Ramifications

August 16, 2022
 by Paul McGowan

Ain’t it the truth? Everything we do has its consequences or, as the headline of today’s post suggests, ramifications.

When we transfer digital audio out of a computer there are ramifications: noise from the computer is transferred to the DAC’s clean power supply grounds with noisy ones. And so we insert devices to fix those problems and they too have their ramifications.

Or transferring analog masters to the vinyl cutter. There are no great choices for doing this without ramifications.

Or mixing DSD. One must choose from any number of less than perfect means for making the best of it.

Or investing more than the price of your loudspeakers to purchase a cable to connect them to your amp. Ramifications.

In the end, we make the very best choices we can under the circumstances. And that’s the key.

It all matters, but sometimes a little dash of common sense helps us choose the right path.

Subscribe to Paul's Posts

28 comments on “Ramifications”

  1. I thought it’s the role of manufacturers to identify these ramifications and engineer them out of audio equipment. Make your enemy your friend. The classic example is Class D power, been sold since the 1960s, sounded terrible, but had a key feature better than any other form of amplification, in that it is in theory 100% efficient. It has now been engineered to the point that it dominates the audio market.

    The same with usb. The irony is that usb used to a dirty word (pejoratively and electronically) in home audio and avoided at all costs. However, its selling point is that it can transfer any audio PCM or DSD data rate, unlike optical, S/PDIF or AES/EBU. So just implement it better. Innuos is about to launch a new range of three pure streamers and the chosen output for the top of the line model is usb.

    So I suppose it depends if you look at something as a compromise or an opportunity.

      1. It is factually and mathematically impossible to mix 1-bit audio.

        What has been proposed from time to time (and something I am personally enamored with) is the idea of 3 to 5 bit DSD (PDM). That has the same sonic benefits as does 1-bit but it adds the possibility of being mixed and volume adjusted.

        Is this fantasy land? No. Literally, every modern DAC and ADC chips use an SDM to produce multibit PDM. That is then low pass filtered into either 1-bit PDM and multibit PCM.

        At one point in Octave’s planning, chief engineer Bob Stadtherr along with our mathematical guru Richard Murison and I plotted out using (I believe) a TI ADC chip that made the 5 bit PDM signal available on one of its pins. The project got too complicated so we shelved it for now.

        There are two major problems with multibit PDM. First, it’s a serious pig. Imagine running DSD256 (11mHz) or even DSD128 (5.6mHz) x 5 x 32 channels for recording.

        Octave’s Pyramix DSD system is one of the biggest in the world running as many a 40 channels of DSD256. We’ve had to upgrade CPUs and HDs to keep from choking on the bandwidth. Now multiple that by a factor of 5!

        Lastly, there are no DAW’s that can handle this.

        1. I think I recently read, every „direct to DSD“ recording is in fact a 5 bit recording, subsequently converted to 1bit DSD or any other format inside the ADC, as there is no 1 bit DSD ADC. Is this gradually correct?

          1. Not to my knowledge. The Sonoma system (and Pyramix) are true 1-bit systems.

            If what you’re asking is about the A/D converters that make the 1-bit output then that varies per studio using them. For example, if memory serves correctly the Meitner ADCs are true 1-bit ADCs while the modern Hapi from Merging are multibit.

            1. I just start to learn, but this here is what Ted mentioned in a thread here:

              „At their hearts (with a very few exceptions) those converters neither record DSD or DXD, instead they are usually a small number of bits per sample at a high sample rate and then get converted to whatever.“

              Also Jamie Howard of plangent processes states this when you listen to this video from 6:15-10 minutes. By the way he somewhere else in the video also tells about his favorite DSD to DXD converter which again is a different one than yours etc. lots of interesting stuff, even if, as I assume, on a clearly lower level than Ted‘s.

              https://youtu.be/Owil9hHJbhA

              So finally what I try to find out on a theoretical basis for some time is, what is an as you call „direct to DSD“ recording at the end. And what I found out so far is, that it acually is something like a 5bit delta/sigma to DSD to DXD (or analog) to DSD path, while what most consumers initially thought of is a mic to final DSD media path. Everyone’s free to correct me, I’m just putting jigsaw parts together. Maybe your choice is the exception of this, so I don’t doubt it. It’s just that after the first view behind the curtain of a very general and simple statement, a can of partly contradicting worms opens up and this is what makes me wanting to know a little more to be slightly over fools level 😉

        2. Paul, I am supposed to be working on a project for my wife and I have just spent the last hour totally absorbed in trying to understand 3 to 5 bit PDM! This will have very serious ramifications!

          I googled 3 to 5 bit PDM and what came back was not very helpful.

          Are we talking about what dSC does which is 4.65 bit PCM at the same sampling rate as DSD? That doesn’t seem very novel.

          How are the extra bits used in 3 to 5 bit PDM? When do you out put an extra bit? Is the extra bit stacked on top of the first bit ( does this becomes PAM )? Or does the extra bit go next to the first bit and we end up with PWM?

          Is there a reference paper on multibit PDM that I can find and read?

            1. Mike, Thanks for the link. I long time ago I read a very good article in Stereophile ( I think ) about low bit ADC’s. I wish I could find it. If MoFi makes DSD256 copies of analog tapes they have to have some kind of ADC to get from analog to DSD256. I wonder if they would be willing to share what they use. I am not putting down the guys at MoFi, but they did not seem like the kind of crew that could build their own ADC.

              It seems to me that if Paul could get the direct output of a low bit ADC and then mix it before converting to pure one-bit DSD he would have what he needs. This also begs the question what do the use at Octave records to convert live analog music to DSD today?

              1. We use exactly what Richard was proposing (sort of). 🙂

                The Pyramix system uses AKM based A/D converters to go directly to DSD. What’s interesting about the AKMs is that their chain is basically Analog->SDM->multibit DSD->branch off to either 1-bit DSD or multibit PCM.

                Thus, when using the AKM based converters, the shortest and cleanest path id DSD. To go PCM the DSD signal must be low pass filtered and then rejiggered to as many bits as one wishes.

                DSD is by far the purer format inside these converter chips and this is basically how all modern A/D work. DSD first, then converted to PCM using a low pass filter.

                Which is why we keep everything in DSD and then use Richard’s Zephiir low pass filter to convert to 352/24 PCM for mixing. There’s literally zero loss.

                1. Paul, Thank you. Is it possible to extract the multi-bit DSD from the AKM chip in the Pyramix and then try to mix that and then convert to DSD? I realize that I have never seen a Pyramix and that I am just taking a crazy shoot in the dark asking if it is possible.

                  1. No, unfortunately not. However, there is a TI chip that allows this and we’re working on a side project to build A/D converters that output the multibit DSD bits.

                    The big problem is what to do with them. They are easy enough (in theory) to record and edit (cut, splice, move), but that’s about it.

                    There are no editing programs available. There are no drivers that recognize this format. There are no DAWs that can work with it.

                    One sneaky plan we had was to use an existing open source DAW (Ardour) and piggyback on it this format. It seems possible but it would require a lot of programming chops we do not possess.

    1. I also understood it as FR. Not possible and therefore no gear available. Not sure if using the Sonoma is closer to pure DSD mastering, but it seems to be limited to DSD64.

      The interesting part IMO is to understand, that DSD seems to be an archival (and therefore also recording) format, somehow „(mis)used“ as music media production format in at least a sufficiently successful way, that it’s better than pure PCM (if we limit things to digital for now).

      What we get as readily to consume DSD media is a direct to DSD recording, which usually went either through a forth and back media disruption (analog mixing/mastering/editing inbetween) or forth and back conversion stages (DXD/PCM mixing/mastering/editing inbetween) with different opinions on which is superior.

      Those are as I understood the „less than perfect means“, Paul mentioned.

      As in the beginning of all digital technology innovations, the whole process is often again declared as transparent and lossless, which it factually can’t be, but if it’s the best we can currently do, it should be enough for now.

      Seeing the (at least at first sight) much higher grade of less than perfect means in analog media production, it’s always a wonder why it’s still comparable…and maybe a hint, that were not at the end in digital progress yet, too.

  2. The thing is, we get better the more we know what plays a role. This then has ramifications meaning the need of add. measures, devices, cost…ending in better sound.

    In the analog example you mentioned, this could be the awareness that a one step plating process improves things, or the use of quieter vinyl and higher RPM per minute.

    In the analog realm I never perceived a kind of hybris, that it was perfect or will ever be. It was known as flawed from the beginning and manufacturers in this awareness and public communication constantly worked on what there was and still is to improve.

    Funnily I tend to perceive this differently in the digital realm, where from the beginning of still understanding little about it, until today where we understand a lot more, the claim always was and seemingly continues to be from CD over hires to DSD…the technology and its connected processes are „perfect, accurate, lossless“…until further imperfections were and are discovered, measures for ramifications addressed and necessary equipment or cirquits developed, ending for the moment in a lot of parts and boxes in and around our streamers, disc drives and DAC‘s 😉

    It’s not that there manufacturers didn’t also constantly work on improvements…but the public communication in many cases was not corresponding. The communication usually was as if there’d be no need to further work on things since 40 years.

    As good as this technology has become, it shouldn’t really need this kind of communication anymore.

    The other examples of cabling or power supplies etc. playing a major role…yes, many didn’t believe it or still don’t and it’s not always the best recommendation to invest in the relation of the parts‘ size and looks but in the relation of their contribution to great sound, which can be opposing.

    1. You may not be able to make it better, but you can make it cheaper and prettier.

      Excluding what is available to the super-rich, sound quality may be as good as it will ever be, but make it cheaper, easier and more practical to use and in up-to-date enclosures, and it becomes more accessible and the communication issue disappears. Well designed products should sell themselves.

      The Marconiphone dominated the USA audio market for decades, in the UK we had the Radiogram, and they were beautifully made pieces of furniture often with beautiful mahogany or walnut veneers. People bought them in their millions because they looked and sounded good without explanation. In those days most audiophiles built their own systems from plans and schematics published in magazines and high street audio stores sold the required components.

      I’m not sure much has changed, other than that the consumer gets increasingly better value. Even with vinyl replay, now that it is more popular, there are exceptionally good phono stages at very low cost that were not available 10 years ago (the Vertere Phono for one example).

      Of course most companies just try and improve on the same old technology, iron out perceived flaws, make minor improvements, more power etc., they are much the same just more expensive. Not saying that is a bad thing, Rega have been tweaking the same technology since the 1970s and they are more successful than ever. That said, the Rega Naiad was fundamentally innovative and resulted in the P8 and P10, both hugely successful (the P8 moreso).

      The reason being that probably only 1% or 2% of products are actually really innovative, at best, most are just rinse and repeat.

  3. Prompted by what jazznut wrote at 3.28am perhaps one of the biggest mistakes made when CD’s were introduced was using the tagline ‘perfect sound forever’.

    I just stumbled across the following article. It’s a long read, there are further links within the piece and I’ve not read them all, but from what I have read it’s fascinating, from a historical perspective and comparing what was said then to what we know now.

    https://dutchaudioclassics.nl/Philips-Compact-Disc-Digital-Audio-presentation-27-May-1981-New-York/

    I like the picture of the lady holding the LP and the CD at the bottom of the first page. The expression on her face, what is it exactly that she is trying to tell us?

    1. Talking of pictures, what’s the picture accompanying today’s post trying to tell us, that your system will sound better after downing a few beers? Well of course it will, you’ll no longer notice the imperfections and even if you do, you won’t care. 🙂

  4. Not hanging out and chatting with the right audiophiles has its ramifications. 🙂

    I like being here cause I find many on here are technically minded and genuinely love the hobby that we are in.

    5 bit DSD recording? Good god I can’t imagine what that would be like, but “ life like “ comes to mind. The future has something to build towards. 🙂

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301
1-800-PSAUDIO

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram