Well yes, I made a mistake in this case. When no oversampling, all the job to filter DAC stepped output to be done by analog filter.
Last edited:
1. There is no correlation between equipment cost and its' sound quality – not necessarily, not always.
2. I need no DBT in order to know that I hear what I hear, whether it's hum noise, or the "musicality" of a sound system.
I am reminded of when the recording industry moved from obsolete tape recorders with vacuum tubes to the then modern solid state units.
The classic Ampex 350 tape machine was made both ways. Today you can still buy rebuilt ones. Of course the vacuum tube electronics are worth much more.
So was it a fad to go to solid state or is the fad to go back to the vacuum tube version?
Now the first pro reel to reel machines I listened to a lot were from Scully. After a few years of this I was surprised to be listening to a record and recognized the Scully sound on the recording. I check the label where they did list the equipment and it was recorded on the model that I expected. This surprised me at the time.
Now the change from vacuum tube to solid state certainly was a notable difference, but looking back I think the vacuum tube gear in good condition probably sounded better than the early solid state gear. Today there have been improvements in pretty much all of the other parts used to make sound equipment and those improvements can be applied to vacuum tube, transistor, analog IC's and even digital.
So while it is interesting to follow the arguments over digital, there is constant change. MP3 is a change but not an improvement.
Now the arguments sometimes get important issues confused. Fourier, Nyquist, Shannon, even Bekessy and Olson, etc all have contributed important bits. However all too often I see the wrong name dropped in regard to an issue.
I am aware that there is more information available on the web than anyone can really follow, but there are still quite a few pieces in those antique books that pertain and are not on the web. So the arguments seem to drift.
Currently I am working on a DSP based unit and get to play with sampling and filter types. Interestingly the literature in the area of my current interest points to approaches that sound absolutely horrible! But they look really good in plots!
As this is a commercial project I cannot add any hard data, and will not comment on most of these issues.
ES
The classic Ampex 350 tape machine was made both ways. Today you can still buy rebuilt ones. Of course the vacuum tube electronics are worth much more.
So was it a fad to go to solid state or is the fad to go back to the vacuum tube version?
Now the first pro reel to reel machines I listened to a lot were from Scully. After a few years of this I was surprised to be listening to a record and recognized the Scully sound on the recording. I check the label where they did list the equipment and it was recorded on the model that I expected. This surprised me at the time.
Now the change from vacuum tube to solid state certainly was a notable difference, but looking back I think the vacuum tube gear in good condition probably sounded better than the early solid state gear. Today there have been improvements in pretty much all of the other parts used to make sound equipment and those improvements can be applied to vacuum tube, transistor, analog IC's and even digital.
So while it is interesting to follow the arguments over digital, there is constant change. MP3 is a change but not an improvement.
Now the arguments sometimes get important issues confused. Fourier, Nyquist, Shannon, even Bekessy and Olson, etc all have contributed important bits. However all too often I see the wrong name dropped in regard to an issue.
I am aware that there is more information available on the web than anyone can really follow, but there are still quite a few pieces in those antique books that pertain and are not on the web. So the arguments seem to drift.
Currently I am working on a DSP based unit and get to play with sampling and filter types. Interestingly the literature in the area of my current interest points to approaches that sound absolutely horrible! But they look really good in plots!
As this is a commercial project I cannot add any hard data, and will not comment on most of these issues.
ES
When I had worked for some time in Supraphon recording studios during my university years, Studer A80 was the standard. Pretty good machine, good sounding. This experience had "frozen" my hifi enthusiasm those days 😀
Analog Tape Issues
All,
Tape Heads
Very few if any reproduce tape heads display any output in the bias frequency range, due to the wavelength vs head gap. The first null in response usually occurs near 20kHz for cheap cassette decks and perhaps 50kHz for the best Nakamichi heads. Professional decks are usually worse in this resepct than were the best Nakamichis (especially the Studer reproduce heads...blecch), with the first null occuring near 30-40kHz.
Moving upward in frequency, the raw response of the head at the bias frequency is usually 60dB or more down from it's 20kHz response. For all parctical purposes this energy is virtually eliminated by the HF deemphasis at the first eq. (this is not to be confused with the bias trap found in some three head monitoring decks)
Dolby HX
My experience helping Dolby develop HX showed that it GREATLY stabilized the recorded noise spectral distribution because: The bias point greatly affects the HF sensitivity, and HF program material dynamically modulates the bias. When correctly set up, HX stabilizes the actual total bias (sum program and HF bias signal). This stopped the dynamic noise modulation that ALL non-HX analog recorders suffer from.
HX also gives a huge improvement to the stability of the imaging due to stabilization of the size of the bias bubble at the record head. This eliminates the dynamic longitudinal movement of the recording threshold along the tape which caused time smear. When correctly set up, the difference is amazing.
I believe that the main problem anyone had with Dolby is the main failing with many analog systems (especially cheap consumer cassette systems) compared to digital: relatively low stability and high error %. I am not saying that digital sounds better. I am merely stating the fact that a digital system usually has lower errors and is more stable over time and other variables.🙂
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
1st on the internet
All,
Tape Heads
Very few if any reproduce tape heads display any output in the bias frequency range, due to the wavelength vs head gap. The first null in response usually occurs near 20kHz for cheap cassette decks and perhaps 50kHz for the best Nakamichi heads. Professional decks are usually worse in this resepct than were the best Nakamichis (especially the Studer reproduce heads...blecch), with the first null occuring near 30-40kHz.
Moving upward in frequency, the raw response of the head at the bias frequency is usually 60dB or more down from it's 20kHz response. For all parctical purposes this energy is virtually eliminated by the HF deemphasis at the first eq. (this is not to be confused with the bias trap found in some three head monitoring decks)
Dolby HX
My experience helping Dolby develop HX showed that it GREATLY stabilized the recorded noise spectral distribution because: The bias point greatly affects the HF sensitivity, and HF program material dynamically modulates the bias. When correctly set up, HX stabilizes the actual total bias (sum program and HF bias signal). This stopped the dynamic noise modulation that ALL non-HX analog recorders suffer from.
HX also gives a huge improvement to the stability of the imaging due to stabilization of the size of the bias bubble at the record head. This eliminates the dynamic longitudinal movement of the recording threshold along the tape which caused time smear. When correctly set up, the difference is amazing.
I believe that the main problem anyone had with Dolby is the main failing with many analog systems (especially cheap consumer cassette systems) compared to digital: relatively low stability and high error %. I am not saying that digital sounds better. I am merely stating the fact that a digital system usually has lower errors and is more stable over time and other variables.🙂
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
1st on the internet
The Studer A80 series was a pretty good analog tape recorder. Its mechanics were remarkable, the servos adequate, but their analog electronics, being good to marginal, getting worse with later models.
In making 30ips master recorders for Mobile Fidelity in 1979, and Wilson Audio in 1983, I removed the analog electronics and replaced it with my own discrete fet and bipolar class A design. I, originally, had hope for the early A80's, that they could be simply modified for optimum performance, but later models used more IC's, Class C output stages, and non optimum (for 30ips) bias frequency, and it was easier to start from scratch.
Dolby HR was never considered seriously by me for any serious or classical recording.
Commercially implemented, it was OK for rock, but we could hear problems with classical. My associate Brian Cheney and I attended a demo of it at Dolby Labs, and that is what we heard at the time, although Ray Dolby denied that there was any problem or change.
In making 30ips master recorders for Mobile Fidelity in 1979, and Wilson Audio in 1983, I removed the analog electronics and replaced it with my own discrete fet and bipolar class A design. I, originally, had hope for the early A80's, that they could be simply modified for optimum performance, but later models used more IC's, Class C output stages, and non optimum (for 30ips) bias frequency, and it was easier to start from scratch.
Dolby HR was never considered seriously by me for any serious or classical recording.
Commercially implemented, it was OK for rock, but we could hear problems with classical. My associate Brian Cheney and I attended a demo of it at Dolby Labs, and that is what we heard at the time, although Ray Dolby denied that there was any problem or change.
Hi,
Yet no filter known to mankind does that, plus there are several digital filter implementations that have widely reported as "sounding very good" that do not do that.
That is simply, obviously and patently untrue. Look at any AMR Digital product, they have (if you select the option) neither analog nor digital brickwall filters AND they do not have the 3.2dB Rolloff at 20KHz.
Incidentally they also offer selectable brickwall filters. Generally speaking, owners and other during auditions, given a simple button with which to cycle through the different options are remarkably consistent in selecting the "no brickwall" filter option as what they like best.
Yes, this MAY happen. However, the funny thing is that the fundamentals of ZOSH (zero order sample and hold) take to great degree care of this. I have repeatedly posted measurements of "Non-OS" DAC's without analogue brickwall filter which conform to common theories on the operation of ZOSH Systems by a remarkable failure to display "mirror frequencies, that may lay inside audio band".
So in the end we have a disturbing discrepancy, where the systems with brickwall filters are not universally perceived as better (actually the opposite apprears to be the norm) and much worse, where the system in measured terms surprisingly fails to violate basic laws of physics and electronics and hence fails to show the faults predicted by and warned against by those Technocrates who never let contrary evidence in the way of spinning much FUD (Fear Uncertainty Death).
In the end it is a question of preference. Not everyone likes the same things. So if some like the results produced by "Hi-Bit" DAC's omiting brickwall filters and others like the sound of "Lo-Bit" DAC's with brickwall filters, all I say to each their own, AS THEY LIKE IT.
Ciao T
Yet no filter known to mankind does that, plus there are several digital filter implementations that have widely reported as "sounding very good" that do not do that.
That is simply, obviously and patently untrue. Look at any AMR Digital product, they have (if you select the option) neither analog nor digital brickwall filters AND they do not have the 3.2dB Rolloff at 20KHz.
Incidentally they also offer selectable brickwall filters. Generally speaking, owners and other during auditions, given a simple button with which to cycle through the different options are remarkably consistent in selecting the "no brickwall" filter option as what they like best.
Yes, this MAY happen. However, the funny thing is that the fundamentals of ZOSH (zero order sample and hold) take to great degree care of this. I have repeatedly posted measurements of "Non-OS" DAC's without analogue brickwall filter which conform to common theories on the operation of ZOSH Systems by a remarkable failure to display "mirror frequencies, that may lay inside audio band".
So in the end we have a disturbing discrepancy, where the systems with brickwall filters are not universally perceived as better (actually the opposite apprears to be the norm) and much worse, where the system in measured terms surprisingly fails to violate basic laws of physics and electronics and hence fails to show the faults predicted by and warned against by those Technocrates who never let contrary evidence in the way of spinning much FUD (Fear Uncertainty Death).
In the end it is a question of preference. Not everyone likes the same things. So if some like the results produced by "Hi-Bit" DAC's omiting brickwall filters and others like the sound of "Lo-Bit" DAC's with brickwall filters, all I say to each their own, AS THEY LIKE IT.
Ciao T
I am not in a position to argue with PMA or ThorstenL about the 'fine points' of digital recording, but I would like to say that 'brickwall' filters have ALWAYS been known to be problematic, both objectively and subjectively.
It must be remembered that 44.1KHz is a LOUSY sampling frequency, only put on us in order that SONY could use its video recorder effectively. 50KHz was the original target, from the middle '60's. The hassle of making the anti-aliasing filters was significant, and the transient response was really lousy.
Later, phase compensation and even ideal phase filters were developed, but then they lead to TIME displacement, or Pre-Shoot. This is bad, because it puts some information before the event actually happened.
In any case, earlier research done before digital recording became popular showed that steep low pass filters were audible (in a bad way) but this knowledge was ignored when digital first came out.
It must be remembered that 44.1KHz is a LOUSY sampling frequency, only put on us in order that SONY could use its video recorder effectively. 50KHz was the original target, from the middle '60's. The hassle of making the anti-aliasing filters was significant, and the transient response was really lousy.
Later, phase compensation and even ideal phase filters were developed, but then they lead to TIME displacement, or Pre-Shoot. This is bad, because it puts some information before the event actually happened.
In any case, earlier research done before digital recording became popular showed that steep low pass filters were audible (in a bad way) but this knowledge was ignored when digital first came out.
Dolby redux
Hi John,
Maybe you were referring to Dolby SR, which was their multi-band companding scheme? The audibility of it's action, especially on low-level detail, was clear, although at the time it was released (1986) it was pretty good...not good enough for ultimate fidelity recordings by John Curl or Doug Sax maybe, but compared to most people's expectations for magnetic recording, good enough, and way better than Dolby A when properly set up (a BIG caveat!!).
In my previous post I was referring to the bias modulation scheme Dolby HX, and my work with it was purely as applied to cassettes. When correctly set up for a specific tape type, enabling it was like acheiving proper focus on a video projector; the sideband modulation which plagues fixed bias schemes would drop away and the sound quality was audibly improved, the imaging would lock in and tones became very pure.
Just my one case in point, and if you WERE referring to HX, I wish I had been there to adjust bias servo detector sidechain response and time constants, and have your ears as a judge! Ah well, we are discussing stone-age events, my attempts to retrofit HX to my CD recorder (now also stone-age technology) have been met with frustration...(ahem)
(...why can't I post a winking emoticon??)
Howie
Howard Hoyt
CE - WXYC-FM 89.3
Chapel Hill, NC
www.wxyc.org
1st on the internet
Hi John,
Maybe you were referring to Dolby SR, which was their multi-band companding scheme? The audibility of it's action, especially on low-level detail, was clear, although at the time it was released (1986) it was pretty good...not good enough for ultimate fidelity recordings by John Curl or Doug Sax maybe, but compared to most people's expectations for magnetic recording, good enough, and way better than Dolby A when properly set up (a BIG caveat!!).
In my previous post I was referring to the bias modulation scheme Dolby HX, and my work with it was purely as applied to cassettes. When correctly set up for a specific tape type, enabling it was like acheiving proper focus on a video projector; the sideband modulation which plagues fixed bias schemes would drop away and the sound quality was audibly improved, the imaging would lock in and tones became very pure.
Just my one case in point, and if you WERE referring to HX, I wish I had been there to adjust bias servo detector sidechain response and time constants, and have your ears as a judge! Ah well, we are discussing stone-age events, my attempts to retrofit HX to my CD recorder (now also stone-age technology) have been met with frustration...(ahem)
(...why can't I post a winking emoticon??)
Howie
Howard Hoyt
CE - WXYC-FM 89.3
Chapel Hill, NC
www.wxyc.org
1st on the internet
Last edited:
Conceptually HH, I agree with you that bias modulation could be a good thing with cassette recording. That was a 'breakthrough' if done correctly. Dolby, itself, while not a bad general concept, seemed to fall apart with its MANUFACTURE AND IMPLEMENTATION. I used to drive them crazy with criticisms. ;-)
Many changes in the analog days were made to overcome shortcomings that were basically forced due to the lack of options. When tube was the only choice, people were understandably unhappy being forced to deal with large, hot, fragile tubes that are prone to microphonics. Solid state allowed smaller, cooler (sometimes), sturdier modules that were more tolerant to vibration.
However, sometimes the sound suffered during the transition, even if the non-sound drawbacks had been eliminated. In modern times, if we can find vintage gear we usually have the option to choose between tube or solid state, and thus we have the luxury of selecting for best sound, perhaps ignoring the shortcomings more easily since we're not forced to use tube as the only option. Since good sound is neither guaranteed nor impossible with either technology, it becomes a matter of finding those works of art with excellent implementations.
I am often reminded that the earliest analog synthesizer designs have the richest sound. Unfortunately, this comes along with poor stability in regard to tuning. At the time, your only choice was a synth that wouldn't stay in tune. When revised synths came out with different chip sets that would stay in tune, musicians rejoiced. Were they concerned that the sound was noticeably thinner, not so rich any more? I'm sure they were a little concerned, but having an instrument that was in tune during a gig was far more important. These days, when seeking out a synthesizer sound, we have digital recreations and both modern and vintage analog variations. Now that we're not forced to deal with difficult-to-tune synths, some people in search of a particularly rich sound might choose the earliest analog synth designs even though it requires extra care in the tuning department.
In other words, the perspective of what options are available at a given point in time colors our reaction to particular limitations, tradeoffs, and shortcoming of any given technology.
MP3 is a solution for reducing storage size and therefore data transmission time. It was never intended to 'improve' the sound, so I would not consider it to be a 'failure' in that respect.
I suppose it is similar. Tubes sounded fine when done well, but there were particular disadvantages. Replacement technologies for tubes immediately removed the shortcomings of tubes, but did not immediately deliver the same audio quality until designers fully understood how to get the best sound. Digital audio sounded fine before MP3, but one of the disadvantages was the large amount of data. MP3 solves the data size shortcoming, but with a loss of information by definition. The difference between the two situations is that solid state at least has the potential of sounding as good as tube and perhaps even better, whereas it seems that MP3 is doomed to only sound worse. It certainly cannot possibly sound better than uncompressed digital.
Last edited:
Automobiles were hailed as an end to horse pollution. There actually are more horses in the US now then when the Model T was introduced.
The Ampex 350 did not change size when it went solid state. I never noticed any problems with microphonics in tape decks as they were run at a real line level input. The heat would sometimes be a problem in summer, but it was mostly due to summer. A few 12ax7's really did not get that hot. The power amplifier was the biggest piece of heat producing gear and in those days that would have been maybe 30 watts.
There is a nice old recording of a financial fellow doing an advertisement that had a background noise. It seems it was recorded in the summer in a small studio and being a proper fellow he was wearing a three piece suit. A fan was running to keep him from passing out! Air conditioning not yet being common.
The Ampex 350 did not change size when it went solid state. I never noticed any problems with microphonics in tape decks as they were run at a real line level input. The heat would sometimes be a problem in summer, but it was mostly due to summer. A few 12ax7's really did not get that hot. The power amplifier was the biggest piece of heat producing gear and in those days that would have been maybe 30 watts.
There is a nice old recording of a financial fellow doing an advertisement that had a background noise. It seems it was recorded in the summer in a small studio and being a proper fellow he was wearing a three piece suit. A fan was running to keep him from passing out! Air conditioning not yet being common.
Yes, I've rebuilt dolby sections on cassette decks as perfectly as I could and the results were always -notably better. However, I still ever used the dolby system on any deck, no matter how good it was.
As for the cable I mentioned earlier. Consider: Lenz.
take the copper tube, drop the neo magnet down it.
Lenz, right? A nice slow fall down the tube.
Take a double wall tube and fill this 'hollow cylinder' with the same fluid in the audio cables I speak of.
Drop the magnet down the tube. Zip! Thump! The magnet goes down the tube at the free fall speed.
So much for Lenz and Reactance.
Complex LCR, as engineers know it... is a function of frozen lattice structures with specific polarized systems of locking and energetic interchange at their quantum levels. The polarized spin characteristics of the locking and electron interchange of the specific frozen lattice IS the source of LCR and the given noise spec.
As I said, instead of dealing with the impedance functions and the specific distortions they produce, like everyone else does....In this case they were simply eliminated from the get-go.
Complex LCR remains but the reduction calculation involves a reduction calculation of a spherical (4pi) and Plank and Boltzman connections. The reduction in reactance is logarithmically lower than that of a solid.
Which is why, in a proper lab test... a given 1.2M finished RCA cable of ours measured a capacitive reactance that is similar or lower than that of 6" of 28g OCC bare copper.
I worked on a Scully R2R for a time. Really loved that beast. It was the mechanics I liked best as it was the editing machine. A real work of art, that transport.
This is probably the same reason you need gallons of the precious liquid metal to have a reasonable series resistance.
- Status
- Not open for further replies.