continue.

I'm reminded of the old RCA logo, which showed the dog looking into a phonograph, head slightly cocked, with the label "His Master's Voice." Except now it's not his master's voice that's got his head cocked. It's the aliasing of the higher partials of his master's voice. Perhaps for modern times they should add above his head a little comic dialog balloon with "W T F?" in it. Again, we all agree that aliasing occurs. It's measurable. So, it's the degree to which we can eliminate it that matters. Maybe anti-aliasing is enough. Maybe 192K sampling is better. But let's agree on one more thing:

• What's been working for you is probably good enough, unless your listeners are also cocking their heads like that dog.

Shooshie

I remember offering a mastering guy gum after lunch, which he refused on the basis that the chewing affected his high frequency hearing depending on where he was in the chew cycle. Now that's extreme.

I was just reading a bit this morning in a thread about archivists doing needle drops of ancient material for restoration, with lots of comments about improvements in results at high sample rate, with regards to transparency of restoration processing.

Actually the gum anecdote is not that far fetched. Chewing can impact hearing acuity by creating noise as well as distort your perception of what you're hearing. Another thing people often overlook is ceiling fans literally "chopping" the sound into chunks of dynamic waves. And then there is the discussion about speakers, wires, wall treatments, and other considerations in reproduction and the listening environment.

Hope I haven't chewed your ear off... ha ha ha!