Live Mike Feeds and Consumer Audio
JPH-22 -- Mon, 03/23/2009 - 01:03
It was an interesting piece, one that I never gave much thought - live-mike feeds. These are not the absolute reference that some claim they are - or are they ? I mean, should we even be *striving* for mike feeds ? Here are my thoughts:
High-end audio has been working with signal quality about three levels down from the live reference. We have acoustic sounds, mike feeds, analog tape...then consumer media. Sad to say, but listeners were not even getting *close* to true references, on signal. Nevermind the losses which occur throughout the playback chain !! But in the last few years - and only with the best analog and digital equipment - could we finally match the sound of analog tape, in my opinion.
Upon this advance, the next logical step was to achieve the "live mike feed" quality that RH mentions so much. And maybe, without a moment's notice, we're there already. Reports are coming in of yet another leap up in source component sound - but this time it's *digital* that's blazing the path. The big news with digital has been the downloadable hi-rez files (like from HD Tracks) that have been getting lots of attention. But surprisingly, it's also the old digital standard that so many declared "low-fi" and "not nearly good enough" that's marching towards live-feed quality as well !! In this month's Stereophile, John Atkinson states that Meridian's 808.2 CD player was in a virtual tie with true hi-rez feeds. Did anyone think this was possible ? RH didn't....and is sure to contest this statement.
The happiest thing about all this is that we have digital signals sounding damn close to mike feeds WITH A FLAWED PLAYBACK SYSTEM. I won't go on a rant here, but let's just say that our playback systems, mainly the loudspeakers, are blocking us from hearing whatever is on the recording. Looking ahead, if speakers were to take a leap of the same order as digital in recent years, we all might be in for a surprise.
These playback issues, BTW, also affect the "testing" of mike feeds and recordings in the studio. We don't really know what's on a well-made recording...being that the rest of the playback chain is so (currently) corrupting.
If the latest Meridian CD player can correct errors made at the *encoder*, then anything is possible. So let's not bad-mouth our 20-bit PCM recordings made since the early 1990's, even as stored on CD. These very well could have live-mike feed quality - or beyond that - buried in them...........
And maybe we should start considering horn mikes. There's already one co. in Poland making them.
JPH
All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
I'm looking forward to hearing the Meridian 808.2; I'm writing a comment to accompany Sue Kraft's review of the 808.2 and Meridian's DSP7200 active digital loudspeakers.
Hi JPH-22,
Actually, it isn't the "old digital standard that so many declared 'low-fi' and 'not nearly good enough' that's marching towards live-feed quality". That old standard, 16-bit, 44.1k sampled PCM is, to my mind, today's "cassette". It is the new digital standard of 24-bit recordings sampled at 4x rates (176.4k and 192k) with the best of today's A-D converters.
Your question "should we be striving for mic feeds?" is, I think, a very good one.
My perspective is that ultimately, the playback system's job is not to reproduce The Absolute Sound. Its job is to reproduce the recording.
It is the recording's job to capture The Absolute Sound (unless of course the producer and artists are seeking something else, as is quite common today). For the sake of this discussion, I would say achieving a recording that sounds like the mic feed is a gigantic step forward. There is a sense of Life in a good mic feed (even in a not-so-good mic feed) that until recently, has always been obscured by the best digital and even the best analog recorders.
Now that we can achieve "mic feed" quality with the best current converters working at 24/176.4k or 24/192, what I've long thought of as "the last frontier" comes once again to the fore: the recording itself. To me, this is usually a far weaker link in the playback chain than any electronic component, source component, loudspeaker, cable, etc.. I've always found it interesting that a really good recording is identifiable as such, regardless of the system it is played back on. Even in the car. Even a (well made) cassette copy or (well made) MP3 (!) played in the car can still reveal something positive about a really good recording.
In my experience, 90-95% of the ultimate sound we hear has already been determined by the time the signal leaves the microphones.
This is part of what motivated me to start my own label.
Horn mics? This is new to me. Not sure why anyone would want to have a horn on a mic. I lean toward small diaphragm omnis out in the open. Seems to me a horn would only add multipath at the mic's capsule. Perhaps I'm not picturing it correctly. Must look into this.
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
Barry,
Apparently, you are not aware of what's happening with state-of-the-art CD playback. I am not alone in my assessment - the editor of Stereophile says this month that Meridian's 808.2 CD player (with good recordings) was virtually indistinguishable from 24-bit hi-rez audio. More and more listeners are saying the same thing - hi-rez is not better.
The key to all this is the recording - from good labels like Chesky, Hyperion and Harmonia Mundi. They record at 20-bits then use dither to drop the (perceived) noise floor down a notch or two, on release. I'm sure you're aware of this process. As a result, CD is not low resolution. It never was anyway - I mean, 16-bits is already a good deal more than a symphony orchestra can produce...and higher sampling rates like DSD's, load the spectrum with ultra-sonic noise. This is one reason SACD does not sound better than a well-made CD.
In the end, however, it was "mastering losses" that hurt the 16-bit recording standard, so they raised it...and subsequently dithered the consumer copy. This solved the problem, more than 15 years ago.
If mikes themselves are lossy, then I guess high-end will never achieve its goal of live acoustic sound. Even if "mike feed quality" will be the next best thing. The horn mike BTW, is from Simpson. It could be a breakthrough.
Hi JPH-22,
Would it be possible for you to state your point of view without telling me about what I have and have not experienced?
Unless of course, you have been at the listening sessions. I don't recall seeing you there but perhaps I just missed you.
Do you have any idea of what I use for CD playback? Clearly you must if you are going to tell me I am "not aware of what's happening with state-of-the-art CD playback". Since I have not disclosed the particulars, when did you visit my studio without my knowing?
And have you heard the Meridien yourself? If so, what other devices have you compared it to that led you to reach your conclusion? What CD masters did you have on hand to compare with the playback to verify its accuracy?
For someone with your extensive knowledge, there are a few things you've neglected. First, no one uses 20-bit ADCs any more. Ask Chesky, Hyperion and Harmonia Mundi and you'll find they're using 24-bit ADCs. Chesky for one, also releases high res versions on HDTracks. Ask David Chesky if he thinks there is no significant difference. (I've spoken with him and already know his feelings in this regard.)
it appears you are also assuming digital word length is related only to signal-to-noise ratio or dynamic range. Its primary advantage is resolution. How many recordings have you compared using multiple digital captures of (say, 16/44 vs. 24/192, etc.)? How many recordings have you made where, in the interests of preserving full dynamic range, you allow the low level sounds to be captured at a low level and then compare the sound of those low level instruments in a 16-bit version vs. a 24-bit version? I'm interested in your direct experiences in this regard and how you reached your conclusions based upon these.
Are you aware that a low level sound, say something that registers at -30 dBFS (not at all uncommon in an uncompressed recording of dynamic music) will effectively use only 11 bits of your 16-bit converter, while still effectively using 19-bits of your 24-bit converter?
How about the software used to record, mix and master? How many applications have you compared? Do you hear differences in the sound of these, using the same source file with each of them?
Perhaps you do not hear a difference between 16/44 recordings and 24/96 or 24/192 recordings. That's fine.
You know what you hear or do not hear. What you don't know is what I hear or do not hear.
As to "mastering losses", how many originals have you compared with finished product? How many different replication facilities have you sent identical masters to and later compared what came back with what you sent the plants? How many dithering algorithms have you compared and what did you find in those comparisons? Do you know the sort of dithering algorithms that existed "15 years ago" when you say the "problem" was "solved"? What sort of converters were being used 15 years ago and how did they solve the problem?
You make some sweeping generalizations about what is possible and what is not, what is an advantage and what is not. I would very much like to know the basis for these. By the way, ask John Atkinson what resolution he records in and how he processes. Ask him, regardless of the Meridien review, which he would rather listen to: 16/44 or 24/96. Please let me know what he tells you (and I'll see if it is the same thing he told me).
I have no argument with your sensibilities and your preferences. Please understand the difference between personal preferences or personal experience and Universal Truth. I don't know anyone who has a corner on the latter.
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
16-bits capture about 25db more dynamic range than a symphony orchestra produces - fact. But it still wasn't enough - due to the intermediate calculations done after the original capture - what I call "mastering losses". 20-bit recording, begun in the early 1990's, solved the problem - along with dithering the consumer copy to complete the process.
And please tell me how resolution is *not* related to dynamic range and S/N - I'd like to hear it.
There is no such thing as 24-bit ADC's - they are only functional to 20-bits. You didn't know this ?
If CD was lower in sound, why does John Atkinson put CD (playback) on equal footing with hi-rez ? Why are many other reviewers saying the same thing ? You apparently are not aware of the hot happenings at the playback end, nor do you seem to care......
I will not argue with you JPH-22.
You have not answered a single one of my questions, you invoke John Atkinson's viewpoints based on one review instead of actually speaking with the man, etc. Perhaps you don't want to hear his answer and prefer to ascribe to him your own interpretation, as you prefer to declare your listening biases and level of acuity as being Universal Truth.
I will defer to your extensive (though undocumented and unsupported by your responses) experience and knowledge of the recording and mastering process.
And I guess I'll have to simply ignore future posts from you until and unless you learn some manners.
Over and out,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
I must thank you, Barry, for saving me about an hour of writing.
I don't think I could have said things any better than you did.
JPH-22 has been making the same sweeping statements for several months, and regardless of how many times other posters site sources and information that contradicts his statements, he continues to forge ahead undeterred by facts counter to his arguments...
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
Both of you,
Please tell me we're I'm wrong. Wrong about the *real* performance of 24-bit ADC's ? Wrong about the dynamic range of a symphony orchestra vs. that of a 16-bit capture ? Wrong about dithering (re-dithering) of a recording to drop its noise floor ? Wrong about reviewers opinions concerning CD vs. hi-rez ?
On this last point, Atkinson is not alone - read reviews of Metronome, Zanden and Reimyo CD systems (inc. online) over the past several years. Hi-rez, according to them, is not better on playback.
I don't know what brings on your denial of CD's sound. But I've noticed that one thing you have in common is that you're both recording engineers. Which means you know what's going on in the studio...but in the dark when it comes to home playback.....
Hi Steven,
Thanks for your post. And you're welcome.
JPH-22, aside from not realizing (or apparently caring to realize) how his tone comes across, continues to display what I can only interpret as a lack of confidence in his own listening biases - lest he would simply declare what he does and does not hear, rather than make general pronouncements of Universal Truth (i.e what he does not hear, no one else can possibly hear).
Since he didn't answer a single one of my questions, I can only assume he has no recording experience whatsoever and thus has never spent any time listening to the results from different converters. He sees word length in only one of its many dimensions and even at that, hasn't any idea that a symphony orchestra's quieter moments will give him a 10 or 11-bit recording on a 16-bit converter, even after I tried to point this out to him.
Further, he thinks dither is used in a recording to "drop its noise floor". Perhaps one day he'll spend the ten or fifteen seconds required to read the first paragraph of any article on dither. He just might figure out dither adds noise, hence raises the noise floor.
He uses individual sentences from reviews he has read - but will not inconvenience himself by actually asking the reviewer a direct question as I suggested. (I'm sure John Atkinson and David Chesky receive email.) Then again, asking the question might get him an answer other than the one he wants. So would a minimal amount of actual experience. Wouldn't want to cloud the issue with facts. ;-}
I am not surprised he draws similar thoroughly uninformed conclusions regarding audio engineers. Ignorance by itself is innocent. It is when it mixes with arrogance and rudeness that it becomes sad. He is after all, an Internet audio "cowboy". As Randy Newman sang, "Big hat, no cattle".
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
I can't answer your questions because you're not making any sense - nor are you answering any of *my* questions....
Have you ever heard of the Meridian 618 ? Or Apogee UV-22 ? Or ANY re-dithering device, for that matter ? Apparently not. Because no-one that has used these devices say they *increase* the noise floor. They drop the quantization noise down, via noise shaping, for improved sound. They have been used since 1993 - I don't know what planet you're living on.
And to think that 24-bit ADC's record at 24-bits - WRONG. It's actually 20 - ask Bob Katz. This, BTW, is the maximum level we've recorded at since the early 1990's. Problem solved, years ago.
Lastly, I'm not picking "individual sentances" from the reviews I'm reading. It's the *conclusion* of the reviews that I'm relaying. If you don't want to believe these writers, then don't. But by doing this, you are in denial of what CD is doing on playback.....
"Have you ever heard of the Meridian 618 ? Or Apogee UV-22 ?"
I've used (and owned) both.
Sorry, 88.1/24 and 96/24 from Grace's Lunatec A/D and even the internal A/D in a Marantz PM-6700 is better sounding than any recording (using the same Schoeps microphones) than I've made with my Apogee UV-22 44.1 convertor (even at 20 bits). Not theoretically better (as if I care) but sonically better.
I suspect that Barry has done a few more recordings (and tests of gear in his recording chain) than you have AND than I
have put together...
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
Hi Steven,
I suspect my long dead grandmother has more done more recordings (and tests of gear) than a certain cowboy we know.
It doesn't occur to our range riding raider that like many engineers, I might own several flavors of the Meridian dither/noise shaping (bet he didn't know it comes in flavors) along with Apogee, Pow-R, MBIT and about a dozen others, that he has either only read about or never heard of at all.
He doesn't answer a single one of my questions because both he and I know what his answers will be (as I'm sure, would anyone else who has read his posts). He just doesn't get it and rather than try to see that, he digs in harder. The arrogance makes me now think of lyrics in a composition by Mingus: "sad clown with the circus closed down"
But enough of rodeo clowns.
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
So you (both) have used these dither boxes, eh ? That means they work. Why would you use them if they didn't ?
At this point, most listeners (not recording people) are saying that CD is as good as - or better than - hi-rez (Soundstage on Zanden, Positive Feedback on Metronome, Atkinson on Meridian, etc.). You are both in denial and are not even *listening* to what quality CD playback is doing. Right ?
You may disagree with me...but do you disagree with Atkinson ? How about TAS's Jon Valin and Anthony Cordesman putting CD, not hi-rez, on equal footing with the world's best LP playback ? And Michael Fremer at Stereophile doing the same ? Remember, hi-rez is 10 years old - why are so many disagreeing with your agrument ?
And for a recording engineer to think that dither *raises* the noise floor when it does just the opposite (via noise-shaping) is just plain silly. Then to think that 24-bit ADCs record with 24-bit resolution when THEY'RE ONLY FUNCTIONAL TO 20 means that this person is not a well-informed audio engineer. I wonder what else he doesn't know.......
Your flights of logic defy rational discourse....
I used the Meridian 518 and Apogee UV-22 because they were the best available AT THE TIME. That time is LONG past...
Judging by your responses to our arguments, YOU are the one in denial. 44.1/16 is only a medium of convenience and standardization, not optimal sonic quality. Any other conclusion is merely wishful thinking.
Both Barry and I make recordings which we listen to very carefully, more carefully, I dare say, than most recreational listeners. We both know how the mic feed and the original room sounded when the recording was made. Do you, or any other recreational listener have that same information for reference? If not, please stop telling us what is better, because frankly, you don't know.
LP playback - listening to ancient recordings that may or may not be properly EQ'd and mastered, doesn't interest or concern me much. With the exception of my ex-boss John Atkinson (for whom I still have the utmost respect as a sonic critic and recording engineer), none of the pundits you've mentioned make their own recordings (high resolution or otherwise),so their opinions are irrelevant to our argument. And why, if JA likes conventional CDs so much, does he bring the original high-resolution 88.1 24 bit WAV files of his recordings to CES to listen to gear?
I know this because we sat together and listened to his latest recordings together at CES...they sounded wonderful - better I dare say than the down-sampled 44.1/16 versions he could have played.
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
Some people in that group at RMAF - I know for a fact - said that JA's hi-rez files sounded *different*, not better, than dithered CD. Some admittedly did say that hi-rez was better. But now JA, who you respect, says that Meridian CD is virtually the same as hi-rez. Who's in denial, again ??!!
And I - along with many other audiophiles - are much more than "recreational listeners". We listen to a lot more playback equipment than you do because that's all we do - is *listen* to the results. This, through cutting edge (CD) playback gear. And like the audio writers who heard this stuff, I do not agree that hi-rez is better on playback - agrue all you want. And even if they're listening to LP, they *still say* that dithered CD is better than hi-rez. Their words, not mine.
If you trust Atkinson, then read this (esp. paragraph six):
http://stereophile.com/asweseeit/907awsi/
We don't have to be "recording engineers" to hear the results of a recording process. If anything, you guys are more on the *objective* side, not us. Subjective criticism reins supreme - and on this score, I won't even waste my time debating.....
Heresay from RMAF is not the same as actually hearing the recordings, is it?
I realize that you will never admit that you are as ignorant as you appear to be.
Actually, you DO have to be a recording engineer to hear what is really happening in the recording process. Otherwise you're just guessing, and in your case, guessing wrong.
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
So Atkinson is wrong in his '07 editorial ? And NOW on Meridian ?? That's ignorance 101.
And audiophile's opinions don't matter ? That's ignorance 102.
You DON'T have to be a recording engineer to hear what matters most - the final product. If you want to be in the minority concerning playback - so be it. That suits me just fine.......
Atkinson's editorial is about disk-based consumer formats of hi-rez SACD and DVD-A, not WAV files.
I have lived with the Meridian 800 series gear for years. Its very good on 44.1 , but not as good as high rez WAV files directly from a recorder or PC source, sorry....
And yes, you do need to know how the original mic feed sounded to be able to judge the recording's ultimate fidelity to that source. Otherwise you are STILL just guessing, but then you seem to excel at guessing.
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
So now it's about the physical properties of the media ? Mmm....
Instead of guessing on their intent, I'd like to see the writers *themselves* say what you think they mean.
Besides, wouldn't hardisk benefit *CD* as well ? I mean, much of the Stereophile crew seems to be using the Soolos music server and are (all) quite happy with Red Book.
If I'm wrong, then all these other listeners (who you dismiss) are wrong as well. Until there is a consensus that hi-rez is better on playback - and there isn't - then I stand by my statement concerning CD. I go with the majority........
You didn't read the editorial very well...it is about the commercial success of high def media, not the quality of the sound of high def.
You MUST have been a college debater since you inevitably change the subject whenever you can't make a "point" with your arguments. Anyone with a modicum of reading comprehension skills will see as they view this thread how desperate you are to be right, yet how miserably you've failed in that regard.
Listen to your CDs. Enjoy them. But don't be so ignorant as to assume they are the best sound you will ever hear...they are already as obsolete as a buggy whip.
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
Please - for God's sakes - read paragraph six from JA's editorial. Hi-rez, according to him, had issues with *sound quality* on recording and playback. You're guessing (again) at Atkinson's intent - now you say it was about "commercial success". Before, it was about "disc media only". Talk about putting words in people's mouth !!
And why, in 2009, is Atkinson putting Meridian CD on a level playing-field with hi-rez ? You haven't answered that one yet - but I'm sure you'll make up another reason......
Yes, I will listen to my CD's - THROUGH A HIGH-END MUSIC SERVER, just like the crew at Stereophile.
If you want to discuss the pro-end of things in audio, I suggest you go to another forum. This one is for audiophiles and their home sound systems.......
Thank you for telling me where to go.
And BTW if you read the latest Atkinson review of the Meridian 802 carefully you'll comprehend that he clearly states that it does not sound as good on 44.1 material as that same material on his 88.2 master tapes.
I would tell you where to go as well, but why bother? You're already so lost that clear directions won't help...
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
It see that it was "hard to tell" the difference between CD and hi-rez. Only towards the end of the review does JA state that hi-rez was "slightly better". But it seemed like quite a stretch for him to say that.
But like always, you draw your own conclusions....
If you want to discuss the pro-end of things in audio, I suggest you go to another forum. This one is for audiophiles and their home sound systems.......
Yes, I will listen to my CD's - THROUGH A HIGH-END MUSIC SERVER
Well, hi-rez for the home is here for those who want to bother a bit, and around the corner for those who do not want to bother. I have no intention whatsoever to put my CDs on a music server: it's a bore and I would rather spend my free time listening to music rather than feeding a server. I will get a server for hi-res music when the material released is more to my liking. I've heard hi-rez a few times, some in controlled a/b environments, and it was obvious to anyone present that this is way way better than the CD or than the CD ripped on the same music server.
"it was obvious that (hi-rez) was way way better than CD"...
I just wish more would agree.....
Hardisk is better on playback - that's the big reason why audiophiles are using them. Less jitter. Read reviews.......
I have on my fanless, drive-less PC server the same tracks at 44.1kHz/16-bit with HDCD encoding and the original 176.4kHz/24-bit files from which the CD-compatible signal was derived. One example is the third movement from Rachmaninoff's Symphonic Dances. I can immediately switch back and forth between them at matched levels, decoded by the same D/A converter. It's fair to say that the 44.1kHz/16-bit version is, if not the state-of-the-art, at least very close to the state-of-the-art of what 44.1kHz/16-bit can do. It was derived from a high-res source, converted with the best possible technology, and enjoys the huge advantage of HDCD encoding and decoding. It also has the advantage of being read back from solid-state memory, not a spinning optical disc. If only all 44.1kHz/16-bit sources were made with this level of care.
Nonetheless, the 176.4kHz/16-bit playback is significantly better in low-level resolution, realism in timbre, spatial recreation (soundstage size, depth, air between images, precise placement of images in the soundstage), dynamics, greater clarity of individual musical lines during dense passages, and an overall liquidity to the sound. Beyond these specific sonic attributes, the high-res is more involving. It promotes a sense of relaxation and immersion that makes me forget about the sound. I've noticed another phenomenon regarding high-res; whatever music I listen to the night before in high-res, I have on my mind—vividly—the next day (and sometimes for several days). Although I often replay in my head music I've recently heard, it is never as powerful or present as when I've listened to the piece in high-res.
Barry touched on an important point that bears repeating: The resolution of a digital system is not determined by the number of available bits, but by the number of bits being used at any particular moment. This is particularly crucial in classical music, where quiet passages might be encoded with just a few bits. A 24-bit system will encode those quiet passages with more bits than would a 16-bit system, thus increasing resolution in a meaningful way.
It's also worth noting that Bob Stuart has written extensively on the subject of how much information needs to be encoded to make the channel transparent. His conclusion is that 44.1kHz/16-bit is insufficient to encode the entire range of sounds humans can hear, thus his championing of high-res PCM and the DVD-Audio format in the late 1990s.
I thought the Spectral CD player you auditioned was close to hi-rez. This would in line with JA's impression of Meridian's 808.2 - saying it was virtually as good as hi-rez.
And there is still a shortage of writers saying that hi-rez is "better than CD", on playback. It seemed - for a while - that lower-priced SACD equipment was better than similar-priced CD. But this has changed due to CD players improving so rapidly.
Maybe we should have ripped and burned DVD-A to hardisk and compared *that* to CD (on the same hardisk) by now. True, DVD-A was encrypted...but this was overcome years ago. Why didn't anyone ever do this ?
Hi JR-1,
I have not encountered the "shortage of writers saying that hi-rez is "better than CD" you mentioned. Every writer I've spoken with who has had the opportunity to make a fair comparison (not as common as one would wish) would say high res is better than CD and that in fact, in the end, there is no comparison, so great are the differences on any good system.
There is no need to rip a DVD-A to hard disk if one does original recordings directly to hard disk. I've often experimented, recording at different sample rates and word lengths (with a number of converters). One great test is to take an excellent analog master tape and encode it several ways (different converters all doing the same sample rate/word length combinations with several variations for each of the converters). Most folks don't get the opportunity to hear such a comparison and fewer have access to an excellent analog master tape. Comparison with the original makes for some very interesting listening. This is what makes me refer to 16/44, good as it can sometimes be, as the "cassette" of the digital world.
JA's complimentary review of the Meridian notwithstanding, asking the man himself for his views on how high-res relates to CD will be much more informative regarding his viewpoints than a review of a single component, written by necessity, in the context of that component.
Same with RH. His complimentary review of the Spectral notwithstanding, we have his take on high-res vs. CD in several posts in this very forum.
No one I've spoken with who has had much exposure to high res, particularly with direct comparisons to 16/44 versions *known* to have been identically mastered (this is rarely the case and is why most folks don't have a basis for fair comparison) would choose the 16/44 version for their own listening.
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
It's not normal to ask "the man himself" concerning reviews. That's why we *have* reviews - to read what they say, in detail. When two (top) audio writers say that CD is close to or virtually the same as hi-rez, we have a problem. And CD is STILL IMPROVING.
As for DVD-A, my question goes to reviewers, who had the time (and equipment) to compare hi-rez and CD. The digital-out limitation was overcome on DVD-A and as such, should have been burned to hard disk and compared to CD - on the same server. I have not heard a cogent reason (yet) why audio writers have avoided doing this.
Hi JR-1,
When I'm curious about someone's perspective, I ask them. I don't know whether this is "normal" or not but it certainly is possible.
The problem with comparing a rip from DVD-A to a rip from CD is that these, much more often than not, are not mastered identically. Often, they are mastered by different engineers. Any comparison therefore, becomes more a comparison of different masterings than it is of the formats themselves.
In an effort to provide something of a start toward a remedy for this, a while back, I put up samples on the Soundkeeper site of the same file, one at 24/96, the other at 16/44. Both are from the same mastering session and the only difference therefore, is the format. To ensure a fair listening comparison, one must know they're DAC is not upsampling, downsampling, applying dither or in any other way altering the native files.
Some folks tell me they can barely hear a difference. Others hear quite significant differences. I hear the latter, specifically in overall focus of the space (the recording was made with only two microphones, direct to hard disk) definition in the bass, harmonic complexity and naturalness in the other stringed instruments and vocals and the overall ease in the sound. Note, this is at 96k and not at 192k, which I use for my current recordings and for which I find the differences appreciably larger.
The samples can be heard here.
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
Barry -
Comparing two different recordings has *always* been a problem. Whenever a reviewer says that one format is better, it's almost never a same-session comparison, I don't think. That's why they listen to many different recordings...and level-match - if they do.....
Hi JR-1,
In my view, level matching is useful for evaluating two different masterings but becomes irrelevant when one is comparing two different masterings in an effort to evaluate two different formats.
It is indeed a problem because the folks pushing certain formats don't appear confident enough to supply a "CD layer" from the same mastering.
Any reviewer that reaches a conclusion about how formats compare by using a different mastering in each format is at best, talking through his hat.
Comparing many different recordings, each a different mastering for each format, does not get any closer to any logically defensible conclusions about the formats. It is simply making the same error many times.
Using many different recordings in a fair comparison is always a good idea as this broadens the scope of any observation. But these must be from the same mastering session if the evaluation is of different formats. Anything less is a comparison of masterings, not formats. Any reviewer that doesn't know the difference needs to learn this before proceeding with a review. And their editors must be asked why they would subject their readership to that sort of incompetence. (In all fairness to the reviewers and editors out there, I believe many --but certainly not all-- of them have not even considered their format comparisons are really mastering comparisons.)
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
Barry -
So, when Robert Harley says that CD is "close" to hi-rez through Spectral, his comparison was flawed. But John Atkinson, who said CD was "very close" to hi-rez through Meridian, *did* use the same master session for his comparison. So his demo was "done right", right ?
But CD was (still) almost as good as hi-rez.....
Hi JR-1,
Why are you arguing? If you enjoy CD and find it virtually equivalent to high resolution, I say enjoy yourself.
See what Robert has to say in his post below. He didn't simply say "close", he said "as close to high res as I've heard from the CD format". There is a difference. Notice the mention of "density of fine detail" missing from the CD, even on the Spectral. Notice the mention of "natural ease" missing from the CD, even on the Spectral. You can take a few words (out of context) from one line of one of his reviews and draw wholesale conclusions from that if you want to. Or you can ask the man himself how he feels --or simply read what he has already posted.
As to John Atkinson, several years ago, he played me one of the first high-res vs. 16/44 comparisons I ever heard. The difference on a cello recording was that the 16/44 sounded like a kazoo in comparison, so devoid of the harmonic complexity was the low res version by comparison.
Send him an email message and ask him. You'll find out he works in high-res because he finds it sounds better.
"Almost as good"? Depends on how you define "almost".
A 1957 Chevy Nova is "almost" as fast as a Ferrari. Almost.
It is certainly as "close" to the Ferrari as I've experienced in an old General Motors economy car. ;-}
Have you compared the files in the link I provided earlier?
Perhaps you'll find the 16/44 almost as good as the 24/96. That's okay.
However you enjoy your music is good. However, if you're going to make pronouncements of Universal Truth (as opposed to how you hear it), I suggest you make a few recordings of your own using some good gear and listening on a good monitoring setup. I'd be interested in what you find.
Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com
John Atkinson was hard pressed to hear a difference. Only at the end did he say hi-rez was "slightly" better. "Slightly" meaning...just what it says. And this while using the same session tapes for comparison !! I know - you'll tell me to "ask the man himself" when this is folly. It's right there in the ink.
And this is right in line with what *I* heard through the same player *and* what Robert Harley said about this player last year in a show report. Are we all wrong ??
The Spectral SDR-4000 playing CDs comes as close to high-res as I've heard from the CD format, but it doesn't have the density of fine detail as the 176.4kHz/24-bit. High-res just has a natural ease that makes you forget the playback system.
Based RH's and JA's reviews, it sounds like CD is creeping up next to hi-rez. This shouldn't be the case - CD was suppose to be "low rez". Remember, CD players can't make a recording sound any better than what's on the disc. Maybe they have fancy ways to describe it ("resolution enhancement") but the bottom line is that they can only *reveal what's buried in the 16 bits*.
The point about burning DVD-A to hardisk / memory is a great one. RH (and others) should have tried this years ago !! Think about it - 24-bit audio is 24-bit audio, no matter how high the ADC's go. Why aren't we doing this ?
There is an easy way to render this discussion academic: lay your hands on a Berkeley AlphaDAC, hook it up to your music server, and play back the HRx sampler files, or the free HDTracks hi-res samplers, and compare them to the equivalent Redbook files. If your system is at all resolving, and you have some experience doing critical listening, you will quickly reach the depressing conclusion that your thousands of recordings on CD are simply not likely to ever sound as good as they could have if encoded in higher-resolution digital.
I speak from personal experience, obviously. I have the advantage of very resolving speakers (Magico V3), but I'm also 50 years old, can't hear a lick above 11 khz, and have slight medication-induced tinnitus in my left ear. Even with those handicaps, I'm telling you, the differences are stark, and very much as Robert describes. Good as the AlphaDAC is with Redbook - and it is very good, quite astonishingly better than my previous Meridian 861 - Redbook is still closed in, opaque, and mechanical by comparison with the hi-res files. By contrast, the one track I have in both 24/176 HRx and 24/96 HDTracks format shows the two hi-res formats to be fairly close - at least to my less educated and older ears.
It looks like it depends on the playback equipment we use. Opinions vary *wildly* right now on CD vs. hi-rez. You say CD sounds "mechanical" - it certainly didn't sound like to me when listening to the Meridian 808.2. Here, Red Book sounded like "hi-rez" - with none of the flaws you describe.......
Well, if I'd not heard the hi-res files, I would have described the CD format playback of the AlphaDAC as "hi-res" too - in fact I don't think even hi-res files sounded as good through my 861. It was only by comparison with hi-res files on the AlphaDAC that the remaining artifacts of Redbook playback became evident. If I hadn't ripped those darn HRx sampler files into my server, I could have happily gone on believing the sound I was getting was state of the art. And don't get me wrong - I'm still happily plowing through my whole collection, spending several hours a day in the listening room, getting sucked into my recordings to a degree I've never experienced before. But that doesn't mean there isn't even better to be had.
It's possible that the 808.2 is that much better again on CD playback - although somehow I doubt it. We'll know soon when RH gets to listen to it. In any case, the relevant test is not how good the 808.2 sounds with CD, it's how close CD playback through the 808.2 is compared to hi-resolution playback through the same unit (I guess it would have to be an 808.2i). Although, I guess that test wouldn't be absolutely definitive, either, because the Meridian won't accept sample rates greater than 96K.
Scott, thank you for your contributions to the discussion.
The only way to determine of something sounds better or worse than something else is to compare it in an A/B manner as you have done.
That's the only way I judge the audio quality of different formats on my own recordings - I make parallel recordings and compare them.
Both Barry Diament and I have tried to make it clear that we both have done these kind of evaluations and found higher resolution recordings to be superior to the same recordings made in 44/16. Robert Harley has also weighed in on his findings, as have you.
But until JR-1 and JPH-22 hear for themselves they will never be convinced.
I fervently hope they both have the opportunity to do their own listening evaluations in the very near future so they won't have to rely on anyone else's opinions...
Steven Stone
Contributor to The Absolute Sound, EnjoytheMusic.com, Vintage Guitar Magazine, and other fine publications
I've heard the Meridian 808.2...at length...and I can confidently say that it *is* as good as 24-bit hi-rez. But I'm not alone - Robert Harley said this unit "sounded like hi-rez" in his '08 CES show report and now John Atkinson says the same thing (more or less) at Stereophile.
Deny it all you want - but at this point, I would listen to this player before you bad mouth it any further..........
Did you listen to the 808.2 with hi-res files? If so, this is an important data point. If not, your statement is without reference, and is therefore not useful as a standard of comparison. As I said, the AlphaDAC sounded so good with CDs in my initial sessions it was essentially beyond reference for me. I refer you to this thread, in which I was almost embarrassingly lauditory of the AlphaDACs performance on Redbook:
http://www.avguide.com/forums/benchmark-bryston-and-berkeley-labs-dac-pc-audio-systme
Please stop with the references to John Atkinson's review. It is obvious from reading that review that Atkinson was praising the sound quality of the 808.2, in part by comparing it to the sound of a completely different player playing back SACD files. If Atkinson's primary goal was actually to compare Redbook to hi-res, he obviously would have made the effort to compare his own hi-res files to his own Redbook masterings on the same player. To suggest that Atkinson, of all people, would engage in such sloppy, unempirical judgements borders on slander. His point was not about hi-res vs. Redbook, it was about the excellence of 808.2 CD performance - about which, I have no doubt. I've owned, by my count, 18 significant pieces of Meridian gear between 1994 and now, and I consider Bob Stuart a hall of fame audio engineer. The only thing I question - strongly - is the ability of the 808.2 to negate the artifacts of Redbook playback to such a degree that Redbook truly sounds like high-res digital. That is not "bad mouthing" the 808.2 - it's just recognizing what I hear, and what apparently every recording engineer I respect hears.
I think you should listen to the lecture Bob Stuart gives at this address:
http://www.aes.org/sections/uk/meetings/a0812.html
and parse what he says carefully. What he says, in essence, is that 16/44.1 has theoretically enough resolution to be transparent to the source, given correct noise shaping/dithering. What he also says, mostly by implication, is that actually implementing a transparent 16/44.1 system in a universe that has a fundamental sampling rate orders of magnitude greater runs very much afoul of the need to filter the input (anti-aliasing filter) and output (reconstruction filter) of the digital system.
In the context of digital playback systems, the problem is to create a sonically transparent reconstruction filter that passes the audio band, but completely attenuates all frequencies above 22.05 khz. Here's the thing: there is no way - no way - in the physical universe in which we live to build such a reconstruction filter that does not create large artifacts in the output signal. You can allow large amounts of ultrasonic noise to escape, or create substantial amounts of pre-ringing or post-ringing, or both, on transients, and/or create large group delay (phase shift), or restrict the response to less than the theoretical 20 khz audio band, or trade off all three. The primary approach in the past has been to use "linear phase" filters which preserve frequency response to 20khz, while creating very substantial pre-ringing and post-ringing on transients. What almost everybody in high end digital has discovered recently, mostly due to the availability of field programmable DSP chips and the resulting ease of programming your own digital filters, is that linear phase filters don't sound so good - presumably, because of the transient pre-ringing created by such filters. Such pre-ringing was supposed to be inaudible, because its frequency was above 20 khz. If that pre-ringing had truly been inaudible, you would now only be able to buy turntables at estate sales.
So Meridian, and Ayre, and Spectral, and Berkeley, and MSB, and Cambridge, and many others have recently discovered the virtues of "minimum phase" filters, in which the only ringing created is after the transient - at the expense of significant group delay (phase shift). Ringing created after the transient seems to have much less audible effect than pre-ringing. But the tradeoffs involved in minimum phase filters are far from clearly established as yet. For instance, the 808.2 uses a traditional, very steep filter that creates a lot of post-ringing - see the first graph in Atkinson's 808.2 review (and consider - is this the response you would like to see from an "ideal" format). Ayre, on the the other hand, thinks that kind of filter is not audibly as good as one that trades off some ultrasonic noise leakage and frequency extension for less ringing:
http://www.ayre.com/PDF/Ayre_MP_White_Paper.pdf
The point to all the above is this: 30 years after it was established, some of the Red Book format's most fundamental implementation issues are being addressed by a variety of high-end engineers - but only to the degree possible. You can fiddle with the tradeoffs to your heart's and ears' content, but you can't eliminate them, and you can't make the signal fidelity of an end-to-end Redbook implementation match the signal fidelity of, say, 24/176.4. Whether those signal fidelity differences translate to audible differences is a judgment that can only be made by ears. Your ears. If, of course, your ears have had the chance to make that judgment - using the same recordings, in different formats, played back over the same equipment, and judged as much as possible without agenda.
Atkinson was comparing 24-bit audio from the same session. And had a *hard time* telling the difference between CD and hi-rez. Deny it (or try to deny it) all you want.
I refer to JA because he's a known...and respected...audio engineer. And I trust his ears. If you say CD sounds "mechanical" compared to hi-rez, when no-one familiar with high-level CD says that (not RH, not Jon Valin, not anyone !!) then *you're* the one out of line, not me....
Yes, of course I've listened to hi-rez - that's why I say "like hi-rez". And the 808.2 (that I spent considerable time with, BTW) had a kinship of hi-rez that was startling. Besides JA, Robert Harley said this player "sounded like hi-rez" in his CES show report last year.
16-bits still code far more than a symphony orchestra produces, never mind a rock band. This, no matter how many "bits are available" at any one time. I've yet to see someone demonstrate how 16-bits does *not* code the audio waveform. The reason we needed more than 16 bits, according to Bob Katz, was due to signal processing done after recording. But we started recording at *20 bits* in the early 1990's, then dithered the consumer copy to ensure best sound. With Meridian, we are finally hearing the result of not only a great recording concept (digital) but the adjustments made to it nearly 20 years ago.
And no matter the techniques involved, playback cannot add resolution (or sound quality) to a recording - only reveal what's on it..........
I'm not denying that Atkinson said he had a hard time telling the difference between CD on one system, and hi-res on another. I'm also not denying that you find them to be similar. I am forcefully denying that these are valid tests of the audible differences between Redbook and hi-res; they are, in the lingo of scientists, "uncontrolled" experiments in which more than one variable is changed at a time. And I'm sure if you were to ask him, Atkinson would agree. You, of course, are another story ...
Your point about the sufficiency, or lack thereof, of 16 bit depth for audio coding is actually pretty much irrelevant to this discussion, for reasons I mentioned above: the 44.1 khz sampling rate of Redbook guarantees that you will see objectively large artifacts added to the audio signal by the necessary input and output filtering mechanisms, regardless of the bit depth. These artifacts are very significant distortions of the original audio signal, by any reasonable definition (look again at that impulse response graph for the 808.2). The whole raison d'etre of the 808.2, and similar products, is to try to rearrange these artifacts so that they are less audible - but even Bob Stuart doesn't say they are eliminated entirely.
For what it's worth, SACD has effectively less than 16 bit resolution in the upper octaves, although about 20 bits or so in the midrange. But the extremely high sampling rate of DSD means that filtration artifacts are essentially nil. This set of tradeoffs produces sound that some people swear by, and others, like Barry, swear at. I will say I've never hear SACD sound as good in my system as the hi-res PCM files I've got, but it's not a fair comparison since I'm not using the same DAC for both.
It is not a good idea to use the same player when comparing formats. It is well known that you must use a *dedicated player* for each format to hear its true potential - disagree as you will.....
The "artifacts" that arise in digital recording must not be that serious, if they remove them on playback. Many recording engineers inc. Peter McGrath have stated that 44.1 sounds as good as higher sampling rates. Bob Stuart has officially called for a 58kHz sampling rate - far from the 96, 176 or 352kHz systems being touted. So, even if there isn't much evidence that higher sampling rates sound better, they record it these levels anyway and have since ca. 1993. This (apparently) helped ease the phase shift that was occuring when recording right at 44.1 - so the theory goes.
We were only waiting for *playback* to catch up to the breakthroughs in recording and happily, that wait is over.
Finally, I agree with your take on SACD. Too much ultra-sonic energy loading the system AND low-rez in the highs. But this still didn't stop many folks from saying that it was "vastly better" than CD. It wasn't, it least when compared to high-level CD playback.......
So if you were to use an 808.2i, and play Redbook and hi-res PCM versions of the same recording through the digital input, that wouldn't be a valid comparison? Seems to me exactly the kind of rigor needed to actually buttress some of your assertions.
The artifacts of digital filtering can't be removed on playback. The artifacts created by the output filter are introduced at playback - see the graph in Atkinson's test again. You can, as the 808.2 does, use a steep output filter with a knee frequency below that of any pre-ringing likely to be in the recording itself, but the cost of doing so is a lot of transient post-ringing, and significant phase shift. That's one example of what I mean by "rearranging the artifacts".
I meant the *audible* artifacts of recording - that can be removed. This is one reason for Meridian's great sound.
And no, using the 808.2i would not be valid, since that player, apparently, was not maxed-out for 24-bit media. That's why Meridian calls it a "CD player". Each format should have its own dedicated player.
That's utterly ridiculous. Meridian themselves would completely disagree with you. Discussion over. Time to listen!
How do you know what Meridian would say ? If you think Atkinson was wrong, that's fine.
Good luck listening and do try to hear the Meridian or any other (new) player that uses the "apodising" filter...