DP could be the first to DSD multitrack

[Timeline]

DSD man. I bought a Tascam SD-3000 and have done the comparison even to 192k-24 and DSD is better to my ears. I think its time for MOTU to jump in and support DSD 256 and below standards so we can start recording multitrack and mastering with DSD. Just my .02c

Regards, GB

[supersonic]

Definitely a good idea though the hardware would take time to catch up I think.

[Robert Randolph]

Doesn't pyramix do DSD multitrack?

[Timeline]

Doesn't pyramix do DSD multitrack?

Not sure. Seems the IOs don't care or do they? I noticed on mixes i did with DA-3000 that the stereo files were 4 times as big as 96k 24 bit but it sure sounds great! Maybe the computer throughput would take some time to catch up as mentioned especially where allot of tracks would be involved.

Would like to just get 16 tracks working first and try some projects. If we could have mixed formats, 96k to192, maybe we could get there sooner. RME allows interem formats and it sounds great too.

[Timeline]

Looks like MOTU is behind the curve actually. I found this:
http://www.supremepiano.com/product/serenadedaw.html

You can have a complete playback multichannel & mix mastering system for about $1200. (Mac&PC)

If you already own an Tascam SD-3000 that has DSD64 &128 and costs street about $800
your good. Total 2K. This company is also making a record IO but have not announced the price. It will do DXD,, Extreme DSD, up to Highest formats and 32bit 384. Can't imagine something like this hitting the floors of AES or Namm by MOTU for a while but building a DSD IO might push them into considering it. Maybe the Sony format is too expensive royalty wise, do not know.

I'll keep hoping though.

[Shooshie]

For anyone needing some background on this stuff, as I did, DSD is Direct Stream Digital. Commercially, it is embodied in the Super Audio CD (SACD), a product originally from Sony and Philips, an "improvement" on the old Compact Disc technology of the 1980s, yet not without its own current set of limitations. (See the quoted paragraph down a little ways below here on the difficulty of using DAWs to mix in this format.)

DSD is:

a sequence of single-bit values at a sampling rate of 2.8224 MHz (64 times the CD Audio sampling rate of 44.1 kHz, but only at 1⁄32768 of its 16-bit resolution). Noise shaping occurs by use of the 64-times oversampled signal to reduce noise/distortion caused by the inaccuracy of quantization of the audio signal to a single bit. Therefore it is a topic of discussion whether it is possible to eliminate distortion in one-bit Sigma-Delta conversion.

Before you delve deeply into DSD, you may want to refresh your memory on PCM (Pulse Code Modulation), which is stuff you've known for years, but sometimes forget that you know.

You may also want to read about one of our all-time heroes (at least, one of MY all-time heroes), Claude Shannon, who is responsible for so much of what we do that we are practically his family. His work is directly responsible for PCM, and ultimately DSD. Really, any kind of signal processing is a derivative of Claude Shannon's work, and I recommend a book called Grammatical Man, by Jeremy Campbell, to understand his work better in laymen's terms. (didn't even have to look that up. I read it 33 years ago, and still smile when I think of it.) For example, when we discuss aliasing, anti-aliasing, Nyquist limits, etc., we're tip-toeing across a field that Claude Shannon planted, fertilized, watered, and harvested before most of us were born. Then there's the CPU in your computer... etc.

Here is a FAQ on DSD that has some helpful information, although it's principally about marketing some specific DSD products.

I find the following two paragraphs very interesting:

Because of the nature of sigma-delta converters, one cannot make a direct comparison between DSD and PCM. An approximation is possible, though, and would place DSD in some aspects comparable to a PCM format that has a bit depth of 20 bits and a sampling frequency of 96 kHz.[22] PCM sampled at 24 bits provides a (theoretical) additional 24 dB of dynamic range.

Because it has been extremely difficult to carry out DSP operations (for example performing EQ, balance, panning and other changes in the digital domain) in a one-bit environment, and because of the prevalence of studio equipment such as Pro Tools, which is solely PCM-based, the vast majority of SACDs—especially rock and contemporary music which rely on multitrack techniques—are in fact mixed in PCM (or mixed analog and recorded on PCM recorders) and then converted to DSD for SACD mastering.

So, DSD may be an improvement over 16 bit / 44.1K sampling, it's not necessarily an improvement over 32 bit FP / 96K sampling, other than the final output being somewhat smaller. Also, some past hardware based on 1-bit Pulse Code Modulation has been less than successful, because of (IIRC) calibration issues. I believe early attempts at an audio CD, which were not digitally processed but PCM technologies depending on the densities of the pulse codes, were the losers to the Digital Compact Disk, also by Philips and Sony.

So, I think it is probably premature to jump into DSD thinking that it's a mature technology for digital hardware. Maybe it is, and the articles just haven't caught up to the reality yet, but I'd do my homework before betting all my money on the future winner of this race.

Shooshie

[Timeline]

One thing in DSD's favor is the fact that there is no harmonic distortion from DSD 64 up.
When I ran Alpha Studios Burbank, Sony brought in a DSD 64 system and tapped into a live mix by an artist. They showed me the results of setup prior to the session by inserting a 10K square wave in and scoping the result. NO RINGING at all. If you did this even at 192 the Square wave would be deeply deformed. Some might say well, we don't record square waves but if you look at a trumpet, that is quickly proven wrong. The elements of attack of an instrument are greatly affected by transient response so I am one who would jump in quickly to this new, revised technology, DSD 128 and above and it has my vote.

I also noticed that when mixing to my ATR100 at 30ips, things sound purer. A square wave although tilted by effects of low end roll off, has little ringing and thats '60s technology.

[Shooshie]

If you did this even at 192 the Square wave would be deeply deformed. Some might say well, we don't record square waves but if you look at a trumpet, that is quickly proven wrong.

Not trumpet. Square wave is more at Clarinet.

Shoosh

[Timeline]

No, actually both:

https://www.google.com/search?q=Trumpet ... 97&bih=603

Trumpet has a much sharper attack though and harder to record on loud bright parts.

[Robert Randolph]

What the waveform looks like is completely irrelevant. A square wave is nothing more than a summation of odd harmonics, with the signal 'looking sharper' as you increase the maximum frequency. If you have infinite bandwidth, you will have an infinitely 'sharp' square wave... but we don't have infinite bandwidths and our ears are significantly more bandwidth limited.

If you bandlimit a square wave into the audible spectrum, you end up with that 'ripple' on the attack that looks incorrect. Regardless of this, it is what you are capable of hearing. Go try it if you want. Generate a square wave at a very high sampling rate then observe the output as you low-pass it. Now generate 2 samples at a high sampling rate, low-pass one at ~24khz and leave the other. Do an ABX test. You won't hear any difference unless your low-pass filter is very, very poor.

It could even be said that if your square wave isn't 'rippled' then you are seeing a very malformed representation of what is physically possible in the realm of acoustic sound, and especially what is audible by humans.

There are arguments to be made about DSD's quality, but how well it visually presents a square wave is definitely not one of them.

[Shooshie]

You can find flute, clarinet, sax, and trumpet waveforms that are hard to tell apart visually, and yet audibly they clearly impart the essential sound of their respective waveforms. Flute is essentially a sine wave. Clarinet is a square wave. Trumpet and sax are essentially triangle waves with steep attack ramps.

Individual players modify their sounds by learning to add specific overtones, such that you'd be hard-pressed visually to identify some of their sounds, but you can instantly tell them apart with your ears, which are supreme fourier processors. Not that this has anything to do with DSD, PCM, or analog tape. They all do a decent job of representing those wave-forms.

Shoosh

[Timeline]

I agree mostly Robert it is not the important aspect when/where frequency response, (bandwidth), in involved. But the proper reproduction of complex waveforms requires no artifacts to be added or taken away from a sound regardless of hearing limitations. What you have said is correct otherwise from my teachings/understanding and hearing experiences, filters removed helps this conversation quite a bit regardless of format. I have to date only experimented with identical mixes192 and DSD128 with my DSD recorder/player and found positive results for DSD128. I think, my opinion, and for my experience of 46 years in recording, it has come of age and is due us.

[Shooshie]

I agree mostly Robert it is not the important aspect when/where frequency response, (bandwidth), in involved. The proper reproduction of complex waveforms requires no artifacts to be added or taken away from a sound regardless of hearing limitations. What you have said is correct otherwise from my teachings/understanding and hearing experiences, filters removed helps this conversation quite a bit regardless of format. I have to date only experimented with identical mixes192 and DSD128 with my DSD recorder/player and found positive results for DSD128. I think, my opinion, and for my experience of 46 years in recording, it has come of age and is due us.

One question, Gary: when you are comparing those identical mixes, are you doing so with your ears alone, or are you supplementing the comparison with an oscilloscope?

Shoosh

[Timeline]

Headphones only. I agree its non technical. I can have a day where I can't hear clearly because of a cold or something but when I go back in a few days and listen I notice a clearer top and less dark result from DSD128. Wish I could try higher formats but the boxes are not there yet with exception to Sonic solutions. It seems though that DSD128 works fine for this.

[Robert Randolph]

I agree mostly Robert it is not the important aspect when/where frequency response, (bandwidth), in involved. But the proper reproduction of complex waveforms requires no artifacts to be added or taken away from a sound regardless of hearing limitations. What you have said is correct otherwise from my teachings/understanding and hearing experiences, filters removed helps this conversation quite a bit regardless of format. I have to date only experimented with identical mixes192 and DSD128 with my DSD recorder/player and found positive results for DSD128. I think, my opinion, and for my experience of 46 years in recording, it has come of age and is due us.

The part I bolded I disagree with. If you can not sense it, it's irrelevant. There's no point needlessly using storage for data that can't be perceived.

DSD as a format has plenty of positives and negatives, and one of those negatives is that for the high-frequency components you're discussing: DSD is less accurate than PCM! This is due to the massive amount of shaped noise present above the audible threshold that takes the place of PCM's nyquist filters. Thusly DSD also takes advantage of the fact that in-audible information is irrelevant to the perception of the signal.

I don't really want to debate which is better, but there are simple technical facts that exist regardless of which format is preferred.