The SACD vs Tape Challenge

[ceved]

"Done right' is the operative and overall contolling aspect which results in the sound of music many of us in this forum seem to prefer regardless of format.
I suspect that without too much trouble even the TP titles which we enjoy could be made to sound 'bad'.
Perhaps the real 'shootout' is in the steps leading up to the final product including raw materials and the care taken in the physical fabrication of the medium to say nothing of the recording, engineering, mastering and the like.
If you are going to limit yourself to analog,LP/tape not only are you are going to miss a lot of really interesting music well performed and recorded in the here and now you will also not be able to as easily confirm your loss of visual acuity as you can no longer read the CD notes.

The most recent Jefferson Starship album is a perfect example of music that I would have missed were I fixated upon format rather than content.

While I prefer good sound, I insist upon good content regardless of genre.  If it isn't worth listening to who cares how good it sounds?

[Ben]

I find with the few SACD's I have, I am limited to still 44Khz sampling rate for stereo.
As long as that is standard, 4 track, 3 7/8 ISP is good match. I'll trade tape hiss
instead of digital artifacts. Ben.

Footnote. I just finished a 6A5G PP amp, with cheap Hammond transformers.
I still can't decide if it is better than the SET 6B4G amp I have, but it seems
to clean up CD's better than I expected. This is using 50's technology here,
but with matched tubes and resistors and caps. I am guessing bandwidth is about
40HZ  to 18 KHZ but  I need to check that with the scope. The only part to get is
now is vintage feet and 2 track vintage music. :)  I'll continue to experiment
with my sound system for a while, since with analog you can still piece together
you think is a good system. So far it looks to be Dynaco SCA 35 clone in the future
with options added back that saved money and optimized for real triode use and
real bandwidth on the inputs.

 

[Tubes n tapes]

Arian,

regarding compatability with DSD, i have a few comments.

on the playback side; i've not heard an SACD player that is superior on one label but not on all labels. it's either better or it's not.

OTOH certain pro DSD ADC's did sound better than others. i was an EMM Labs owner for 7 years and there is no doubt that SACD's recorded thru EMM Labs pro ADC sounded better than the dCS pro ADC. and the EMM Labs mastered discs did sound better on EMM Labs consumer DAC's. dCs mastered discs also sounded better on EMM Labs consumer DAC's. Ed Meitner was involved with Sony/Philips in the creation of the first pro audio gear for DSD recording. Andreas Koch (who also did digital design for EMM Labs) worked with Sony/Philips in the creation of the DSD format.

even though SACD's mastered on the dCs were not quite as good (in general they had less pleasing high frequencies), those SACD's still sound very good. i think that over time there are fewer 'non-optimized' SACD's. i've not purchased one for 3 or 4 years which had any edge. i think that dCs has improved their ADC.

i think another problem with SACD is that some DAC's convert DSD to PCM prior to analog conversion (mostly this was older SACD players). it's not hard to find out which players do what and how.

like anything; there is SACD 'done right' and SACD 'light'. PCM also comes in many different flavors of filtering and upsampling, it's not unique to SACD.

mikel

Mike,

I have played for a while with both the EMMlabs and the modified Luxman in my system at the same time and it was very clear that some SACDs performed better on the EMMlabs while others performed better on the Luxman. The noise shaping algorithms of DSD on the recording side is vastly different from one SACD recording to another and the different playback systems handle that very different. This was very audible and also very measurable. I could measure noise floor differences of up to 20dB. Most of the time in favor of the EMMlabs but sometimes the Luxman was just as good, depending on the SACD used.
I was making analog recordings of SACDs at that time using my virtual surround process. Except for the noise that was very different, also the depth of the soundstage was remarkably different between the EMMlabs and the Luxman and again depending on the SACD the EMMlabs or the Luxman could be clearly better in that respect. So I ended up making some of the recordings from the EMMlabs and others from the Luxman.

Concerning DSD to PCM conversion, I think that is one of the flaws of DSD. As far as I have been able to find there are no true DSD DACs. All of the existing DACs will convert DSD to some level of PCM in the actual DAC section. I have spend quite some time trying to design a real DSD DAC, but I failed because there is a practical limitation you cannot get around. The only way is to start filtering in the digital domain and abandon the true DSD stream. Again if there is a real DSD DAC out there I would really like to learn more about that.

[steveidosound]

my experience is that CD is great for no hiss, but the dynamic range, depth and soundfield sucks compared to a good clean LP...I'm hoping for a similar experience with reel to reel audio, but I know there will be hiss with my Revox A77 at high speed...however, I feel confident that a half inch or inch reel at 15 ips is going to blow any CD or SACD, or any digital recording away regardless of sampling rate and resolution

Most everyone on here would agree with your "by ear" assessment for the most part, especially the "depth and soundfield" part.
You should know however, even if it is abused and misused on most pop recordings today, that plain old CD 16 bit / 44.1 KHz
actually has superior dynamic range to any analog media, if by dynamic range you mean from the point the signal is clearly dominating the background noise (in digital there should theoretically be no noise) to the point at which distortion becomes unpleasant (more gradual with analog - sudden and awful with digital)
All the soft to loud in between is the dynamic range or "signal to noise ratio".  Whether you can hear hiss on tape has everything to do with how loud the relative record and playback levels are for the track. If what you are recording has a 30dB dynamic range and you want to record it with no compression, and you drove the record levels to "zero" on the machine, the softest parts would come back at -30 and the loudest peaks at "0". Depending on how dynamic the insturment was, the worst being something with loud high frequecy peaks much higher than it's average, and long pauses, you would hear some hiss. Typically what happens in practice is that the instrument has a dynamic of over 30dB, but is compressed, and then the track is pushed harder - it varies as to what acceptable distortion is, so that the track might only be down to -15 on its quietest parts.
I don't know about 1/2" or 1" stereo tape at 15 or 30 ips., but 1/4" is pretty lucky to get into the mid 60dB signal to noise range and vinyl pressings somewhat worse than that. A freshly cut lacquer from a state of the art system might get to be as good as the 1/4" stereo 15ips. master.
Paul could probably quote specs for the 1/2" or 1" at 15 or 30ips. I think I recall an ATR demo. saying they had achieved the same dynamic range as a CD.

A big part of what people care about is how an analog or digital system behaves at it's noise floor or distortion limits. Good analog systems just rise slowly in distortion or distort a bit sooner at very high or low frequencies, thus giving the impression that they still have more left before severe distortion takes place.
Digital is a brick wall of distortion that you never want to hit, but which most pop mastering engineers flirt with all the time in the interest of making  the CD "play loud". At the other end if you turn up an analog system loud enough with unmodulated media going past, you will hear the noise floor. The music will just fade slowly into this noise floor as the recording volume is lowered. There is a last on or off limit to the CD, but the range is so wide that other things like analog microphone preamp noise will show up before you hit the digital wall on the "soft" side, but that range "wall to wall" is over 90dB.
How dynamics are handled or mistreated by recording engineers is the subject of great debate.
I am old enough to remember when people who cared were fighting for a couple more dB from their analog systems. Now we have all this range and in some cases we are using only the top  1-2% of it.     

[ironbut]

Well put as always Steve.
I don't get too wrapped up with worrying about tape hiss. Most of the old recordings we all love have enough hiss on the master tape that the 468 that the TP tapes are made on are well below that.
 One of the reasons that even very good digital transfers of these old master tapes don't have as much depth and soundstaging is that they've been processed. In an effort to remove tape hiss, drop outs, and any other audible problems many studios rely on software that repair these automatically. I use restoration software all the time on live recordings to do things like removing steady state noise like motors or hum, clipping or chatty neighbors. Removing talking/coughing without effecting the music is much more doable than removing the noises that last a long time. The software that removes those steady state noises are constantly making math based decisions on what is noise and what is music. The subtle cues that our brains have learned to associate with distance and location end up being altered or eliminated. Something that an engineer would never knowingly do just to reduce hiss another couple of dB. But the software has it's orders and that barely audible reverb that's right in the hiss's frequency range has got to go. A classic baby with the bathwater effect. For this reason, I've stopped using this type of de-noisers unless it's a really awful.
When I was doing demos of these tapes I had several guys who had owned reel to reels before. Almost all of them commented on how low the hiss level was.
What I feel is well worth the trouble is getting the timbres and subtle sounds just right. That's why so many of us really go all out to refine what comes after the tape.

[steveidosound]

...One of the reasons that even very good digital transfers of these old master tapes don't have as much depth and soundstaging is that they've been processed. In an effort to remove tape hiss, drop outs, and any other audible problems many studios rely on software that repair these automatically....The subtle cues that our brains have learned to associate with distance and location end up being altered or eliminated.... But the software has it's orders and that barely audible reverb that's right in the hiss's frequency range has got to go. A classic baby with the bathwater effect.... What I feel is well worth the trouble is getting the timbres and subtle sounds just right. That's why so many of us really go all out to refine what comes after the tape.

air...space...dimension...depth...

Sounds more like religious experience, but when it is right in a recording you know it and are transported.
The nuts and bolts might reveal higher phase consistency between the channels at high frequencies or a very low noise floor, but your ears say ahh, this is closer to what it is like in real life. At present our old analog technologies can, in their state of the art form, capture more of this detail than normal CDs can. I have stayed out of the higher bitrate digital discussions on this thread because I don't have enough experience with them to comment. I do like the sound (or lack thereof) in SACDs I have better than normal CDs, but I don't think I would rate it as better than the best analog sources I have heard.

[Ben]

I agree about the sound processing affects the depth of the soundstage.
Now we all need to do is  agree on what is the best amp for the music to be heard
for a modest budget. :)

[Hiro]

Concerning DSD to PCM conversion, I think that is one of the flaws of DSD. As far as I have been able to find there are no true DSD DACs.

just two examples luxman D-06 and esoteric SA-50 - both support DSD natively
http://puresuperaudio.blogspot.com/2009/03/new-high-definition-pure-audio-players.html

IMHO all original master tapes should be brought to DSD.

[ironbut]

Hiro, the grim reality of what actually happens in the case of archiving and production of master tapes/files is, what leads to the best sounding product doesn't always make the cut. I'm certainly not arguing with you. IMHO, these master tapes will all need to be transfered eventually and I'd like to see them archived in the highest quality that technology is capable of. To me this means with no error correction (if errors occur during the transfer, it should either be repeated until there are no errors or the archivist should correct the code manually) and no dither or other interpolation added. So, if I were to have a bone to pick with your statement, it would be that even 5.6 mhz DSD may not be good enough for archiving master tapes.
What is probably the most important factor in choosing a digital format for any kind of archiving is that it will be accessible to a wide user base into the foreseeable future. When we talk about one of a kind sonic treasures whether it's music, spoken work or newsworthy events, they need to be accessible to the folks that can bring them to the public in general. From an audiophiles perspective that may seem like a simple matter but even though the digital age of recording has only lasted a few decades, the number of formats that have come and gone and are still in use makes this vitally important. On a much smaller scale, if you look at some of the reformating software like Max, it can read over 20 file types and sometimes that isn't enough. There are already formats that have been out of use for 15+ years whose files can't be read since they're based on operating systems that were native to computers that went by the wayside. So, whose to say that of those 20 file types that Max can read today, how many of those will remembered in 10-15 years. My guess is 5 or so.
So, what it boils down to is just that,.. who is to say?
Harvard and Indiana Universities did a joint study into these very questions. They took advantage of parallel research from other universities and organizations such as The British Library, and the Library of Congress all of whom entrusted with preserving the worlds audible legacy. If you'd like to see the results of their research, here's a link to download "Sound Directions".   http://www.dlib.indiana.edu/projects/sounddirections/papersPresent/index.shtml
So, there's a lot more to archiving than meets the casual eye. I've put a little thought into the question of how these master tapes should be handled and I'd like to see a higher standard developed for these and other items whose exact details aren't entirely understood (are we certain that we've mined all that can be had from these tapes even today?). Maybe a physical scanning of the oxide particles would be more exact than playing their results? I don't know. But I do know that all digital formats that I've heard aren't up to the task.

[steveidosound]

This archiving thing is cool and should be another thread.
I don't think one can EVER justify the disposal of original source material... ever.
It is sad when it happens due to natural disaster or deterioration through improper storage, but to say that it has been captured in perfect form and so the original need no longer be saved is just the worst sort of fallacy.
For example, at the beginning of the analog tape era record labels decided to transfer all their pre-tape material cut directly to discs in the new state of the art tape format. At this point would we rather go back to the original disc masters or those 50+ year old tape transfer duplicates? The disc masters might have proved more durable than their tape replacements. And the quality of the pickups and turntables used to make those transfers, not to mention the transport, tape and head technologies have improved too. Compare an early 50s  Lp release of 30s era 78 material to that same disc master restored today. Need I say more. Who is to say what a transfer from that same disc master cut in the 1930s might be like 50 years from now?
But you say, that was then - NOW we can make much better restored copies and store them in space saving forms.
Restored according to whose values and standards, and on which digital format that will last forever? 10 years ago that might have been CD/ pcm / wav. files, but not so much now. 
Even though we are living in what seems to be an ideal age where every modern digital recording technology is pursued and pushed to higher limits and all the old analog ones are being researched and developed to a degree not thought possible in the day when they were the only technology around, we can never say we have arrived.
I wonder how much more can be known and understood that will make analog tape and disc better than the state of the art we have today?
So,
 Make the best duplicates and restore from that copy but never EVER toss the original source if it can be saved somehow and preserved.

[ironbut]

Truer words have never been spoken Steveo. I've seen a few folks get dragged down to the flour for suggesting that something should be tossed once a transfer is done. You never know what the future will bring and just like some of those mystery movies have taught us,.. sometimes the most important thing about a message is what it's written on!
The LOC (Library of Congress) just opened a special facility in Culpepper Va. that is primarily for the restoration and storage of delicate audio and video artifacts (musical instruments included). It was designed from the ground up (which isn't quite right since it's partially underground to save energy maintaining cooler temps) for that purpose alone.
I first got interested in this stuff when Mikey Fremmer visited the LOC and wrote about it in his Analog Corner column in Stereophile. Just the thought that every recording, book or anything else that is copywritten by the US has been required to send several copies to the LOC to have on hand since Congress is where copyright law is debated. Of course once a collection of that magnitude gets started, everybody starts to send stuff (hence the instruments). Collections of ungodly proportions are donated to them every year in hopes that they'll be displayed with the patrons name next to it. Just trying to keep things organized there must be akin to filling a bottomless bucket.
I confess that I just know the surface stuff when it comes to archiving and restoration but it is very interesting and important.

[steveidosound]

"If you'd like to see the results of their research, here's a link to download "Sound Directions".   http://www.dlib.indiana.edu/projects/sounddirections/papersPresent/index.shtml "

Fascinating stuff ! At least to me.
If I could choose a new career direction at 50+ it would be to work professionally as an analog audio historian/archivist.
I have been fascinated  by old media and transferring stuff from disc to tape and tape to tape since I was a boy.

[Tubes n tapes]

just two examples luxman D-06 and esoteric SA-50 - both support DSD natively
http://puresuperaudio.blogspot.com/2009/03/new-high-definition-pure-audio-players.html

Hiro,

Both the Luxman and Esoteric SACD players are fine machines. They support DSD, but the actual DAC chips (PCM1792A and AK4392) first use digital filtering to control the high frequency noise and then convert the data to a PCM like multi bit sigma-delta data to convert it back to analog. The PCM inputs on those chips go through a different digital filter but end up on the same multi bit sigma delta DAC. I'm sure both chips do a good job, but it is unclear whether these DAC chips will perform equally good on all DSD coded data streams.

When I said "there are no true DSD DACs" I meant that as far as I know the original beauty of DSD didn't fly because you cannot convert DSD back to analog as intended.

The original thought behind DSD was to create a 1 bit data stream that basically contains the audio information in analog form, but since it was digital data you could still store it on digital storage media. 
In the basis that original thought is still valid. If you take the digital DSD data stream and you hook that up to the analog input of you amp, you will actually hear the music. (Try that with PCM.) However the trick is in the details. DBX brought out a 1 bit sigma delta digital recorder somewhere in 80's that used a non-dithered and non-noise shaped A/D process. The drawback was that the S/N ratio was only about 50dB. But since this was DBX they combined that with their compander noise reduction system to get something like 100dB. It worked but was not suitable as a standard.
DSD is basically the same, but it uses enormous amounts of negative feedback and dither to move the noise below 20kHz to the area above 20kHz. From a digital point of view that works OK, however it has now become practically impossible to distill the analog information straight from the DSD stream. Even the slightest timing and transition imperfections will bring a significant amount of the noise that was moved away from the audio band back in. In the end if you can get 60dB S/N you're doing very well.

So what I have seen as practical solutions in all DSD DAC chips as far as I know is that the DSD principles are abandoned and the DSD stream is first converted to a multi bit PCM like delta sigma signal exactly the same as is done with high res PCM.

All these tricks make DSD about as vague and uncontrolled as our beloved analog media. Great fun, but not very practical.

[Hiro]

just two examples luxman D-06 and esoteric SA-50 - both support DSD natively
http://puresuperaudio.blogspot.com/2009/03/new-high-definition-pure-audio-players.html

So what I have seen as practical solutions in all DSD DAC chips as far as I know is that the DSD principles are abandoned and the DSD stream is first converted to a multi bit PCM like delta sigma signal exactly the same as is done with high res PCM.

All these tricks make DSD about as vague and uncontrolled as our beloved analog media. Great fun, but not very practical.

SA-50 and Luxman support DSD natively just like for example Pioneer PD-D9 (WM8741 DAC) - the DAC's have 'the option' to use DSD conversion to PCM but they can also play DSD natively bypassing PCM conversions...BTW take note that most of today's PCM DACs are really delta sigma, not to mention most of CDs available on the market were recorded with the use of sigma delta 'pcm' recorders


To me as a music lover the most important part in DSD is the sound.. but people working in the studio seem to be satisfied with the results they are getting from DSD too, BruceB from audioasylum says clearly "DSD128fs sounds identical to the master tapes to us."
http://www.audioasylum.com/forums/hirez/messages/25/256485.html#postfp

[docb]

Sorry guys, this has gone way too far afield and by now has nothing to do with tape. I have locked the thread.