Digital....Disappointment

[mshilarious]

I only see two ways to resolve the issue of whether or not this is a technical issue: have Fleetwood Mac record in your studio, or shell out the bucks to record in a top-tier pro studio. Once we get the same musicians using two different setups, then we'll know.

I'd like to help, but Christie McVie and I had a falling out a few years ago, and now she won't return my phone calls

Getting back to technicalities, there are a few cost barriers that never go away. For high-end tube gear, you need a large PSU transformer and a good signal path tranny too. You can blow over $100 right there. Add $20 for a decent tube (or two), the cost of 1% resistors, big fat caps for the PSU, a nice box and a VU meter and pretty soon you're at $300/channel just for parts, and a retail price well over a grand.

Converters though, I'm not sure what the barrier is there. Probably just volume. Sure, the analog path components are a little more costly for nice ones, but nothing like a tube circuit.

 

[mshilarious]

The sample rate for digital is still inferior in the upper registers. There are transients and harmonics digital will not even come close to capturing until we get more towards the megahertz sampling range.

Second time I've posted this today, but Dan Lavry disagrees with you:

http://recforums.prosoundweb.com/index.php/t/2997/0

Tape also has a natural compression to it and you also don't have to deal with messy A/D D/A conversion. Only the most expensive D/A converters get close to representing the true signal. Digital sounds crisp and clean, but there is still alot missing.

OK these are two conflicting statements. If tape has natural compression (which it does), then it doesn't represent the true signal. Then you say digital sounds crisp, which I have to presume means transient response and/or high frequency response. So what is missing? Desirable analog distortion? OK, but let's not mistake our inherited biases for a technical deficiency of digital audio.

The most expensive converters get close to the true signal not because of a difference in sample rate--there is none--but because of pristine analog signal paths and very low jitter.

 

[c7sus]

Yeah, this thread has been romin around some as normal conversation has a tendency to do as well.
To wrap this up what this really is about is Fidelity. Plain and simple. My digital rig and probably most of everyone elses in the home camp is no where near the best analog studio of recent years back was. How important is that in the world of MP3 and IPODs? Probably not very.
Now does NOT having the super analog fidelity as that on the Tango recording mean all is lost? Not at all. For instance, The Mamas And The Papas recordings were not really great fidelity wise. If you compare the M + Ps to Tango the difference is plainly there. Its the musical nature of the performance that shines through and no one cares about the fidelity. If you were listening to the Ms + Ps or Tango's Seven Wonders on your car AM radio you would enjoy them equally, even with some power line static in the background. The lack of pristine, mastered fidelity will not stop anyone from creating truly musical art or your listener from enjoying it. I was a little dissapointed that after my research and hard work with this, its still not even in the same sandbox as a good analog studio. Those guys that make the really high end gear know stuff about fidelity that has not leaked out into the main stream yet. Their monopoly on that and low volumes of hand made gear keep the cost of it high. As time goes on I believe the propietary tricks they apply to add that extra edge of fidelity will be common place. The fidelity of home recordings has gone up some from five years ago. It is difficult to overcome the cost of manufacturing great tube gear that operate at proper voltage levels so I don't forsee change there. I do expect the fidelity issue will end up improving on the semiconductor side similar to how other product areas of consumer electronics have improved as the technology evolved. And make no bones about it, home recording is swiftly becoming consumerized - gee..like the personal PC, once the domain of geeks who used slow and cryptic machines in college labs. This area is only getting better folks. We're going from the ice box age to the refrigerator age. And the things a refrig can do these days!

Keep trackin'
Bob the mod guy

Man, you need to listen to the Anthology Of American Music on Folkways recorded by John and Alan Lomax.

It ain't all about fidelity.

It's lot's of little touches. Garth Hudson adding ring modulator to horn parts or Brian Wilson or Aaron Neville stacking vox and putting down what THEY hear in their head. You can't buy that anywhere.

Keep swinging.

 

[Guitar Jim]

Ok, lets say that right now all this ultimate fidelity that some folks seek somehow becomes instantly available and affordable for all........what then?
Will a poor composition suddenly become better?
Will a less than adequate voice somehow improve?
Will an instrumentalist gain greater command of an instrument?
Will a performer gain the necessary "charisma" for success?

It kinda reminds me of a millionaire who wishes to retire for good after making the first million bucks.... does he retire? Of course not..... he then longs to make the second and then third and then fourth million etc etc.
It's the same for audio quality....... the "technician" will never be happy. He will always seek improvement, no matter how good it gets, because there will always be the next challenge or hurdle to overcome.

I believe, and this is just my personal opinion ( I'm not criticising anyone for their opinions ), that the real underlying issue involved here is in fact not audio fidelity...... but rather musical creativity.
Musical creativity can overcome virtually anything whereas perfect audio fidelity will never save a composition or performance that doesn't cut it. Just my opinion though.

 

[PhilGood]

Second time I've posted this today, but Dan Lavry disagrees with you:

http://recforums.prosoundweb.com/index.php/t/2997/0

mshilarious,

That was a great read! Sorry I didn't catch the previous posting, but this article makes it much clearer. Thank you!

The basis for my earlier statement was a conversation I had with Tony Merrill in his car some years back talking about how he hated CDs compared to vinyl. He gave me his theories and I jotted them down mentally. I'm in the computer industry and always referred back to our conversations when thinking about sampling. I'm sure he's aware of all this now, but I haven't spoken to him in years either. We were both pretty young then. I'm not making excuses for my ignorance, nor putting any blame on Tony.

As a person dealing with computers I discovered some moot points that Mr. Larvey talks about concerning file size and transmission speeds, but the rest of the article covers all the bases for the Nyquist theory. It's really solid. I'll have to pass this info along to some people I know.

My comments about transient responses refer to the fact that some transients may happen to fast to be captured accurately by digital equipment. This article bases all it's samples on generated wave function and not real world signal effects. It's like saying a mics response chart tells you how a mic will sound compared to another when we know this isn't true. When you start recording a performance there are some signals that don't behave the same as a laboratory test. The article seems to expain this well enough, though.

Again, thank you! I keep finding this forum more and more valuable in my education.

 

[mshilarious]

mshilarious,

That was a great read! Sorry I didn't catch the previous posting, but this article makes it much clearer. Thank you!

Sorry, I was getting annoyed at Rado, didn't mean to take it out on you.

My comments about transient responses refer to the fact that some transients may happen to fast to be captured accurately by digital equipment. This article bases all it's samples on generated wave function and not real world signal effects. It's like saying a mics response chart tells you how a mic will sound compared to another when we know this isn't true. When you start recording a performance there are some signals that don't behave the same as a laboratory test. The article seems to expain this well enough, though.

OK, that is a critical point: no audible transient can be comprised of an ultrasonic frequency. Thus is it entirely appropriate to use sine waves when modeling digital audio.

I'll look for a link on that later.

Also on rereading his paper I might take back my comment on the relative cost of a high-quality analog circuit in a converter. Lavry's design shown in the paper uses 3 OPA627s per channel

 

[wilkee]

The strange thing is if I record one off my prized LP's onto my hard drive using an Emu 1212M soundcard and a decent turntable/cartridge, when I play it back it sounds just like a great analogue recording.

Its all just down to producers/performers, IMO Pro-Tools just makes it far too simple for the producer to Fu*K It Up which most are only too happy to oblige.


Tony

 

[PhilGood]

Sorry, I was getting annoyed at Rado, didn't mean to take it out on you.

Just looked at some of those posts. That guy's thick!! I don't think I could take the word of someone with spelling that bad.

OK, that is a critical point: no audible transient can be comprised of an ultrasonic frequency. Thus is it entirely appropriate to use sine waves when modeling digital audio.

I'll look for a link on that later.

Also on rereading his paper I might take back my comment on the relative cost of a high-quality analog circuit in a converter. Lavry's design shown in the paper uses 3 OPA627s per channel

Question: Given a high frequency wave at random timing (as in a live performance), say 20khz, at 2 samples per wave, how can you be certain you are sampling the wave at its apogee?

 

[mshilarious]

Question: Given a high frequency wave at random timing (as in a live performance), say 20khz, at 2 samples per wave, how can you be certain you are sampling the wave at its apogee?

Answer: you can't, and it doesn't matter. We are sampling that wave twice, which isn't going to give use a nice smooth wave, it is? It doesn't matter if we hit the apogee and null or two other points somewhere in the middle--the resulting 'digital' wave is going to look like a stairstep, or a series of square waves. As you know, square waves sound bad.

However, a square wave can be broken down into component sine waves: a fundamental, and overtones. Thus, although a sample of a 20kHz sine wave would look pretty bad, when played back it would sound like a 20kHz sine wave with a bunch of ultrasonic overtones.

Those overtones are eliminated by the DAC's oversampling anti-imaging filter, so they are not present in the analog output.

 

[PhilGood]

Answer: you can't, and it doesn't matter. We are sampling that wave twice, which isn't going to give use a nice smooth wave, it is? It doesn't matter if we hit the apogee and null or two other points somewhere in the middle--the resulting 'digital' wave is going to look like a stairstep, or a series of square waves. As you know, square waves sound bad.

However, a square wave can be broken down into component sine waves: a fundamental, and overtones. Thus, although a sample of a 20kHz sine wave would look pretty bad, when played back it would sound like a 20kHz sine wave with a bunch of ultrasonic overtones.

Those overtones are eliminated by the DAC's oversampling anti-imaging filter, so they are not present in the analog output.

OK. I understand that. However, if you sample the wave at somewhere between the null and the apogee, aren't you decreasing the amplitude?

(This should probably be on another thread if it gets longer. I'll stop asking questions and let it get back to the original topic.)

 

[mshilarious]

OK. I understand that. However, if you sample the wave at somewhere between the null and the apogee, aren't you decreasing the amplitude?

(This should probably be on another thread if it gets longer. I'll stop asking questions and let it get back to the original topic.)

Well it's still sort of the same topic. Your questions are about Nyquist theory principles, so you might want to check out a book on digital audio theory.

To answer your question, yes, there is attenuation at and below the Nyquist frequency due to the anti-imaging filter, which is why the sample rate is a bit higher than 2x the audio band.

 

[Bob's Mods]

So it would seem the accuracy at the higher end of the sample rate spectrum is reduced, which is a form of distortion of the original waveform. Hence it won't sound axactly the same on playback as it the original take. This would help to account for the improved high end definition of a higher sample rate.
I wonder what the bandwidth is of a good analog studio is? From -3 dB point to -3 dB point in a quality tape based system the waveforms are captured in their true likeness. In a digital world, the waveform is only a collection of points rather than a mirror image copy of the original so it is easy to understand how the accuracy begins to be reduced by some factor as one approaches the upper bandwidth limit. One of the inherent deficiencies of digital that analog does not share.
To my ears, 44.1 always sounded warmer than 48 and this could very well be from the reduced accuracy in the upper range of the 44.1 sample rate. Very interesting point.
Digital is a bunch of dots, similar too a childs connect the dots book (hokey example). When the dots are connected, an image appears, the more dots, the more accurate the image. Its still dots though. A picture would be a truer copy of the original and this would be an analog medium.

Bob

 

[lpdeluxe]

Still looking at this as a technology issue, aren't we?

Good point, along with many other good points, Mr Chessrock.

For me, I stayed away from this thread as long as I could, because here's the issue:

"I am SO disappointed in digital!"

--meaning, why doesn't this dadblamed plastic thing I bought do everything for me? What the hell's wrong?

There's only, ever, one reason your recordings are disappointing:

You screwed up.

I get what I consider to be wonderful results with digital. So sue me. I also spend a lot of money and a lot of time on gear, coaching musicians, mic choice and position, and you name it. Yeah, I'm disappointed a lot, too, but dammit, I'm mature enough to know the name of the guy responsible. It ain't the arrow, it's the Indian, and I'm the Indian. God grant that before I die I make a recording I'm satisfied with.

Every disappointment leads to the next breakthrough.

Or do you believe in the tooth fairy?

 

[paddyponchero]

All equipment introduces distortion, the highly prized analogue classics happen to do it in a way thats pleasing to the ear.

Now if only there were an easy way to cut and paste 2" tape I'd certainly consider that route

 

[Bob's Mods]

Maybe I need to clear the air here. My original post was a technical statement related to an observed response when recording tracks on my own digital system and how they compared to those recorded in a great analog studio. The quality of the musical pieces on the medium is another subject. I happen to be content with my musical skills.
Everyone has a different place where they set the limbo bar when it comes to their gear. There are always posts about who likes what mic and what pre and why. A mic, a pre, is technology that translates into a paint brush for the artist. A clave is an instrument, a musical paint brush for an artist to use to sculpt his sound. Gentlemen, all these things are tools and each one of us has differing expectations from them. Ipdeluxe hasn't spent alot of time and money on his gear and art because any old thing will do. He knows his expectations and he keeps fiddlin until he achieves whats in his head. Those who do not I suspect are satisfied with mediocrity.
If gear doesn't matter then maybe we should all be recording with SB16s, microphones that came with cassette recorders and Behringer modeling amps.
Gear does matter to anyone who cares about his sound.

Enjoy,
Bob

 

[mshilarious]

So it would seem the accuracy at the higher end of the sample rate spectrum is reduced, which is a form of distortion of the original waveform. Hence it won't sound axactly the same on playback as it the original take. This would help to account for the improved high end definition of a higher sample rate.

It could; it depends on the implementation of the converter. A crap converter is going to suck more than the theory describes at 44.1 and 96, so higher sample rates aren't a panacea.

I'm still going back to the analog front end, that's where most of the high-end is going to be lost. If your mic loses 6dB, and your pre loses 3dB, then that 3dB rolloff at 20kHz might be the final straw. However if you preserve the high end going to the converter, then your final product shouldn't suffer.

It might be illustrative to run a 16kHz or higher test tone through your entire signal chain and see what happens.

 

[freshmattyp]

Bob - are you mastering your own stuff at home or sending it out for professional mastering?

 

[chessrock]

My original post was a technical statement related to an observed response when recording tracks on my own digital system and how they compared to those recorded in a great analog studio.

Yea, but what some of us are having a hard time with is that you're comparing your own work to those done in high end comercial studios.

... and somehow you've made this in to an "analog / digital" thing.

It's just kind of a leap, that's all. I mean ... come on. You're not a professional audio engineer or producer. You're not running a comercial facility. I'm assuming you don't have the kind of budget Fleetwood Mac had when recording Tango.

Give yourself a more realistic benchmark for comparison.

 

[Bob's Mods]

Fresh, its just stuff I do at home.

Chess, those are observations I haven't considered. My original thinking was that mods and a pristine audio chain should give the definition I'm hearing on mixes that were recorded in good studios. The local folk station here in Boston is always playing mixes that were done by folk artists who recorded in a real studio. Alot of struggling artists still use studios rather than record at home. Those mixes are really good. No doubt digital studios and probably mastered too. I don't give up easy though. My gut tells me that the pristine quality of really high end can be had on modest budget. Its easy enough check out my recording chain with a frequency generator. That will tell pretty quickly where any problem areas are. mshilarious thinks looking for a roll off is a good idea and I do too.

Don't get me wrong, my stuff does sound respectably good. Hell, never in my wildest wet dreams could I have imagined I could play in my own band, sing, write, produce, arrange and have any instrument at my disposal...all by myself! Midi and digital audio make it possible. Its the perfect hobby for someone who likes music and technology. It keeps me out of trouble too.

Bob

 

[Guitar Jim]

By the way Bob, I think that maybe my meagre contribution on this subject with my several posts may have sounded like I was critical of you. I didn't intend it that way. I suspect that when we listen to a recording we probably "hear" it in a slightly different way. I'm a right brained person..... I can barely change a washer on a tap.