I have to admit that I do not like all those “improving gismos” between my digital sources and DA processor and particular all that crap that Hi-End industry offers. Therefore it is kind of atypical for me to pay attention to those toys, not to mention the toys that come from pro-mastering world, the world to which I arrogantly do not have a lot of respect generally.
Over the years I was in a very non-aggressive and very mild search for a new version of processor to power my delay channels , I have seen nothing stimulating and then yesterday the “K-Stereo” unit was recommended to me. It was recommended because the different reasons but when I begin to read about the “Ambience Recovery Processor” it picked my interest. I might rend this thing for a weekend if my local pro shop will have it available to see if it “works”. Frankly speaking I have little expectation …. but it might do very well for delay channels processor…or who knows

The unit is here:

http://www.digido.com/misc-content/k-stereo-2.html

The description of the processor sounds slightly idiotic and it is MUCH better in the PDF file of owner manual.

http://www.goodsoundclub.com/PDF/K-Stereomanual.pdf

It might be very interesting to hear the actual sound of this thing. Did anyone have heard it?

peter foster wrote:

The feature that I find the most useful is being able to select the soundstage characteristics.  There is a noticeable difference between the various settings that assist with stereo imaging of players / instruments…. As I understand it, the ambience recovery mechanism takes an attentuated digital copy of each channel and copies it to the other channel with a slight delay….

Peter, this is very interesting.

If you read my articles “Preamplifiers: keys to mystery. (Lamm L1, L2)” at:

http://www.goodsoundclub.com/Forums/ShowPost.aspx?postID=257#257

…then you might pay attention to the next post: “Are the secrets of Lamm X-factor resolved?”

In there I was hypostasized that some of the key to the Lamm’s amassing X-factor might be the spatial expander algorithms that Lamm uses, as no other know to me “straight preams” has no X-factor effect. The spatial expander patterns might be different, perhaps Lamm did found one that works and implementable at analog domain, perhaps not the best implementation but it was just a first (perhaps even accidental for Lamm himself) step.

Now it looks like there is a digital device that claims own space in the ceremony of spatial re-rendering. I do not know how it is done and have no expertise to evaluate it on theoretical bases the concept that K-Stereo proposes but the fact that someone is looking at those things I found superbly positive. May I ask you:

1) Does this device allow modifying the death, amplitude and “polarity” of the spatial extraction and injections?

2) Do you detect any changes to sound (sounds) besides the injections of spatial information?

3) Does the positioning of the images are changing or the images juts got slightly different spatial presentation at the very same position?

4) Do you detect any change in sound when the K-Stereo unit in a by-pass mode?

5) Does the K-Stereo unit has the I2C-bus input or it uses AES only with further AES to I2C conversion?

6) If you run the K-Stereo “spatial extender” multiple time on the same digital file then what happen?

Romy the Cat wrote:

Do you detect any change in sound when the K-Stereo unit in a by-pass mode?

drdna wrote:

I used this method in my car and my house, but the destruction of the sound with DSP let me to abandon all these efforts and to focus on simply reproducing the original Sound.

Yep, the very same observations are here: as soon DSP engine touches digital stream/file then sound immediately goes to toilet. The numerous, and in many case very complicated, DSP processings that take place inside our AD/DA processors I would conveniently leave un-mention as it would screw the elegant picture of my DSP’s hate… :-)

What however is interesting in all of it is that there is rule: analog can’t delay, digital can’t filter. If what Peter describes is correct in what the K-Sterio possessor does while it is working then it should activates some kind extraction and delaying of injected signal. Delaying is just a temporary storing in a local memory, something that has no impact to sound at all. Delaying is not a problem for digital. Sure, then injecting massaged signals from another channels they will be hugely  attenuate at digital domain – something that digital can’t do but still it will be at very low level and it might not be not bad.

Of cause they are all juts groundless speculations and should be interesting to learn about the transparency of this thing… (Look for "Reviewing preamps by imbeciles.” In the list of the related links). Also, it is interesting to learn what exactly from the K-Sterio processor is being licensed by Daniel Weiss…

peter foster wrote:

I suggest the following, if you are interested:  Select three recordings of varying quality; copy ten minutes of each of the three recordings to Redbook CD; send to me by air mail the CD that contains three ten minute tracks; include instructions to me for each track of the K-Stereo settings that you wish to have used; I will process the tracks using my K-Stereo according to your instructions; and I will send the results back to you on CDs by air mail.  That would make the 3 original tracks plus 3 x 3 or 9 processed tracks making 12 tracks in total.  Let me know by PM whether you want to do this and I will then forward to you my mailing address.

peter foster wrote:

K-LEVEL can be set from -90.0 to +6.0 and unless BYPASS is used then all setting (both -ve and +ve) will add some level of ambience recovery with -90.0 being completely imperceptible.  The K-LEVEL +0.0 is that level where Bob Katz considered a typical reference point for his use in the recording studio when mastering.

peter foster wrote:

As the K-LEVEL is increased there is also an apparent related small increase in volume.  Perhaps one way to describe the increase in volume is to use the example of a CONTRAST feature of a color television.

peter foster wrote:

3) The WIDE feature will separate the instruments to make them sound further apart in width.  The DEEP feature will separate the instruments to make them sound further apart in depth.

peter foster wrote:

4) My K-Stereo unit is always in the playback chain.  BYPASS mode passes the digital data absolutely transparently in my opinion.  On a side note, in my opinion the digital signal volume control on the K-Stereo works blamelessly.

peter foster wrote:

6) Don't know because I use the K-Stereo unit in playback always on the original digital source recording.

Dear Romy,

A.  Recordings of different quality are worthwhile.  Let's take three examples:

1) A stereo recording that you feel is as good as it gets.
2) A mono recording of a beautiful piece of music that is well played.
3) A stereo recording of a beautiful piece of music that is well played but where the soundstage is imperfect.

B.  I do not see Redbook conversion issues at my end because the conversion would be Redbook CD digital data input; K-Stereo digital processing; K-Stereo digital data output; hard disk recorder storage; Redbook CD production.  No dac or adc involved, just purely working in the digital domain.

C.  I would use an Alesis MasterLink hard disk recorder as the source for the digital data.  It supports sample rates of 44.1kHz, 48kHx, 88.2kHz and 96kHz.  It supports word lengths of 16-bit, 20-bit and 24-bit.  BUT my preference would be to receive recordings that you are intimately familiar with and in Redbook format because Redbook is the lowest common denominator for everyone.  That way, anyone can download the source tracks and processed tracks and listen to them and compare the results to form their own views, on their own playback systems, presuming of course that you would post them on goodsoundclub.com.  Note that I am not really in the mastering world but I have the necessary equipment for private study and hobby recording and hobby mastering using some equipment which is as good as the professionals use so the results should be appropriate.  The project should enable you to form your own judgement using material that you know intimately and I think this is very important.  I already know the results for my own music collection, and it is obvious that I think it is beneficial because the K-Stereo machine is a permanent part of my playback system.  However, we have perhaps some different tastes in music so it might be more helpful for my music collection than yours.  Also, we have different playback systems and listening rooms.

D.  Amount of ambience recovery can be selected and also the type of ambience recovery to the extent that WIDE or SMALL is one , DEEP or not is two, AMBIENCE HP/LP FILTERS is three, AMBIENCE BELL is four, INPUT LEVELS (mid and side) is four, DITHER is five, etc.

E.  The amplitude of spatial modification is, in my opinion, correct and detectable whether at -10dB and -50dB.

F.  One of the interesting tests would be whether the results of using the digital volume feature carry some of the penalties that many believe are produced by digital volume controls.  Another interesting test would be how compound processing of the digital data ten times using the same settings affects the results.  I will think about the K-Stereo settings and PM you to see whether we can come up with a set that are informative.  In the meantime, select your three tracks.

This really is old technology, since Bob Katz had been using the methods that he built into the K-Stereo machine perhaps ten years prior to building the machine and the machine was released in 2001.  So that makes the technology perhaps 15 or 16 years old.

Regards, Peter Foster.

peter foster wrote:

3) A stereo recording of a beautiful piece of music that is well played but where the soundstage is imperfect.

The subject of my “hate of Soundstage” is well known and documented for years, including this site. I was vocal and persistent Soundstage (as it sold to audio people) abolishener years ago at AA, from where the reviews-writing-idiots  took my context, wrapped it  into their typical epistolary-nothingness and then were selling my anti-Soundstage propaganda to the Audio-Morons as “the new wave” of audio marketing.  I well knew it and I played my little game with the subject: I did not tell “everything” and I intentionally imbedded some “mistakes” in some of my keys pointers. All of the “industry new views” were published with my original “mistakes”. Do you remember the wave audio revivers in 2002 that “suddenly” degraded the value of soundstage? Be advised that they took place after very large and extended debate by me the subject at AA and then, in whole topic were deleted from AA … 2 moths after it was deleted from archives and the fragments from there made to publications, right along with imbedded “mistakes”.  (I mention it not because my ego, at least not because the only my ego but rather because to have a pleasure to point out one more time the primitivism and  impotency of the official audio)

Yes, the Soundstage is not the property of live music but explicitly the property of sound reproduction efforts but there is a LOT more to it. Although Soundstage is a nice “trick” (deep, wide the geometry) but there is more important moment then de facto Soundstage. There is also spatial discrimination between the implicit center of transient pitch and the way how the space outside of the pitch’s source reacts to the pitch.  This moment do not directly relates to Soundstage but it is modifies the Soundstage values. A Soundstage might be big but “dead”, with the “near spaces” do not react properly to own transient irritations. A Soundstage might be flat and narrow but if the “near space” is set properly then the flat would not bother a listener, unless he is an Audio Moron who instead of listening runs across a room with measuring tape, assessing the dimensions of the bogus Soundstage. The proper reaction of the “near space” affect the inner-dynamics of imaging and it is way more important factors then the geometry of Soundstage. BTW, Soundstage does not exist in real world but the “near spaces” reaction does. Did you ever have experience to monitor how good quality playback sound outside during a nice sunny summer vs. a few minutes before rainstorm. If you do then you will know what I mean what I say “near space reactance”…. If you want to stady this subje in audio then get yours an old Lamm L1/L2 preamp – you will learn a lot.

So, what is the more interesting in the results of  K-Stereo unit  would be to learn if it deals with spatial reconstruction via making Soundstage “bigger” or if interfere with imaging at the level of the “near space’s  inner-dynamics”.

Antonio J. wrote:

then it will be worth exploring. I have always missed in my playback the way the air in a concert hall gets filled of sound, as if notes and specially their armonics were flowing upwards. The TU-X1 does that trick quite nicely, but it's not the same really.

Romy the Cat wrote:

BTW, pay attention that TU-X1 while doing “space” does not change the size of imaging. Pay also attention how sensitively TU-X1 portrays the acoustic environment of each performing hall… BTW, if you use TU-X1 then try to get as best LF section as you can…. If the K-stereo’s processing algorithm can do the same with digital stream then it will be wonderful. Though my ultimate 9and very personal) goal would be do not do it “before DAC” but to find a way to “smartly” randomize phases and inject ambiance at… speaker level. I did some experiments in past with artificial faking RT60 and cross-injecting of my right and left the channels. Something was very promising, something was not good… I still am planning to return to this subject in future…. Rgs, Romy the Cat

Romy the Cat wrote:

Does K-Stereo unit deal with Soundstage or with REAL spatial reconstruction?  What is the more interesting in the results of  K-Stereo unit  would be to learn if it deals with spatial reconstruction via making Soundstage “bigger” or if interfere with imaging at the level of the “near space’s  inner-dynamics”

drdna wrote:

There is no mystery in how the machine works, really.  A very crude analogy is how they used to put a dark blue filter on the camera to simulate night scenes in movies.  It was cheaper and more practical to shoot during the day hours.  However, if you looked closely, you would see everything casting shadows despite the supposed night time.  But just like early moving pictures fooled the viewers, I am sure this machine will create a very interesting soundstage effect that eventually the listener may become able to discern.

I know what you are saying and it is might be true. Well, it might be true in terms of artificial soundstage reproduction as soundstage is the artificial subject itself. Still, if the K-Audio device works in trim of spatial improvement then it might be interesting as we have a lot of room to go in spatial audio. The spatial information is recorded, preserved and reproduced by “other” means. The mainstream audio does not deal with spatial complexity of recordings. Too complex, too expensive, too difficult to sell, too difficult to monitor…. etc, etc…

I certainly do not automatically presume that K-Stereo devise might operate at the level of “imaging without soundstage”. Looking at what they say and promise it looks that their awareness still is confided in soundstage prison. However, they say what they are able to sell and in the language that might be understood by their consumer.  They name a frustration around which the audio person’s mind is rotating – namely “a wide and deep soundstage”  - and they offer a devise that produces an improvement in that area.

Does the K-Stereo screw something else? Does the K-Stereo deliberately or not affects the spatial presentation?  I do not know. However, I very much welcome any ideas that toy in this direction. I like idea of microphone that “process” the near-sound and far-sound with a certain level of “smartness” (similar to what out hearing does). I like electronics that deal with space like a hot scooper deal with ice-cream. I like loudspeakers that destroy division between the “beginning of soundfield” and “end of the soundfield”… So, to me, the more devices that expand our “dangersness” in spatial reproduction the merrier. You might be correct and many of them will be usable juts conditionally or be just turn out to be plainly bad but it is how the experience is building itself up – with exposure to many different opportunities, trying to educate ourselves with what “works”, why it “works” and what might be the algorithms for success…

Dear Romy,

The engineering basis for the K-LEVEL feature of the K-Stereo machine is described in Bob Katz's US Patent Number 7,076,071 "Process for enhancing the existing ambience, imaging, depth, clarity and spaciousness of sound recordings".  Here is a link to that patent.

http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&u=%2Fnetahtml%2FPTO%2Fsearch-adv.htm&r=86&f=G&l=50&d=PTXT&s1=katz.INNM.&p=2&OS=in/katz&RS=IN/katz

In answer to an earlier question, perhaps this, amongst other technologies, is what Weiss have licensed from Bob and why.  You will also notice from the Digital Domain web site that the K-Stereo machine looks like it has been built using a Weiss chassis and displays.

Regards, Peter Foster.

Antonio J. wrote:

Anyway, going back to the subject of the thread, why would you implement the ambiance retrieval at speaker level and not at the "pre DAC" one? After all we'd like to have from CD's something closer to what TU-X1 does.

2) I feel that it is job of acoustic system to handle acoustic of reproduction enrolment. I wonder how the K-Stereo’s effects would be “rendererable” by acoustic system’s efforts and in case the delays were done by modulators and UHF analog lines. I really would like to see the spatial potency as a natural capacity of speakers instead of encoded signals. The signals might/should be “clean”. The K-Stereo unit might be a perfect device to prove the concept and to model the concept but I see the concept (if it worthy) as a final solution implemented at speaker level; implemented completely integrated with a given listening space… or even with the speakers that are initially designed with spatial reconstruction awareness.

3) Since I am a “speaker person” and it is the region where I did something and have my relative understanding, some proven success and great personal interest, then constantly in my mind am updating my view of THE Ultimate Loudspeaker Solution. Without going into details what it might be I would just mention that it should have the RT-60 mimicking delay channels among other thighs… So, it is quite possible that some kind of spatial extenders might be a part of the Ultimate Loudspeaker Solution. That is a subject that always interests me….

Peter, thank you very much for posting the link. I read it, I love it and I will read it another few times. I need to leave with this idea for a while. As you understand my objectives are not like K-Stereo proposes: “to make big is audio business” and not “how to get big sound” (I have it already) but rather to figure out “what else is possible”….

Dear Romy,

I do not know for sure.

The best that I can tell from various internet articles describing WaveLab is that the Bob Katz features incorporated into WaveLab are for calibrating, monitoring and preventing excessive limiting.  These are important but different issues to ambience recovery.  See the following excerpts:

Finally, the level/pan meters now have K-Metering options, as defined by renowned mastering engineer Bob Katz, which you use in connection with calibrated monitor levels to ensure repeatable high-quality mastering results without compromising dynamics or squashing transients.

---

It now also supports the K-System metering standard developed by mastering legend Bob Katz.

---

A welcome improvement to the VU meters is the addition of K-System's loudness metering (developed by Bob Katz) as a means of better calibrating and monitoring excessive limiting.

Authoratative answer would come from Bob Katz.  Send him an email with your questions.  I have always found him to be courteous, helpful and prompt answering my questions.  His contact page is http://www.digido.com/contact/index.php

The only company I know that has integrated K-Stereo ambience recovery features is WEISS.  See

http://www.weiss.ch/gambit.html

DNA1 Denoiser/declicker/ambience regenerator
POW-R Wordlength reduction (redithering)
SFC2 Sampling Frequency Converter

http://www.weiss.ch/dna1/dna1.html describes the K-Stereo integration into the DNA1 machine.

Interestingly, the Digital Domain K-Stereo machine has itself integrated features of the WEISS POW-R and SFC2 machines.  WEISS machines have a top shelf reputation in pro audio industry.

Regards, Peter Foster.

Yes, Peter,

I know that in the Steinberg WaveLab use only Bob Katz’ metering system, still I thought that if they are cooperate then Katz’ Ambience Recovery might be available as a standalone plug-in. It is kind of suspiring that Mr. Katz was able to patent it as the notion of “kinky” channel-crossfeeding in order to make stereo to simulate the true binaural recordings are very old and used by headphone people for year and years. I always very welcome those experiments as I feel that stereo was very faulty concepts to begin with, not to mention how it is being used by the industry idiots. As I can see, it order Stereo to work properly it should be substantial amount of acoustic phase randominisation that takes place between a loudspeaker and listener. In case or “not helped air” I would estimate that it is necessary 25-30 feet minimum in order all Stereo mathematical algorithms got polished by phase randominisation and stop to me artificial. In case of our nearfiled listening (it is the way how I prefer because of many reasons) stereo is effective but incorrect was to reproduce live sound. It is very possible that some kind of K-Stereo processing might help a little…

Still, I do not think that Mr. Katz invented the channel-crossfeeding he rather juts parented the algorithms of digital extraction and injection. How I see it, being implemented it is superbly simple operation for digital and the whole processor, it is operates at I2C-bus level should be just one small DSP chip. It would be a perfect plug-in. The complete K-Stereo use would have power supplies, the data accusations stages, the stages that convert AES/EBU to I2C and then back to AES/EBU… I mean a lot of things that might not need to be done. The K-Stereo unit has also some custom dithers. My experiments with them suggest that good dithers are very-very useful tool at 16bit but since we go to 20-24 bit then all those dithers might become unnecessary.  Not to mention that most of my recordings as FM broadcasts where the HF FM noise does acts as very effective dither.

So, let to talk about the hypothetical “perfect” processor for a person who do digital only (I would LOVE to have this toy it at preamps or at speaker level - who would be up to the challenge?!!!). The K-Stereo unit might be very fine but look at the DNA1 unit from Weiss. Daniel Weiss’s machine should be even more interning (twice price) as it has many source-denoiser functionalizes.  Robert Witrak, the guy who run HDTT use it and he is absolutely ecstatic about the DNA1 performance. Though I have my disagreement with some aspects of Robert’s sound but his tape-hiss denoising is very high quality. Would the DNA1 be the ultimate toy to try? Well, not really. Application of custom dithers and other DSP processors crates own noise and here is where the Lavry noise shaping techniques might be very useful. You might read about Lavry’s Acoustic Bit Correction® at:

http://www.lavryengineering.com/white_papers/dither.pdf

The application of his noise shaping techniques along with custom dithers might be a good tool as well. 

OK, now the ironic part that I learned during my experiments with sound of some of those and the similar thighs. That is very important moment from my point of view.  When you have a better amplifies, phonostage or a cable for instance then you learn about their capacity what you plug them in. With the DSP devises and particularly with the devised of spatial reconstruction you learn about their benefits when you bypass them. This is a paradox that I have been thinking for years, I have no explanation to it but it is how my mind works.  The spatial reconstruction effect when it kicks-in impacts me as very positive but then what I shot down the effect then I begin to observe what did I sacrificed in order to have the spatial reconstruction. Still, I never heard the way how it is done at K-Stereo level and it might works very different then what I have experienced before…

I can wait what I put my 12 active channels playback back to the fully operation mode and will be able to try the K-Stereo opportunities…

Dear Romy,

K-Stereo receives a digital data feed, performs a bit-rate conversion in preparation for internal processing, then as a last step optionally performs dither for the reduced word length digital output.  The machine contains an internal (24 or 32 bit?) processor plus the WEISS sample frequency converter (see SFC2) plus the WEISS word length reduction dithering algorithms (see POW-R).  Such an integrated set of hardware and software functions is better suited to a dedicated machine rather than as a plug-in for software mastering programs running on a workstation.  It appears to me that the K-Stereo is a combination of all three WEISS machines but without the denoiser/declicker.  I have found the results to be flexible, optional, non-destructive and add refinement to music reproduction on my system when using digital sources.  In other words, I am satisfied with the results for my music collection because they are a plus.

You make some very interesting comments about stereo and distance.  I believe that the very best music from my system is heard when I crank the volume up, open the wide doors to my study and take a walk outside about 20 metres from my study.

I noticed that some people very much prefer the sound of TU-X1 as a benchmark.  There is something about that machine which is very pleasant to listen to and especially with live performance FM broadcasts.  However, when I have the same recording on CD as is being broadcast to my TU-X1 then I would much prefer the CD on my system every time.  The TU-X1 sounds to me as though there is a dolby processing happening even when the dolby feature is off.

I am familiar with the Lavry dither algorithms on my Lavry Blue 4496, such as Acoustic Bit Correction, ABC-1 and ABC-2.  It is a 2 channel microphone preamplifier, clock and 2 channel analog to digital converter.  There are a number of dither choices on the Lavry and a number on the K-Stereo.  Their dither algorithms are different and good results can be achieved using either machine.   The best results, for me, are achieved by recording on the Lavry Blue in 24/88.2 or 24/96 on the Lavry Blue, then performing no dither within the Lavry Blue and passing the bits to the K-Stereo and only performing dither as the last step prior to conversion to analog output.  Note that the Lavry Blue (and Gold) also have a feature for massaging clipping that adds its own puss, as you might say.

The link that that I sent in my previous post contains original and processed samples of the WEISS DNA1 machine and there are some highlighting results of its ambience recovery which should be the same results that are achievable from the K-Stereo machine which I think are assembled in the WEISS factory anyway.

http://www.weiss.ch/dna1/dna1.html

They are at the bottom of that web page.  They might help you to form an opinion of destruction or not of the music of ambience recovery.

Regarding destruction with DSP, this is a typical risk/benefit analysis.  Is the result a net gain or a net result, that is the question.  For example, in converting my vinyl collection to digital I soon stopped performing noise reduction because too much music was lost.  So I live with the vinyl noise on my vinyl to digital conversions.  But then I have not used the WEISS noise reduction machine so maybe that would have been a solution.  Also, your Pacific Microsonics machine and your Lavry Gold machines are DSP machines too, so the results can be very good using a quality implementation in a correct way.