Analog vs. Digital, revisited

[ThomasOK]

I recently wrote a piece for dealers that covered my feelings on some changes going on in the industry. I felt that a slightly modified version might be an interesting starting point for discussion here. All opinions on my thoughts are welcome.

I’m sure we are all aware that the analog vs. digital question was first raised with the onset of digital tape recording and became a common question with the introduction of the CD. I also expect we all know that popular opinion has swung back to the superiority of vinyl records over CDs (something which certainly pleases me). Downloads and streaming are replacing CDs as the way most people listen to music and vinyl records are on the rise with new pressing machines being manufactured to keep up with demand. So the CD seems to be quickly fading. This could lead to a renaissance in analog music listening, and in some ways it has, with improvements in turntables, arms, cartridges and electronics (especially Lejonklou electronics, IMO) bringing us ever closer to the music.

However, the analog vs. digital question has been brought back into discussion by the advent of high-resolution digital downloads and the equipment to stream and decode them. The idea of digitally transporting this all the way to the speakers has also gained more adherents. Interestingly, some of the companies who argued most strongly of the inferiority of the CD when it was introduced, and who loudly decried the idea of “Perfect Sound Forever” as an advertising slogan, are now preaching the same message about high-resolution audio claiming the ability to bring the music to your ear “losslessly”. To accomplish this companies are recommending the downloading high-res digitally encoded music files, running them through a streamer, a digital preamp with digital volume control (DVC) and often digital crossovers and amps so that things only get processed back into music either before or after the power amp. (Now there are many variations here with some using digital amps and others not, some claiming that DVCs outperform any analog preamp while others say the best analog volume controls still outperform digital ones.) The reasons for this trend are fairly simple: 1) There are a number of younger digital engineers taking positions of power in Hi-Fi companies who feel they can overcome the problems with this technology. 2) Creating things like crossovers in the digital domain gives all kinds of abilities to adjust things that cannot be done easily, if at all, in the analog domain. 3) Digital circuitry is generally much less expensive to implement giving the possibility of substantially higher profit margins for the manufacturers.

But the big question is: “Does it really work?”

The answer you will get from me, and a number of others, is not really, at least not yet. While I feel that in the long term, with additional understanding, better paradigms and better electronic devices, there is real potential for a mostly digital transport system to be superior musically, I don’t think it is there yet. My own research, as well of that of others, indicates that the most musical systems at present are the ones with the least amount of digital handling of the music. I have yet to hear anything as musically enjoyable as a good record from an analog master tape, played on a well setup quality turntable, through quality analog phono stage, preamp and power amp into speakers with analog crossovers (be it passive or active). Many have found, and I concur, that quality mastered high-res digital, and even well mastered and well ripped CDs, can give very enjoyable music reproduction when played through a quality digital streamer and an otherwise all analog system. (Although still generally not as good as the same music from a good LP.) However, my listening has revealed that even the best DVCs are just not as musical as a really good analog preamp and tend to get worse as the listening level goes down. I have also never heard a truly musically convincing system using digital crossovers and/or digital amps. Also the room correction that these allow tends to make things more even but less musically informative and enjoyable.

As a matter of fact, it appears that the more digital stages a digital signal goes through, and the more transformations it has to make along the way, the more the musical message is damaged. Why is this? Well, despite the theory that “bits are bits” and current systems can transport things with no loss to the musicality, my experience has been the opposite. I, along with a great many others, can hear musical changes brought about by different Ethernet cables, different CD ripping software, different streamer software, different supports under digital streamers, etc. etc. In fact we have yet to find any part of the gear that feeds a digital streamer that does not have a musical effect.

So to sum it up, I feel that digital transport of music does hold promise for the future when the devices and parameters are better understood. But the reality is that we seem to still be a ways away from that time. In the hear (sic) and now we find that the more of the system that is analog (assuming equipment of high quality) the more enjoyable the result. I recommend that anybody seriously interested in music reproduction listen to a Lejonklou HiFi Sagatun preamp (either stereo or the sublime Mono units) compared to the best DVC equipped device you can find to hear the difference for yourself.

[matthias]

Thank you, Thomas, for this excellent contribution,

I would like to add that the most musical digital systems are able to playback both PCM files and DSD files up to DSD256 or higher as well.
For the best performance the digital pathway must stop after the Digital-Analog-Converter(DAC), then a straight analog pathway takes the signal to the analog volume control in an analog preamp like Sagatun Mono...............

Matt

[u252agz]

From a novice end user point of view, I find that some of my LPs, ripped CDs and 24 Bit Downloads sound just fantastic - but others don't reach the same heights, and consequently don't get played much, if at all.

I have no idea how the music has been produced - whether the original is a master analogue tape, or master 24 bit digital, or perhaps something else. I don't even know if the musicians had a really good day when the recording was made or how good the quality of the sound engineer / equipment used etc etc at the recording studio on the day.

Most of the time, I seem to have little choice as to whether I choose Digital or Vinyl, as the music is only available in one format.

It would be great to have all my music in the best possible format - which sounds like it is analogue master tape to good quality vinyl pressing, and a great playback system, but the reality is that if one were to insist on this - the range of music available would be quite limited.

The advantage of having both playback systems is the sheer variety of really great music available in all formats.

I do think good quality simple digital playback systems are easily good enough to really enjoy (at least some) digitally recorded music, and even if the quality might not be up there with the best, allows me to have free rein of the music world and many many hours of pure joy and relaxation.

Having said all this - I am looking forward to the day a Radikal PSU will replace the one in my Majik LP12, as the latter is already unbelievably enjoyable ( with the right vinyl pressing ) .

[Music Lover]

Great thread!

[hcl]

Great thread!

+1

The topic is quite interesting and having the oportunity to have a look on the recording side of things I have gained a slightly different perspective on things hifi. If one thinks there are big improvements to make from selecting the best equipment and tweaking the installation it is very subtle compared to doing the same in the recording situation. For example cables may have a some (more or less significant) impact in the hifi, but selecting the best cable when connecting the mic to the mic pre-amp makes a much bigger difference (not really surprising, but anyway). On the subject of old vs new there are a common opinion among professionals that the best gear, especially microphones are the vintage ones (60th, 70th,...). I wonder how much of the difference between recordings made with those mics when new and when used now when the same mics are rather aged. How much of the difference actually come from the old mics being ... old. Another huge difference between old and new is the way things can be done now a days. Much of the recordings made before where done in a few takes having the musicians playing at at the same time in the studio. Not much editing could be applied, because that also meant that the treble was lost and the noice unbarable high. This lead to the recording engineers taking great care to get it right, all at once. I guess that there are many (most) recordings from back then that was ruined because of bad equipment and bad handling, but it is not those recordings we have come to cherrish (long forgotten). The huge benefit from recording all at once is that the interplay between musicians are in place, just like in the live situation. This can never be done when everything is added one instrument at a time. There are other problems when recording many instruments simultaneously, but in my opinion these problems are much smaller than the musical gain from doing it like that. *My personal opinion* Some music can not be made like that, but the one that can can definitely benefit from such an approach.

Todays equipment enables the engineers to do endless edits where each edit only means a small deterioration to the signal (beside the desired effect). The engineers can also re-take as many times as they have time to and pussle together, edit pitch,... until they are satisifed. To me it is rather obvious that the whole meaning and musical context of a piece easily gets lost when having too much focus on the details. New gear not only means worse performance though. There are a lot of good gear also now a days and I would say the possibilities to make good recordings are better now then ever, but as with hifi one have to choose carefully to get really good equipment and then there are the handling and knowledge required to use the eqipment right. It takes a lot of practice and skills to be able to do state of the art productions. Not many have the time or the critical viewpoint to get there and with the editing possibilities of today comes the possibilities to do something wrong and there comes the problem of good enough monitoring equipment. As with hifi amps and speakers studio monitors differs a lot and although I would say that the most used stuff out there are rather good (compared to most hifi stuff), the best amps and speakers around (targeted for hifi only) are a lot better than what is used in most studios. There is an obvious problem when the mix engineer have worse equipment than what is later used to play back the final result. In the studio one can rather easy emphasise stuff and thus be able to hear into the recording when nessesary, but when it comes to getting the whole mix together the listening skills of the mixing engineer are put to an impossible situation when the monitoring does not let her or him hear all that can be reveiled with a better system. Monitoring of old was not better before, probably the oposite, but instead of leading to worse mixing results it lead to a different approach all together. Equipment and best practice was built up over time instead of one engineer making isolated desicions in front of the monitoring system.

I realise this is not exactly on topic, but I thought it could give some additional info to what ThomasOK initially wrote, which I think carries a lot of truth, but as many have recognised also digital recordings can sound very good, also on vinyl.

[dasher]

I also use recording equipment but in a home studio environment andI agree with everything that hcl says here. I was a recording musician in the mid/late 1970's and we almost always worked as a band in the studio - with separation but often together - leaving additional tracks for harmony overdubs, additional secondary instrument lines and possibly a master re take on the vocal. But studio time was expensive and the part that involved going to tape was horrendously so - even 8 track.

I used analogue and tape in the home environment for years, at one time running an 8 and a 16 track Foster on 1/2" tape. I then went 'half way' and used an Akai 24 track all in one unit and kept with this for years. This was fully digital without going to a Mac or PC. I now use a Presonus interface and Apple Logic Pro. Would I go back - no - it is simply so much easier and less time consuming to use. Most of the instruments that I record are analogue, some of which I still use microphones for ( and out of choice almost always would)and some I go through DI boxes - I tend not to use amp or guitar 'emulators' - if I want a Rickenbacker sound then I use a Rickenbacker.

I do use both digital and analogue keyboards - each has their place and their sound

Interestingly though, for playing back on the HiFi, I tend to master down to a 2 track Otari or Revox and therefore feed an analogue system into the pre amp (Kinos).

It is an interesting discussion - one that has been had thousands of times - but Thomas is not wrong to re-visit it as there have been groundbreaking digital developments in HiFi in recent years. To me there is just something about analogue replay in the home that is just easier to listen to. Digital Replay will improve further - that is the nature of technology and one day the gap will close. One day I may go to a piano recital, come home and listen to the file on my server and get the same emotional involvement - one day, not there yet!

[ThomasOK]

No question that the recording, mastering, etc. plays a huge roll in musical quality. It was not discussed in my original post as it is something we have little or no control over. I do find it interesting that, along with older mics, a lot of the most musical recordings date from the late 50s through the mid 70s. I think there is little doubt that the simplicity of the recording process dictated by two and four track tape machines and simple mixing consoles, combined with recordings that were generally the whole group live, allowed for more musical and visceral playing to be put down on the tapes and therefore records. Of course, there were producers, engineers and musicians who didn't create quality product, as there always will be, but there were also many pushing the boundaries of creativity and sound. Equally, the advent of multitrack recording and giant mixing consoles allowed some significant leaps in music creation, but also allowed some to screw things up more effectively than ever before.

Today there are an amazing amount of tools for musicians, and those who record them, to use in the production of music. In the right hands these can still create some very enjoyable music (Steven Wilson would be one example that comes to mind). But it seems more often these tools are used in such a way that they rob the music of its vitality. Splicing in every little, supposedly perfect, note giving a very sterile sound and correcting the notes for those who actually can't carry a tune in a bucket but are young and pretty (you know who you are). That these recordings lack life and emotion becomes more apparent when you hear a modern record that has them like "The New Standard" record that I mentioned on another thread a while back - a recording of three talented musicians playing live in the studio on an all analog recording. It just makes you realize how much is often lost with the more modern technologies.

[matthias]

I do find it interesting that, along with older mics, a lot of the most musical recordings date from the late 50s through the mid 70s. I think there is little doubt that the simplicity of the recording process dictated by two and four track tape machines and simple mixing consoles, combined with recordings that were generally the whole group live, allowed for more musical and visceral playing to be put down on the tapes and therefore records.

+1
Like there is a renaissance of vinyl and vinyl record players there is one of highend analog mixing consoles as well. I like very much old jazz records from the fifties, most of these recordings are superb. They are recorded and mixed analog, this is in my view the best technique for today as well. The best way to preserve the musicality of these recordings for digital distribution is the transfer to DSD256 in the final step or producing a LP anyway. Higher rate DSD is the most "analog like" digital format and needs very simple conversion techniques to analog. One example of this procedure is the excellent swedish label Opus3, at the moment limited to DSD128.
For playback of these files we need DSD capable streaming devices. IMO, we will see more "Audio over IP" AES67 (Dante, Ravenna) devices in the future. The advantage of Ravenna is that it is DSD capable up to DSD512, Dante is limited to PCM 192kHz.

Matt

[hcl]

I do find it interesting that, along with older mics, a lot of the most musical recordings date from the late 50s through the mid 70s. I think there is little doubt that the simplicity of the recording process dictated by two and four track tape machines and simple mixing consoles, combined with recordings that were generally the whole group live, allowed for more musical and visceral playing to be put down on the tapes and therefore records.

+1
Like there is a renaissance of vinyl and vinyl record players there is one of highend analog mixing consoles as well. I like very much old jazz records from the fifties, most of these recordings are superb. They are recorded and mixed analog, this is in my view the best technique for today as well. The best way to preserve the musicality of these recordings for digital distribution is the transfer to DSD256 in the final step or producing a LP anyway. Higher rate DSD is the most "analog like" digital format and needs very simple conversion techniques to analog. One example of this procedure is the excellent swedish label Opus3, at the moment limited to DSD128.
For playback of these files we need DSD capable streaming devices. IMO, we will see more "Audio over IP" AES67 (Dante, Ravenna) devices in the future. The advantage of Ravenna is that it is DSD capable up to DSD512, Dante is limited to PCM 192kHz.

Matt

I agree that a lot of really good recordings where made pre 1980, but have to say that all of them show huge discrepancies from an acousic live performance. I am really optimistic about the future though and do not think it has to be as limited as You suggest above. Some recordings might well be possible to do the way as You describe, but most wouldn't and to mee that implies that the big analogue desc instant downmix to two channesl and DSD conversion will be a method kept for a very small specialised part of music production. This method also have big disadvantages in that it puts limitations to what can be recorded and how it can be recorded in terms of instrument separation (everyting has to be done "live", leakage between instruments into the separate channels etc). It also require instant - analogue only - application of EQ, compression, reverb (it is and has to be applied, in virtually all recording situations and it has a huge improving impact on the end result - when used right) and other effects. on top of that the possibilities to adjust the end result for the actual monitoring equipment used is rather limited because that too also has to be done instantly.

I would say that the digital mixing options, available through PCM digital processing, in general have such big advantages over the above suggested approach that PCM in most cases will prove to yield a, in all aspects better end result. When the right equipment is used carefully and with a "most played at once" approach. The editing possibilities offered should only be used when and to enhance the end result, not to replace an otherwise good way of recording.

[dasher]

From a listening point of view I much prefer SACD to streaming to the Klimax - so I prefer DSD over PCM in this respect. This is probably due to a psychological preference to a vinyl (analogue) presentation.

From a recording perspective though al my equipment is PCM and it is there that I see the developments coming from. Sadly the R&D and will (probably economic) to pursue SACD and DSD just isn't there at the consumer end.

[lejonklou]

From a listening point of view I much prefer SACD to streaming to the Klimax - so I prefer DSD over PCM in this respect. This is probably due to a psychological preference to a vinyl (analogue) presentation.

That's interesting. I have never heard an SACD that performed better than a CD, when both are played on a disc player. Linn promoting SACD was a mystery to me and one of the first instances where I felt they were loosing their touch.

Streamers are in another league and I've never heard DSD streamed. Unfortunately, I don't see any future in it when it comes to the music. The only thing that makes sense is in my opinion to focus on the best possible reproduction of music in the dominating formats.

[matthias]

That's interesting. I have never heard an SACD that performed better than a CD, when both are played on a disc player.

Yes, me too.

Streamers are in another league and I've never heard DSD streamed. Unfortunately, I don't see any future in it when it comes to the music. The only thing that makes sense is in my opinion to focus on the best possible reproduction of music in the dominating formats.

I have heard high rate DSD and I like it. The fact is that in more than 90% of the DAC chips today PCM is converted to DSD before the converting to analog anyway. In the best chips like AK4490 or AK4497 DSD can be bypassed inside the chip so that the pathway for DSD is much shorter than that for PCM. Further it is beneficial to do the conversion from PCM to DSD not inside the chip but much better by a good software player in a computer. The algorithms for that are much more sophisticated.
IMO, there is a great difference between DSD64, the format of SACD, and DSD256.
I would like to have a streamer that can stream both PCM and DSD.

Matt

[Charlie1]

I'm surprised folks think things are worse now. For me, and this is a sweeping statement, I really struggled to enjoy new music in the 90s, be it radio or CD. I found it really hard work to the point of inpenterable - I didn't have my turntable setup for much of that period - maybe that would have made a difference. I don't know what was happening during that period. 90s recordings are ok'ish now, on a modern streamer or radio but still a low point. Some time around the early 00s I started to find new music become enjoyable again. I don't find new music today on modern streamers (or my LP12) anything like those that period. Anyone else feel the same?

Interesting to read Tom's latest thoughts on digital playback plus HCLs and dasher's comments on the recording side of things. Didn't know the old mics are still thought to be the best.

One other factor to add to the mix, they stopped using whale oil on master tapes around mid-70s too.

[lejonklou]

I have heard high rate DSD and I like it. The fact is that in more than 90% of the DAC chips today PCM is converted to DSD before the converting to analog anyway. In the best chips like AK4490 or AK4497 DSD can be bypassed inside the chip so that the pathway for DSD is much shorter than that for PCM. Further it is beneficial to do the conversion from PCM to DSD not inside the chip but much better by a good software player in a computer. The algorithms for that are much more sophisticated.
IMO, there is a great difference between DSD64, the format of SACD, and DSD256.
I would like to have a streamer that can stream both PCM and DSD.

Have you heard any of the players that use AK4490 or why do you say they are best? I'd be surprised if you've heard the 4497 as it's brand new - and perhaps not any better than the 4490.

You're right that on the reproduction side, almost all of the chips are ready for high frequency DSD. But when designing a product there are inevitable tradeoffs. If you optimise for DSD, it might perform worse with PCM. And if you optimise for 24/192 PCM, it might perform worse with 16/44, CD-quality PCM. This is unfortunate and something the manufacturers don't want to talk about. Their machines are always optimised for everything. I say that's simply not possible.

Far more problematic than the tradeoffs on the reproduction side, DSD is almost nonexistent on the production side. So in order for DSD to make sense, a conversion from PCM to DSD and then from DSD to analogue, both at the reproduction side, must perform better than converting PCM directly to analogue. Do you agree?

I am sceptical about this being the case, but would be happy to find that I'm wrong.

[matthias]

Fredrik,
I understand your position. If you want to optimise the DAC for PCM the best would be to go for a R2R ladder DAC. There are several manufacturers doing this:
MSB in the US
Metrum in the Netherlands
Soekris in Denmark
Totaldac in France
Aqua Hifi in Italy
Some of these are offering OEM modules of their DACs.
The other way is to use one of the chips I described in my last post. They offer a very good PCM playback, but shine with DSD. They seem to operate best at DSD256, so the best way to feed them is to convert all PCM and DSD to DSD256 with HQPlayer in a computer. The second best way is to feed the chip directly with PCM or DSD. They provide a direct path for DSD without volume control.
The third way is to create a DAC with a FPGA, like PSAudio, Playback Designs, Emmlabs, etc.. These DACs are in most cases optimised for DSD and are doing inside the DAC what HQPlayer does in the computer. Because software is cheaper than hardware these DACs are really expensive.

Matt

PS: You have a PM

[lejonklou]

Thanks Matt, but on what are you basing your advice? What have you listened to?

I haven't heard a single DAC that has impressed me. The digital world is full of high tech and clever guys that don't know how to listen.

The secret to performance lies in mad attention to detail. The more details you optimise, the better it gets. You can start with a ten year old chip and outperform everything if you perfect everything around it.

[matthias]

Thanks Matt, but on what are you basing your advice? What have you listened to?
I haven't heard a single DAC that has impressed me. The digital world is full of high tech and clever guys that don't know how to listen.
The secret to performance lies in mad attention to detail. The more details you optimise, the better it gets. You can start with a ten year old chip and outperform everything if you perfect everything around it.

Fredrik,
sometimes a suggestion originates from a gut feeling,
You are right. There are some manufacturers who are using your VC chip in their preamps but their products are not like your Sagatuns. I am sure with your attention to detail you will design a Super Streamer.

Matt

[lejonklou]

Fredrik,
sometimes a suggestion originates from a gut feeling,
You are right. There are some manufacturers who are using your VC chip in their preamps but their products are not like your Sagatuns. I am sure with your attention to detail you will design a Super Streamer.

Sorry, I didn't mean to sound blunt or cocky.

I am truly thankful for your suggestions, because often I am down in a hole, digging, and not paying attention to what others are developing somewhere else on the planet. You seem to always keep up with the latest.

I am, however, rather sceptical and have been underwhelmed by digital products for so long now. I want to create a streamer that makes vinyl feel soggy! But I may have raised the bar too high because it seems very, very hard to reach it.

[matthias]

Fredrik, no problem.
I do not know if you investigated high rate DSD, if not please give it a try.
BTW, I was never tempted to buy a Klimax DS.

Matt

[lejonklou]

No, I haven't. If it convinces even when converted from PCM (where nearly all the music is), then it's certainly interesting.

Then I also need to learn how to best do that conversion. Because digital processing can be bad for the music. Also when the math guys claim the operation is harmless or even lossless.

[TMV]

Analog, PCM and DSD pulse response (distortion in time scale):

(I don't know if I agree on the analog pro and consumer plots...)

PCM and DSD spectrum (distortion in frequency scale):


Since DSD and PCM signals are different it is impossible to convert lossless between them.
If you go from DSD to PCM "A" will be lost, if you go from PCM to DSD "B" will be lost.
Is A or B more important? This could help us to understand what format i "best", but to compare the formats you need to compare analog - PCM - analog with analog - DSD - analog.
Here you have same file in both PCM and DSD: http://www.ayre.com/insights_dsdvspcm.htm

Many people still prefer analog, even in the record industry where analog tape is still used. And vinyl rocks! :)

[matthias]

Sorry, but the graph is related to DSD64.
IMO, DSD64 is outdated in 2016.
With DSD128 or higher you have A AND B!
Please have a look at this comparison:

http://www.computeraudiophile.com/blogs ... erter-627/

Matt

[TMV]

Matt, it was an example to show that DSD and PCM are not the same signal.
And if yo compare DSD128 with PCM 32b/384kHz (or something) you will get the same.

[matthias]

Matt, it was an example to show that DSD and PCM are not the same signal.
And if yo compare DSD128 with PCM 32b/384kHz (or something) you will get the same.

With 24/88.2 PCM you will get maximum 40 kHz and never 140dB.
This is a simplified graph which has nothing to do with reality.
But you are right lossless converting is not possible.

Matt

[hcl]

Matt, it was an example to show that DSD and PCM are not the same signal.
And if yo compare DSD128 with PCM 32b/384kHz (or something) you will get the same.

With 24/88.2 PCM you will get maximum 40 kHz and never 140dB.
This is a simplified graph which has nothing to do with reality.
But you are right lossless converting is not possible.

Irrespective of the information difference between DSD and PCM any conversion between them will result in information loss. Almost all opertions done on any signal will also result in at least a small addition of noice to the signal. This is why signals should not be converted or processed in any way numerous times without caution.

(24 bit PCM can actually carry >140 dB of maximum signal to noice ratio)

Maybe we should stop this PCM vs DSD discussion in this thread, as it is a bit off topic here. I think it can be quite safely concluded that the CD format (16 bit, 44.1 kHz sampling rate) is a limiting factor on ultimate musical performance, but the main limitations from the CD format I would say are already addressed by high resolution digital download and high resolution digital streamers at the consumer end. There are surely more to improve on though, both at the production side and the consumer side and that is reassuring for a bright future. Among other stuff I'm thinking of the work done by Meridian on MQA. The biggest improvements still to gain is, in my opinion, on the production side, addressing issues associated to compromised productions and iPod listening optimized mixes and mastering (defeating the loudness war madness...).

-

So, what we have in our back catalogue is analogue recordings (mainly on vinyl) and digital recordings (mainly in CD format) and I would think that not much can be done to squeese these past the inherent limitations of each recording (may it me from recording, mixing, mastering or from the distribution process/format itself). What is ther is there and the damage already done can not be un-done (at least not with the processing theory known as of now).

Maybe we will (somtime in a faar future) develop algorithms to process or actually recreate the music form a musical understanding point of view, going back to what should have been there - in a musical sense - and totally re-build the signal and create what should have been there. I think this is rather far fetched (impossible!), but I think this is whats needed to *really improve* on what we get today from CDs and vinyl.