Monday, February 20, 2012

Analog Noise versus signal


Copland: Billy the Kid, Rodeo, Grofé: Grand Canyon Suite
Morton Gould and his Orchestra
Stereo/Multichannel Analog
Living Stereo 82876679042








Here is a plot spectrum of the music.














Here is a plot spectrum of analog and DSD noise in between Buckaroo Holiday and Corral Nocturne from Rodeo.














Transfered it to my computer with Audacity using the analog outs.

16 comments:

  1. Assuming error free storage/retrieval, would DSD downloads always sound better than PCM downloads, all other things being equal? If so, is there any reason why there are apparently no plans to make DSD (or DST, i.e. losslessly-compressed DSD) downloads widely available? If piracy is the only concern, wouldn't this be a tacit admission that DSD is superior to PCM, since PCM downloads are available?

    ReplyDelete
  2. I believe DSD downloads may sound superior to high resolution PCM downloads but I have no proof, I base this on my experience with SACDs, DVD-Audios and HRx DVDs.

    It's a chicken and egg thing, there will be more DSD downloads when there is software to play them. iTunes supports PCM up to 32 bit 192kHz, what I would like to see is iTunes include DSD in its readable formats. That may never happen as iTunes still doesn't support FLAC. Of course one needs a DSD DAC to play the DSD signal otherwise the computer will convert it to the highest resolution PCM available.

    Most people are transferring DSD downloads to DVD to create a DSD disc so they can play them without them being converted to PCM. I feel this is only temporary as we should soon have native playback on the computers other than Sony which is the only one I know of offering computers which play DSD music files. I'm a Mac person so we may be the very last to get native DSD playback.

    ReplyDelete
  3. After further searching, I found a forum with a lot of information on the relative sound quality of various formats by some apparently well-informed audiophiles (at http://www.computeraudiophile.com/content/SACD-vs-Hi-Res-Computer-Formats ), as opposed to the typical forums which I equate to "the blind leading the blind in echo chambers." Some seem to think that 24-bit, 192K recordings have a slight edge over DSD, while others seem to think that DSD has the edge over 24/192. So, the difference seems to be subtle and subjective, but it does make sense to avoid converting between DSD and PCM if possible, including by putting DSD downloads onto DSD discs and playing them on compatible SACD players. It's unfortunate that these players don't have a USB input so that the files could be copied to flash drives and played from them.

    Linn Records states that "FLAC files are lossless at various high bit rates, for example, 44.1kHz, 48kHz, 88.2kHz, 96kHz and 192kHz.... The quality is identical to that of an SACD (Stereo only)." However, they obviously wouldn't offer the higher rates if they all sounded the same, even post-interpolation.

    Linn has also just started adding high-res dubs from Decca and DG analog masters to its collection of downloads. I wonder if they plan on making DSD copies too, or if they're actually making the high-res files from existing DSD copies.

    It seems to me that there would be a demand for a service which would download these massive files onto flash drives and ship them out to those who want to purchase the downloads, but don't have sufficient internet bandwidth to download them.

    ReplyDelete
  4. In the course of searching for information on DSD discs, I discovered, through the PS3SACD website (http://www.ps3sacd.com/news.html), that there are a couple of new sources for DSD downloads, including Channel Classics.

    ReplyDelete
  5. While trying to get to the bottom of the issue of whether "on the fly" optical-disc error-correction is as good as its defenders claim, I ran across CD Ripping Terminology (http://docs.linn.co.uk/wiki/index.php/CD_Ripping_Terminology) on Linn Hi Fi's site devoted to this issue. Based on it, I now suspect that data errors aren't the problem, but the effects of the disc transport's servo mechanisms on the power supplies, which end up affecting the outputs. The study, The Numerically-Identical CD Mystery: A Study in Perception versus Measurement (http://www.prismsound.com/m_r_downloads/cdinvest.pdf), indicates that this effect has not been noticed in systems with separate transports and DACs.

    ReplyDelete
  6. High-end reviewer Robert Harley wrote a short article entitled "Robert Harley comments on the Memory player" on this subject in the December 2006 Absolute Sound (http://webcache.googleusercontent.com/search?q=cache:uoYJcpYYfiUJ:docs.hueaudio.com/tas/tas_167.pdf+%22dsd+master%22+%22error+rate%22&cd=11&hl=en&ct=clnk&gl=us), in which he was very emphatic that he knows for a fact that error-correction is essentially perfect, assuming that the drive works and that the disc is in good condition.

    So, this would mean that for cost-conscious audiophiles, SACD rules, now and probably forever. It's difficult to imagine a more ideal system, from everyone's perspective except those who expect perfect sound for free.

    ReplyDelete
  7. I'd like to use an inexpensive SACD player with an HDMI output as a transport, and a receiver with HDMI inputs and DSD-to-analog conversion as a DAC. Yamaha has some interesting receivers which might do this. They accept HDMI outputs from SACD players, and they have a "straight decoding mode." In this mode, digital inputs are sent directly to the DACs with no processing. However, I gather that in the "non-straight" mode, the DSD signals are transcoded into 24/192 for processing, and it has occurred to me that this transcoding might also be part of the "straight decoding mode." It probably doesn't make a lot of difference, but this question is nagging me. I've been searching the internet to try to learn whether this is the case, to no avail.

    Can you shed any light on this mystery? Perhaps it would be sufficient to know whether they use DACs with DSD inputs. (This raises the issue of whether these DSD signals can be accessed, which is almost assuredly impossible.)

    ReplyDelete
  8. It seems to me that if Sony really wants to popularize SACD, they should offer inexpensive SACD DACs with few if any functions other than to extract the DSD from an SACD's output-stream and convert it to analog, and inexpensive SACD players that could be used as transports in conjunction with these DACs.

    The reason for having a separate DAC would be to ensure consumers that the DAC isn't affected by the transport's demands on its power supply, the latter of which I've recently learned can cause audible effects if the supplies aren't adequately isolated from each other. [1] If the DAC and transport share a common enclosure, it's hard for consumers to be certain that this issue has been addressed.

    These inexpensive DACs would be designed to have excellent, but not cutting-edge stereo sound, so that audiophiles would aspire to more expensive gear with better sound and more channels and features. If it threatened to stifle demand for pricey gear, it probably wouldn't be put into production.

    Such a DAC could be made at a very low cost, considering that all that would be required would be to decode the data from the SACD and to convert it to analog. [2] But it would motivate consumers to purchase more SACDs. Eventually, when their budget permits, they would upgrade their systems.

    Another likely reason such a device isn't available is Sony's concern with minimizing access to the master-quality digital content of SACDs. However, it seems that Sony could reduce the decoding and conversion circuitry to a sealed module so that the decoding section would be destroyed in an attempt to access its output. Furthermore, if greater efforts were put into making pirates worry more, Sony could worry less about letting us have such a DAC.

    Notes

    [1] Optical disc treatments evidently improve sound quality in cheap players, not by reducing errors at the output of the error-correction system (which are essentially zero despite rumors to the contrary spread by audio hucksters), but by reducing the amount of work the servos must perform to keep the laser on track. This in turn reduces their effects on the servo power supplies, which in cheap players aren't sufficiently isolated from the audio section's power supplies, so that servo-supply fluctuations are passed to the analog section's power supplies, which affect the output signal.

    [2] DSD is actually two-level analog with the continuous analog information contained in the relative density of ones and zeros. It is "digital" in the sense that the signal has only two levels, and in the sense that the individual "bits" have equal widths. Conversion to analog is performed with a special but inexpensive analog filter.

    ReplyDelete
  9. Oops - "Two level analog" IS digital. What I meant to say is that DSD is a digital signal modulated and demodulated in an analog manner.

    Furthermore, considering that there is apparently little difference in sound quality between DSD and DSD transcoded into 24/192, it seems like the Yamaha receiver with a DSD decoder and 24/192 DACs is a good solution, assuming that it transcodes to high resolution and uses these DACs to something approaching their full capability. I just wish that Yamaha would explicitly state what it does, because sales-speak makes me suspicious that they're hiding some dirty little secret. The reviews on Amazon look good, but I don't know the supposed reviewers. For all I know, they'r written by Yamaha's sales department.

    Another issue is whether cheap SACD players with HDMI outputs send out DSD, or DSD converted to PCM.

    This is why I like to keep things simple and take a purist approach. If Sony wants to sell SACDs, then they need to prove the superiority of the format by offering simple, affordable, quasi-purist products that scream "this is SACD, and it is amazing" in order to erase all doubts about the format's superiority. Once converted to the Church of SACD, consumers would start buying SACDs and perhaps eventually upgrade their SACD-playback equipment.

    ReplyDelete
  10. Last year, when UK's HiFi World bench-tested Yamaha's RX-V567 (in the same "7" series as your Yamaha receiver, but lower down the line), it was found that a direct DSD>>>analog conversion was being done.

    from http://www.sa-cd.net/showthread/78922/79010

    ReplyDelete
  11. Can't get a straight answer on AV receiver capabilities

    My investigation into whether inexpensive AV receivers with HDMI inputs and DSD decoders convert DSD directly to analog has turned up an inconclusive answer from Yamaha's support department and inconclusive descriptions in an Onkyo manual.

    The Yamaha rep wrote that the output of the DSD decoder in "convered directly" to the speaker output, but he didn't specify what he means by "directly" or specifically rule out a PCM conversion. Is he hiding something, or just ignorant of the distinction between types of conversions?

    The Onkyo TX-SR313 manual, on page 32, differentiates "DSD" mode from "Direct" mode by listing them separately, although both are described as a lack of surround-sound processing. But this really doesn't prove anything, since they might have been designated differently to give the impression that they are different.

    A note listed on the same page indicates that SACD players with both DSD- and PCM-output modes might sound better in the PCM-output mode when played though the receiver's HDMI inputs, indicating to me that the receiver converts it to PCM, and perhaps to 44.1K or 88K, but perhaps only with the DSD mode off.

    So, after all that effort, I really don't know anything more than I did before.

    ReplyDelete
  12. The easiest way to get an answer might be to get a copy of the service manual, and note whether it has DACs with DSD inputs.

    ReplyDelete
  13. SACD "tranport/DAC" setup doesn't have to cost a mint

    According to Yamaha customer service, even their lowest-price receivers have 5 CHANNELS OF DIRECT DSD-TO-ANALOG CONVERSION! Here is the quote:

    Here are the spec's for 3 models. (Yes they are the same.)
    RX-V371 DSD Spec:
    DSD 2/5.1ch, 2.8224 MHz, 1 bit SACD
    RX-V671 DSD Spec:
    DSD 2/5.1ch, 2.8224 MHz, 1 bit SACD
    RX-A1010 DSD Spec:
    DSD 2/5.1ch, 2.8224 MHz, 1 bit SACD

    So, an inexpensive Yamaha receiver, used in conjunction with something like a Pioneer DVR-610, which sends all 5 channels of SACDs in DSD format out via HDMI [1], would approximate an expensive SACD transport and DAC. This setup might not have the cutting-edge sound quality of a Lindemann $20K SACD player, but it would probably be very good, or at least very analog-like, without having to wonder if the transport is jerking the analog section's power supply around.

    Notes
    [1] I base this claim on the following from Pioneer DV-610AV-K Owner's Thread (http://www.avsforum.com/avs-vb/showthread.php?t=1310489):

    Today, I've tried the DV610's audio setting at Auto(DSD) via a HDMI 1.4 cable to a Denon 4311ci receiver with the DV610 player set to 1920x1080p output, and the 4311 receiver picks up the SACD audio as DSD.

    DVD-Audio discs in the DV610 show up as PCM 96 kHz in the 4311 receiver.

    If the DV610 audio is set to just Auto, SACD audio shows up as PCM 88.2 kHz in the 4311 receiver. If DV610 is set to Auto(DSD), the receiver just reports DSD without a bit rate (sounds better than the 88.2 kHz PCM signal, but that's subjective).

    Again, this is all with the DV610 video resolution via HDMI set at 1920x1080p.

    ReplyDelete
  14. Howard RothsteinMarch 7, 2012 11:32 AM

    Hi ... i'm just your average audiophile, with love of SACDs, and also i dabble around with computer audio ... based on discovering your blog and Positive Feedback articles, i've just purchased a bunch of Telarcs ... I have a question, what is the system setup for capturing SACD plot spectrum via analogue to Audacity on a Mac?

    ReplyDelete
  15. I use the tape deck loop in my tubed preamp to output SACDs to my computer using what is called the "analog hole" into my Mac Mini computer. On the Mac Mini the audio input is a single mini stereo, if yours is the same you would need a 2 RCA to stereo mini cable. I purchased a Monster Cable branded one at Amazon.

    Record an entire track from the SACD in question, the Audacity manual can help with setting recording levels.

    Follow the Audacity set up for spectrograms and plot spectrums outlined in my article http://www.positive-feedback.com/Issue60/hirez.htm

    Hope this helps.

    ReplyDelete
  16. I am in the process of converting all my SACds into 24/96 PCM and storing in itunes. From SACd output to a Korg recorder and then in audacity do the conversion. While DSD sounds better to me, it alas is a dead end when having to use a music server. Dont know how long the market will support SACD players either. Sad state of affairs.

    ReplyDelete