Is higher better?
Join Our Community Subscribe to Paul's PostsWorking with DSD and PCM has been a real learning experience. The two formats sound different from each other though it is unclear why. Is it the analog to digital conversion processor? The DACs? That DSD is closer to analog? They are all different technologies.
On the one hand, DSD runs at a much higher sample rate than any PCM. Single rate DSD is 64 times higher sample rate than CD quality PCM. Yet, there’s not much more audio bandwidth available because of that higher sample rate.
And then there’s PCM. Few today would argue that 44.1kHz is the bare minimum required for decent reproduction. Anything less and we lose the audible frequency range. But double that, and now we can capture twice what we are capable of hearing. And 4 times that (176kHz) and we’re able to capture 80kHz. Much more than we can hear and more than sufficient for phase linearity.
Yet more seems better: 2X, 4X, 64X, 128X, and so on.
We don’t know if the “better” we are hearing is due to the change in architecture or sample rate or both.
But, in a way, what does it matter?
I have heard CDs trounce 192kHz versions just as I have heard DSD smash anything PCM.
In the end, I don’t think higher is necessarily better any more than I think lower is always better when it comes to distortion.
Gotta listen.
I think the main problem is, that so few really have the chance to make a meaningful comparison, which in my understanding would at first need a real parallel recording in different formats.
The difference between CD quality and hires, ok…the difference between PCM and DSD, ok…but how many then have a setup so resolving to determine differences between various higher sampling rates than CD quality? And what’s the frontend for a meaningful comparison between PCM and DSD…a DAC which converts both to DSD? Or does it have to be a DAC which plays back both natively?
I first would need a definition of a meaningful test environment (DAC to use) and test file creation (I guess many public files in various formats are down converted instead of recorded in various formats in parallel…and down conversion probably does its own harm)
Too many choices & comparisons can make your head spin & leave you with less time to enjoy (listen to) the music.
That’s why I didn’t do it yet extensively 😉
Yes, finding that three month period of utter boredom wherein you could really concentrate on all the comparisons that you’d like to undertake, is a challenge in itself 😉
I did this test. Ripped FLAC from CD (Pepe Romero Guitar Solos). Ripped WAV from CD and converted to 192/24. On my very basic system (Yamaha WXAD10 streaming from NAS) is clearly audible difference. 192/24 WAVs are full rich sound, flac sounds simple in comparison. My wife called 192/24 would sound like “older recording”.
I have performed blind AB tests on a number of albums issued in bath CD and HiRes (normally 24/96). I could always tell the difference, but was undecided whether one sounded ‘better’ than the other. Based on periods when I have been restricted to CD quality I would say that, on average, I prefer HiRes but this is no guide to individual recordings.
When it comes to electronics I would be confident in saying that lower distortion means greater accuracy of reproduction. This does not necessarily mean improved sound. Tube amps tend to have higher distortion than their solid state brethren, but it is a distortion that many find pleasing. For them a tube amp sounds better. Since we listen to music for enjoyment I am completely agnostic about this, although I personally am not drawn to the tube sound.
It depends I’d say. There are CD’s played back with my Harman/Kardon HD 760 which sound “more analogue” i.e. alive, than every HiRes-Format (PCM) played back with my Cambridge Audio Azur 851n and vice versa. But in fact (due to the recording) both versions sound fine. I stopped doing A/B-testing. It costs time and nerves (don’t know, if it’s said like that in english).
I just switched the caps of my vintage speaker-crossovers. That was sound-improvement at it’s best (for ALL analogue and digital playback options). Best regards!
“I just switched the caps of my vintage speaker-crossovers. That was sound-improvement at it’s best (for ALL analogue and digital playback options)”
🙂 I’ll second that! A few years ago, instead of re-capping my 1975 KEF-based speakers I substituted with the well-regarded Radford crossover. Only thing even approaching a night and day transformation I have ever heard!
I love my vintage speakers but they are still original. I have a pair of Pioneer CS99 and a pair of EV-1s by Electro Voice. Can you guys recommend a resource or YT channel for learning to re-cap? I’d love to try some DIY.
djB-O-B,
‘Talk’ (email) to Danny Richie from ‘GR-Research’ down in Texas, he loves helping people upgrade crossovers in loudspeakers.
He’ll even put the all the ‘bits’ together for you & ship ’em.
“Gotta listen”…wise words indeed.
I have to trust your opinion of DSD Paul, as I have not experienced it.
Of course I doubt very much that you would be pouring so much time, money & effort into
‘Octave Records’ if DSD wasn’t da bomb.
However, for a lot of us, there’s no point test driving a Ferrari if your budget only allows for a BMW & if 104 octane gas is rare.
If the music that I love, from the mid 60’s to the late 90’s, had all been recorded in DSD, there’s a very good chance that I would never leave my apartment, & that would definitely create problematic WAF’s that I do not encounter now.
I first heard the full dCS Vivaldi stack in 2013 playing some high quality CDs through Wilson speakers. I don’t think I’ve heard more transparent sound before or since. They pretty much invented all the audio DSD upsampling stuff and that system has by sheer presence been the reference disc spinner/DAC system ever since.
Marantz have been digital pioneers since the start and all their CD/SACD players have included their DSD processing system since 2012, which they explained with a paper here (three processes: Oversampling, Noise Shaping and Dithering):
http://www.marantz.eu/press/Marantz_Music_Mastering_Whitepaper_EN.pdf
The more recent SA-10 literature provides a simpler explanation thus:
“So DSD64, as used for SA-CD discs, is 1-bit at 2.8224MHz, and DSD256, or Quad-DSD, is 11.2MHz/1-bit. This means music stored in DSD is a much more literal representation of the original analogue waveform of the sound recorded, and thus needs much less processing to turn it into audio able to be passed to an amplifier. Or, as the company puts it, ‘DSD is analogue’.”
I have no doubt that remarkably good results can be achieved with 16/44 files through the processes described by Marantz, which I assume are fairly similar to those applied by dCS, PS Audio and others.
All of which is entirely separate from the need for higher sample rate source material. 16/44 has more than enough dynamic range and sampling rate than can be discerned by human hearing. The non-audio press speculate that HD files are mainly a marketing gimick as audiophiles will pay more for it. I tend to agree, although I think it is also a matter of convenience as so much music is mastered at 24/96 and 24/192 and it is just as easy to issue it at those rates. Linn used to master in DSD and were apparently at one time the largest source of SACD, but now everything is issued up to 24/192 PCM.
The undoubted merits of this processing seems evident from the number of manufacturers developing inceasingly cheaper and more powerful machines with this type of processing.
At the end of the day, most people seem to agree that mastering is the most fundamental issue, garbage in garbage out, and you can’t edit in DSD.
The creation of garbage starts with a poor recording technique neglecting basic laws of psycho-acoustics. The mixing console and the tools implemented is than used to minimize the audible garbage and to create a most artificial sound which should be acceptable on most loudspeakers of most different designs. In the Rupert Neve lecture (Link shared in an earlier post) Rupert Neve claimed that his analog console featured a dynamic range of 126 dB – much bigger than the range of RBCD!
I agree mastering is the key issue. Unfortunately, as a profession, mastering is no different than medicine or law or portfolio management. Lots of mediocre practitioners, few at the top. And, when it comes to DSD audio and 1 bit upsampling, the guy at the top, in my opinion, continues to be Ed Meitner of EMM Labs, both for consumer and studio applications.
That may be the case, above my pay grade. I see that dCS and EMM use the same Esoteric drive. Even in one of the world’s larger economies with a strong interest in audio, I don’t think there is a market in the UK for two SACD spinners costing £25,000. As just about every high-end store carries dCS, there is little room for EMM. Only one dealer appears to carry the brand and they don’t stock the transport, only processors.
Discussing such machines is a bit like discussing space travel in that the number of people with experience of each is probably similar.
Reading some comments, for a second I felt sorry for myself. I am one of those poor souls who are “restricted” to CD quality… 🙁
Fortunately…. I love it 🙂 Good enough for me.
IMO HiRes is overrated. I tried a HiRes subscription from Qobuz.
Not worth the extra money, so back to cd-quality subscription. Saves me $10 every month, which I rather spend on cd’s.
BTW, how much of the HiRes music is RECORDED in HiRes instead of “making” it HiRes afterwards because it’s fashionable nowadays ?
jb4,
I’m not sure whether it was you who posted sometime ago that you dabbled in SACD’s but ultimately you decided that they weren’t worth the extra costs.
However, I have noticed that the few HDCDs that I have purchased do sound better, to my hearing, that their original RBCD counterparts…I could list them but that’s not really the point.
Interestingly the HDCD versions obviously came from the same original analogue tapes, but probably remastered.
I’m happy that I bought those HDCDs & it would be nice to get my hands on some more because they do have greater clarity, however, in the main, I’m also in the RBCDs are ‘good enough for me’ camp.
Fat Rat,
I don’t think it was me who dabbled in SACD’s .
And to back that up : I have a total of 2 (two) SACD’s.
In both cases I found out it was SACD instead of RBCD after unpacking at home.
Look, I have nothing against SACD, but when I buy/bought one it’s more accidentally than deliberately.
I buy music because of the music, not because of the format.
And play a well recorded RBCD on a very good cd-player/transport and it sounds glorious.
Nowadays it seems a lot of people are more concerned about the format than about the music.
And don’t forget most of the available HiRes is fake, i.e. upscaled “not-so-HiRes”.
Just like they did (and do ?) with video, from SD to Ultra HD.
That may change in the future of course, but for now….
jb4,
Thanks for doing “2 (two)”…
…I could’ve got that completely wrong
I have an old 14 bit CD player with 4x sampling, another one not as old a 16 bit with 4x sampling, another 16 bit modified with a crown DAC 4x sampling, all Magnavox/Phillips. I also have a Conrad Johnson which is a modified Phillips 1 bit and I believe that has 256x sampling. They all sound different but the one thing they all have in common is they are said to sound more analog than a typical non Philips designed CD player. I have not heard DSD or any other high definition CD’s other than a mobile fidelity which is still the basic format. I’m not going to buy a DSD player when all of my CD’s and there are hundreds are the basic CD format because I’m not starting over again. If I want better sound I will buy a high quality Vinyl first pressing.
Joe,
From yesterday.
I checked out the NHT 3.3’s & 2.9’s on-line & yes they do appear to be quite impressive; the 3.3’s are hard to get a hold of now so good luck with locating a pair of those for yourself.
I wouldn’t be surprised if the PS Audio FR-30’s end up being a similar beast.
The Celestion – ‘Ditton 66’ floorstanders that I had for 38 years were also awesome in their sound & range; 3way with a 12″ & 90dB/W/m…160wrms max.
Notably they had one of the sweetest, most detailed & lively 2.25″ dome midrange drivers ever designed & manufactured.
I had a chance to buy the 3.3 numerous times. Problem is not many seller’s will ship them due to their weight and size. Some will but that runs around 400.00. The 2.9 are designed almost identical to the 3.3 except they are 50 pound’s lighter not as deep and use a 10 inch instead of a 12 inch side firing subwoofer. The plus is the 2.9 are easier to move around and in smaller rooms they are more practical and are not designed to be right up against the back wall like the 3.3 which are extremely deep and create their own corner. Some don’t like the way the 3.3 look being almost as deep as they are high. Might not pass the wife approval factor. I’m happy with the 2.9.
Celestions are known for the beautiful midrange and highs. I owned the original SL 6 and currently own the SL 6si. The NHT 2.9 equal or surpass the Celestions mids and highs while adding a nice low end. No stands needed either and a subwoofer is not really needed especially if you have the 3.3. I have never heard piano or voices sound so realistic than on my 2.9s. The Celestions have that too.
Joe,
Seriously, the ‘SL 6’ & the ‘SL 6si’ are not anywhere near what the ‘Ditton 66 Studio Monitor’ midrange dome driver (MD-500) could produce; I know because I used to sell them back in the late 1980’s/early 1990’s…just sayin’
I believe you. But the little buggers do image well. The midrange on my 2.9 or the 3.3 need to be heard to understand how good it is. My Celestions are in the closet because of my 2.9 and it’s not just because of their great bass. They excel through the entire frequency range. For small rooms the Celestions do well and with surprising bass response for their size.
Joe,
If you ever get a chance to listen to a pair of Harbeths, set up properly of course, do take the opportunity to do so.
Personally, they are the best 3D holographic imaging loudspeakers that I’ve ever heard.
I have heard good things about the Harbeths. I would love to listen to them. I want to listen to Vandersteen too. I heard them but not in any serious listening just when my local Hi Fi shop the Speaker Shop in Amherst NY were breaking them in, the bass was very nice and warm. I love warm bass and an organic midrange. I love great imaging too but they have to do more than just image well. They need to be musical and tonally correct. I wish I can hear a dozen speakers I never heard before. I’m interesting in hearing the Mirage M1. The NHT 3.3. The Infinity IRS V.
Joe,
Harbeth’s are the closest to ‘tonally correct’ that I’ve ever heard.
Alan Shaw, the owner of Harbeth, will tell you that if your main genre of music is Rock ‘n Roll, Metal or EDM then don’t buy his loudspeakers.
For everything else, I would venture to say that, you wont be disappointed.
Although, as always, it’s subjective of course.
After quite a bit of experimentation with the various formats, my preference is the standard 44.1. It’s main limitation has been the application of the technology, not the format itself. For whatever reason, it has taken basically 2 decades for technology and recording techniques to fully utilize 44.1 capabilities.
What I have noticed fairly consistently with DSD is added soundstage. When I compare most DSD recordings to the 44.1 format, most of the time placement of the instruments gets shifted in an unnatural fashion. Frequently, “stuff” in the middle gets pushed to the sides. Additionally, a consistent finding is a difference in tonality and volume of certain instruments. Where certain instruments stood out in the past, they now are now more precise but their sonic weight is diminished.
I’m not an audio engineer, but there appears to be a technology/knowledge gap on how to fully utilize the extra capabilities. With the formats changing so fast, who knows if there will ever be the luxury of time to fully vet out any new format.
Hi Reed,
Thanks for posting!
As someone who has not investigated DSD yet I find your comments very interesting.
As a mostly Rock ‘n Roll listener, I value sonic weight, clarity & dynamics over placement.
Cheers
The recording itself (which includes the recording venue) is the foundation of the quality of the listening experience. As you have pointed out, or implied, in other posts all is lost if the recording itself is not excellent. This seems to be the major motivation for developing your own capability to make recordings.
Can we agree that all formats from vinyl phonograph records, to analog tapes, to PCM, to DSD have the same function, to store and retrieve an electrical signal that is supposed to be an analog of musical sounds within the range of human hearing with the least distortion possible? If that is not the goal than anything I have to say is pointless because I don’t know what the goal is.
Outside of this link in the chain the signal can be manipulated many different ways at practically every process point in the chain. Of the four formats vinyl phonograph records is clearly the big loser. It is obviously defective and less desirable in many ways. Analog tape while better is also no prize. We’re left with two remaining formats, true digital encryption/decryption of which PCM is one strategy and RBCD is one standard for PCM and DSD which as I understand it is pulse width modulation akin to class D power amplifiers. DSD is therefore not a digital encryption method anymore than FM radio is. In fact DSD and class D are akin to FM radio which could be called carrier frequency deviation modulation by amplitude.
My experience is that because of the way recordings are made regardless of the storage format, deviations from flat frequency response back to the recording microphones are all over the place even for the same recording label, the same studio, the same recording equipment setup and even recordings by the same engineers using the same tools on recordings with adjacent catalog numbers. Therefore just to get accurate playback frequency response of a recording, an inverse equalization must be adjusted for each individual recording. Because of the designs of the systems I use, these deviations are more audible and correcting them properly is more critical because those differences are magnified.
There are many ways to test the accuracy of the storage/retrieve mechanism. Paul you have said that you could burn a CD from any vinyl phonograph record and the resulting CD is barely distinguishable if at all from the vinyl while we know the reverse cannot always be done. Why not try the same with RBCD/DSD. Use one for the source and the other for a copy to see how they compare. In one experiment the RBCD disc is the source and DSD is the test. Then reverse the process and use DSD as the source and RBCD as the test. If you do that please report your your results. BTW, in IMO differences of audible results in different bit rates and word sizes for the same recording are due to differences in the analog domain, principally deviation from flat frequency response. Dr. Mark Waldrep’s method for comparing RBCD with HD is so far the only one I consider a valid test method and with over 600 testers he concluded much to his chagrin I’m sure that HD adds nothing that is audible. As for MQA as I see it, technically it’s a joke.
I have listened to the same recording using different bit and sample rates- i can’t say they the music sounded different from
One mode to another.
I recall years ago I had a Cal Audio Icon Mark II and a mid for HDCD became available, which I had done by the company. I really did not experience any dramatic change post modification; I still have the CD player and it’s great!
The only major change I have noticed in this wild hobby was that when my amp was out for repair it sounded much better when it was returned vs when it was not there!
I think I’ll get back to enjoying a performance, period! BTW my home system sounds better than the same performance at the local music hall! Now I know why a night at the opera/symphony is the most expensive nap in town!
Larry
I subscribe to Tidal and listen to MQA audio and compare them to my CDs. In my opinion CDs sound better than the Tidal audio. Have an excellent NAD BluOS Streaming DAC C658 and comparing it with the Denon Surround Receiver via Helos, can’t tell much difference in quality of sound…CD is better in my opinion, much more life.
MQA is a solution that doesn’t work for a problem that doesn’t exist. Whatever your opinion of HD audio its audible differences have nothing to do with its increased frequency response capabilities beyond the audible range and its increased dynamic range capabilities. It is attributable to other factors that are unrelated. RBCD is good enough to perform its function flawlessly. As for MQA it doesn’t work because Audio Origami is a clear violation of the Shannon Nyquist criteria that is universally accepted as valid by everyone in the information technology and telecommunications industries except for advocates of MQA. They are just wrong.
larryro2, I used to have the same CD player and loved it too. I bought a few HDCD copies of favorite CD’s and compared; I could recognize the difference even when my wife helped and we did blind comparisons. The HDCD’s were richer and of course the bass was more defined. I moved on to a complete Mark Levinson system and then onwards to a PS Direct Stream player and DAC. I still remember the useful functions that I had to give up going to the PS Audio players. Ah well, the pleasure of a SACD is worth it.
OK, so let’s roll back a the years a bit. I’ve been listening to SACD’s for over 20 years now. But I had a simple Pioneer “Universal” (DVD/SACD/DVD Audio/CD) machine that dumbconverted the SACD layer to 88.2/24 Bit connected via RCA cables into a Denon AVR-1082 Receiver, played back through an Adcom 555II and my Homebrew Towers. I easily heard the difference between SACD and CD (Super Bit Mapped) layers.
When I mentioned this to Gus during my visit 1.5 years ago, his answer was. “It’s all in the mastering”. Whether it’s him, Bob Ludwig, Steve Hoffman, and all the other greats out there. Makes total sense to me.
Lots of terrific comments on here and this is why a few posts back on MQA I brought up some points about the gimmicks Of upsampling. Someone has already mentioned Mark Waldrep. I think it is great for people to listen to what he has to say, especially in regards to “bit bucketing” recordings that aren’t truly recorded in at least a 24/96 sample rate.
Right now I feel the notion of hi res audio can be so deceptive. I’ve even tried and own just about every sample rate except 32bit/384k and doing critical listening it sounds like a simple remastering.
Blue ray Audio is another matter. Audio that is recorded with video like a lot of my concerts blu rays definitely have a true difference in the 24/96k region compared to some redbook discs. A good example is my simple minds 6 disc BOXSET that includes a dvd audio of the album New Gold Dream in 24/96k LPCM. If I compare the the 24/96k version to the remastered cd at 16/44.1k I definitely get better results, however having the original non remastered CD compared to the DVD audio at 24/96k is quite…well…different. I find one really isn’t better than the other! it is a matter of taste because once again it just sounds to me like a different mastering job. However on the other hand if I had a ps audio SACD player that could be quite the game changer. Currently I get killed with the copyright protection on my wonderful chord dac. The only way I can hear DSD truly on my home system is through my OPPO SACD /bluray player. It is not nearly as good cause the OPPO transport has a mediocre cirrus logic 4396 DAC in it.
Lastly I’ve got around 35 of those MOFI discs, but I primarily buy them because of the great mastering and the cd layer on the hybrid discs are terrific, especially with my system Capabilities.
Long story short. I strongly think that mixing and mastering well, has more depth and importance than sample rates.