I'm so glad CD's came out in the early 80's. Because we were able to get a lot of good music put on a digital file before they decided to remaster old records to raise the volume and brick them. When I buy older music, I always have to buy the first release they did on CD. Because the later remastered versions usually always sound like crap and distorted.
In the case of one of the first direct to digital recordings Donald Fagen’s The Nightfly. Roger Nichols was quoted saying they made the decision to record the Nightfly on digital equipment. He said "We booked the Village Recorder in 1981 to cut tracks for Nightfly and decided to try the 3M digital machine. We ran a Studer A-80 24-track analog machine in parallel with the 3M for the test. After the band laid down a take we performed an a-b-c listening test. The analog and digital machines were played back in sync while the band played along live. We could compare the analog machine, the digital machine, and the live band. The closest sound to the live band was the 3M digital machine. We re-aligned the Studer and gave it one more chance. The 3M was the clear winner. We rolled the Studer out into the street, (just kidding) and did the rest of the recording on the 3M 32-track machine. When it came time to mix, we mixed to the 3M 4-track machine." The original resolution of the 3M recording was 24/50 and best you can get is 24/48. Comparing the 24/48 version to the 16/44 CD version is a good A/B test and cool way to test the resolving power of your system.
Back in 1971 or 1972, I had an open reel tape that was cut from a digital master. It was digitally recorded by a company named, I think, SoundStream. I think they were the first company to digitally record music, at least on the commercial level. I got it from the old Windsor Tape Club back when I was a student at UNC (class of 1973).
A bit of trivia. The 0.1 KHz of the 44.1 KHz is only there because the 16 bit Digital audio frames at 44.1KHz fitted nicely on each scan of the recording heads on a Sony U-Matic video cassette which was at the time a convenient way of transporting the digital master to the CD mastering/pressing plant.
Back in the 70's, I used to buy "half-speed" mastered LPs. As I understood it, the pressing master was cut by turning it at half speed (33 1/3 divided by 2), with the master tape playback running at half speed. These were about twice the price of a normal LP, and they definitely had more detail.
They did, but they often lost low end as bass started to fall off below the low frequency limits of the equipment. For example, high speed duplicators would lose the high end because the high frequencies fell beyond the capabilities of the duplicator.
I think the thrust of the question was really more about older analog recordings transferred to high resolution digital vs Red Book, more than about upsampling earlier digital recordings. SACD and DVD-Audio came out in around 2000. In the early days of SACD, probably 95% of the releases were from older analog masters.
It's interesting to say, however, that the older analog recordings were very high resolution. As far back as the 50s, the recording equipment and techniques were astonishingly good. If you could get uncompressed copies of the masters, you'd think they were among the best modern recordings.
There are definitely a lot of great recordings from decades ago but I've found that the best of analog starts in the 1970s. Which recordings from pre 1970s are astonishingly good? Right now while writing this message I'm listening to a 96kHz 24 bit digital version of the master of Apr 30, 1960 Satchmo Plays King Oliver Louis Amstrong St James Infirmary from Amazon HD Music. It's not bad, but it's not really awesome either. But hey, on a good audio system, music from the 1960s sounds much better than modern music on the average Bluetooth player or earbuds that the majority of today's younger consumers are using.
@@graxjpg In the 70's WBUR had a program every Sunday afternoon called Adventures in Sound. Edgar Vilcher used play tape masters of classical albums he owned and send them to the transmitter using a microwave link instead of the telephone lines that were usually used between the studio and transmitter. I used to record those on my Revox A77 two track deck and the sound was very good, better than vinyls of that same recording. Now mind you this was on FM which is limited to 15khz and the Revox, while pretty good, was not the best available by a long shot and yet they sounded better to me.
Jazz at the Pawnshop. An audiophile standard, recorded in 1976, by a professional recording engineer who dragged his equipment to a bar in Sweden to record a pretty decent jazz combo. The music is good, imaging is perfect, and the ambience is real with clinking glasses and quiet conversations in the background. The recording was made on analog tape. The vinyl record is a sonic delight. Naxos has re-released it in a 2x (128 bit) DSD format. The DSD makes the pure analog vinyl sound like there is a curtain between the listener and performers. If done properly, there is absolutely merit to upsampled releases, and DSD is the King of formats!
I have it on SACD, and it sounds amazing. It's a good demo for when you want to show off the benefits of SACD/DSD. Without digging it out of my storage boxes, I don't think it was Naxos, and I certainly didn't think it seemed there was a curtain between me and the performers.
Great explanation. I feel like I have spent half my life trying to explain digital audio to 'audiophiles'. Just one thing to add that once you have recorded in DSD it can be converted and compressed (so long as it is a lossless alogorithm) for storage or transmission but it should be converted back to DSD before conversion back to analog audio. It shoudl be said that the so called '1 bit' digital audio is just another way of encoding the analog signal: it is a system known as Pulse Density Modulation (used to be called 'pulse width' when I was a boy). CD used Pulse Code Modulation (PCM) which sends the bits to a DAC in chunks of 16 at a time. At the time CD came out the frequencies required for pulse density modulation were beyond the capability of the digital electronics (8 MHz was a very fast computer processor back then; now we use processors at 4 Ghz). The trick is you can convert PCM encoded audio to PDM (SACD 1 bit) before converting it to analog. The reason for doing this is to get the benefits of using the much simpler converter electronics. Since the analog signal is actually present in a PDM bit stream the converter is little more than a low pass filter with a very gentle roll off ! If anyone wants to understand how this works look up Fourier Analysis as it relates to square waves (It is actually how the human ear does it !). Last thing if you record to PCM you do loose some benefit compared to DSD recording but at modern sample rates for PCM the difference is negligible; far greater harm is done if you try to do any processing on the digital data ; it must be converted back to analog for processing then stored as DSD again to preserve the benefits. Sorry to rant a bit but I can't fit this on a T-Shirt (something to do with data density ?)
Trying to follow the thread of this answer is an exercise in futility... it's partly the fault of the question, which encompasses several topics that can easily sidetrack Paul - which is pretty easy to do. What the questioner really wants to know is whether a 192 kHz digital master taken from an analogue tape recording compared with a 44.1 digital master taken from the same analogue tape is better. There's an assumption that older tape recording techniques are inherently worse. Listen to a high res version of "Rumours" and tell me that that recording is worse than one done today. It sounds better than many recordings being done today on digital equipment.
Paul actually did answer that, stating that although the 192kHz won’t add musical information, yes it will sound better as it allows the filter to be more benign than the brick filter used for 44.1kHz.
Analogue has a higher resolution in time as well as in signal value.[1] It also has imperfections like noise that can "hide" sound and ambience, like on these old cassette-tapes that were recorded without noise reduction. I was surprised at how much detail was present below the noise level, after I applied a program that removed that noise.[2] If I can still get this program to work, I wonder what this can do to old studio tape recordings! [1] However, it likely also has more dynamic issues like non-linear distortion, compression, slew-rate, tape saturation, speed-issues[ like wow and flutter, and noise or hum. These are audible imperfections, but not always bad. They can sound so good that digital studio workstations have tools to mimic them: "tube sound" or "tape sound". We like it! [2] It is rather technical, but if you remove the noise in the time-domain in a sensible way, the resulting sound is much better than removing it in the frequency domain(which most programs do). And yes, I first had to digitize the recording, but the noise level of a tape recording is probably about 12 of the 24 bits, so there is something to gain.
This reminds me of movies as well. Some film restorationists don't believe that there isn't much benefit to 4k or 8k scans of old 35mm films because the optics and film available at the time weren't so high quality. You'll see heavier film grain and optical flaws. However there are some benefits such as medium format such as 70mm or VistaVision, restoration work and archival work. Though some will argue that 35mm film has a resolution equivalent to 8k.
SACDs derived from an analog master tape have the capability to sound far better than CDs provided that there is not an intervening PCM digital step in between.
This is correct. What we find though is that with most popular music recordings from the late-80s onwards is that the 2-track master was digital. Most likely a DAT at 44.1 or 48kHz. There were a few other solutions that might've given a few extra bits in the Digital word-length. What we need for a true HD re-master is for the analogue multitrack tapes to be digitised at say 24/96 or using DSD and a new 2-track master produced. When we get into the 90s, we have more and more multitrack recordings made in the digital domain. A new master can be produced, but there's no more detail to tape to be had from the source recordings. Some of the big 80s artists recorded in a mix of Digital and analogue and their re-masters have been elaborate affairs / labours of love.
I have a few SACD‘s ...but non of them sound really outstanding! Whereas most of my regular CD‘s (with and without HDCD) sound amazing!! It’s the recording not the medium
@@bikemike1118 It's true, they can show up a lot of the creaks and joins in recordings from the 1970s. I'm thinking specifically of the Queen back catalogue that was released in Japan on SACD.
@@bikemike1118 In your case, it might be that the master recordings may not be very good to begin with, or the mastering of the recordings for the SACD are poorly done. SACD is not necessarily about sonic spectaculars (though it can deliver that in spades), it's more about getting closer to the best sound that a particular recording can offer. If you don't mind sharing, which SACD titles were you not satisfied with, and does the CD layer of the SACDs you have (if applicable) sound much different in comparison to the SACD layer?
Paul, unless I have misunderstood you, I think the brickwall fltter is placed at the upper end of the audio band not above the sampling frequency; that is, for RBCD the filtering has to take place between 20kHz and 22.05kHz and that's why it has to be so steep..As far as early SACDs are concerned, I bought a number of those very early Sony's and as I had many of them already on either CD or LP I was always amazed at how good those SACDs sounded - better, in fact, that many more recent ones!.
So is the DSD format our gateway to hearing original, uncompressed master tape quality? And is the original, uncompressed version embedded in a PCM? Because we aren't having to deal with a physical product anymore, it would just be about the limits of storage, which are relatively unlimited nowadays. It's almost like we finally have as close to a perfect way to record now. It's very exciting.
many companies in charge of these things dont anymore to them its just a fast cash grab since the majority of the listeners wont even notice the change anyway sadly
isnt that common knowledge? it can never sound better then when it was recorded so if it was recorded in a bad way or mixed in a bad way no matter if you have a 50 million sound system it will ultimately sound bad
Again, Paul does not answer the question and bangs on about DSD. The answer is that copying an old recording to another digital version cannot recover detail lost in the original recording. It may be a good copy but it cannot be better.
DSD is intriguing, but I listen to mostly rock. DSD recordings for sale online are jazz & classical. HDTracks has Beach Boys, B-52s at 24/192 , but there is no info on how these were mastered. I don't know if they are upsampled or legit remasters for hi res. It would be nice if there was a site that rates the audio quality for HD files.
oh god yes totally agree i recently upgraded my DAC and unintentionally got MQA support so i figured hey why the hell not to try Tidal and imported my spotify playlists and was casually listening then a few albums just sounded Terrible it was like my youngest son who is 4 years old had mastered them not sure why or where but something on that "Master / MQA" track was just off
If the original file recorded or copied in the studio is 24 bits and you change it to 16 for cd, and changes it back to 24 bits there is no loss. it just restructures what was there. "Let it be a signal, limited in band, and its time interval divided in equal parts, so that intervals are obtained such that, each subdivision comprises an interval with period {\ displaystyle T} T seconds, where {\ displaystyle T} T is less than {\ displaystyle f_ {m} / 2} {\ displaystyle f_ {m} / 2}, and if an instant sample is taken arbitrarily from each sub-interval, then the knowledge of the instant amplitude of each sample plus the knowledge of the moments when the sample of each subinterval is taken contains all the information of the original signal. " The first 24-bit recorders began to be used in the late 1980s, I think. We have many albums that we can be improved to listen to in streamers.
This didn't actually answer the question I was hoping he would answer - my question is that we know old recordings weren't done digitally - they were done analog and presumably when they reissues a "remastered" SACD version of the same music that was previously released on CD format, they are digging out the old analog recordings and remastering them with modern digital equipment. The question is, how much better should we assume that this is going to be than the 16/44.1 version from 30 years ago? How good were the analog recordings? How much better is the modern digital conversion gear? My guess is that the answer will be "it depends" but I'd still very interested in Paul's thoughts here.
short answer is stay away from remasters. you have a good source one you like the sound of you can dabble in new masters and versions but there is no guarantee that it will be better Long answer is there is no way to know if they even used the original recording so it wont really matter what they do or call it if they used the original recording and if that recording was good there is a high chance the new version will sound good. but there is many issues with this who did it? we all put our own flavour on things even unintentionally so really no matter what it will be different some for better some for worse. The digital gear is better then the analogue that's for sure but you are 100% right it all depends on how they do it and who does it and how experienced he or she is
What about by favorite album Brothers In Arms by the Dire Straits they recorded the album on PCM for CD, but in the 1980 what where they recording at 16 bit 20 bit 24bit what and if they recorded it at 16 bit or 20bit why use DSD?
It's like Boston's first album. Recorded on some of the best analog stuff out in the 70's. Sounds great on CD also. SACD not so much. If I ever wanted a SACD to shine it was that album but was a huge disapointment.
yeah some re-releases on new tech just sounds bad i upgraded my dac a while ago and unintentionally got MQA support so i figured id give tidal a try to see what all the fuss was about and was casually listening and some of my favourite albums sounded like something my 4 year old son mixed together just killed the music and i went to the CD and compared it just to make sure i didn't remember wrong and nope something somewhere got messed up
Thanks easy street photo but It's everything to do with the Nyquist theorem. A sample rate of 2f is just the bare minimum and more is better. The 50MHz refers to the analog sinewave bandwidth. The newer version of my oscilloscope is still 50MHz analog sinewave bandwidth but with 2GSa/s. Paul is seemingly incorrect about the aliasing problems with higher sample rates - aliasing occurs with insufficient sample rates particularly with modulated signals where incorrect peak voltage readings can occur. (i.e. the sample rate is such that the peak voltages are never sampled.) The only disadvantage of having higher sample rates is the higher amount of storage memory required for a given period but that's no issue in this day and age.
@@PebblesChan rule of thumb with scopes is 5f, not 2f. but your sample rate isn't your bandwidth, you're getting 20 samples a cycle - that has nothing to do with nyquist, your bandwidth, any of that. you just get that many samples. that's not bandwidth.
@@easystreetphoto2401 Apologies for my poor language. English is a second language for me. Let me try to clarify for some practise. The marketed bandwidth specified for an oscilloscope is the bandwidth of the analog front end prior to and independent from the sampling section. Most likely the bandwidth specified is the -3dB for that front end. The same oscilloscope is sold with many different bandwidth analog front ends ranging from 20MHz through to 200MHz - all feeding the same 1 or 2GSa/s sampler. The Nyquist theorem specifies that the sample rate has to be at minimum of 2f or greater of the observed signal frequency. The rule of thumb of 5f for scopes for the display of more complex waveforms than just a constant sinewave where there are higher frequency components present such as but not limited to triangle, saw, pulse and square waves so that the displayed signals are acceptable. This can be demonstrated by displaying a 20MHz square wave of on a 20MHz bandwidth oscilloscope vs a 100MHz bandwidth oscilloscope - The 100 MHz scope will display squarish wave and the 20Mhz scope a distorted representation something akin to a sine wave.
Paul never answered the question , which is yes high resolution files from the 70 and 80's can sound better than red bool CD's especially if they remastered it from the original multitrack tape which had a higher dynamic range than vinyl records could handle . Some recordings even had frequencies above what CD's are capable of and above human hearing but research shows we can somehow perceive them anyway and some high resolution files of music from back then preserves them.
Magnetic tape masters from 40 or 50 years ago will probably be in very poor condition. The only useable masters would be an early digital archive taken when the original tapes were still playable.
@@geoff37s38 Tape stored in an environment with temperature and humidity fluctuations help degrade tape much faster than tape stored in a controlled environment .
@@geoff37s38 I have nearly 40 year old cassettes and if they lost their HF it's not apparent, but they were CrO2 tapes not Fe02 . Anadialog has compared a 50 year old Pink Floyd DSOM cassette to vinyl to CD to SACD and in the show more section he provides High resolution files to compare. There was no hint degradation of HF on the tape, check him out on TH-cam.
@@leekumiega9268 I used to have a Brenell reel to reel tape deck. A really good bit of gear. The version I chose had 3 speeds up to 7.5 IPS but there was a 15 IPS version also. Apparently each of the Beatles owned one. They also made studio recorders. I gave it away many years ago.
@@MrBiggmartin over 50 years ago when I was studying electronics engineering the term Nyquist Criteria AKA Nyquist Criterion is defined just as Paul said. He didn’t remember the second word so I filled it in on my comment.. www.analog.com/media/en/training-seminars/tutorials/MT-002.pdf
@@MrBiggmartin what is the difference between the link I sent you in my last reply and what you found. This is not a matter of who is correct. You can call it what you will they are all the same definitions. I was taught the term Nyquist criteria in engineering school. The link I provided you from Analog devices which is a very large company uses the term criterion. The word theorem can apply as well because what’s your name. I only tried to fill in what Paul could not remember. Nothing more
Recently i decided to build my first non powered speaker system and watched alot of videos.And still do at this moment.for speakers im in 300 euro max range.I look at specs and things like that(and they dont tell the whole story).And so far one brand im considering is Wharfedale specificaly Diamond 11.1 speakers.Found them on sale for 229euros suposedly usualy costing 350.I live in europe.So Any answers will be greatly apriciated.
Hey the only answer will be to find a store that has them on the Demo floor and have a listen to them i personally have bad experience with Wharfedale but some like them also 300 euro wont get you far for that money i would recommend headphones and save up money unless you can find something in the used market. The reason i say go and listen to speakers with you're favourite music is because no one can tell you what is good or bad its you're ears not mine only you know what kind of sound you like one cheap company that is Swedish is Audio Pro their old Ego 515 were the best under 200 euro few years ago and i got 4 of them for my kid for 80 euros on the used market and they sound very similar as Bowers & Wilkins 603 S2 for 2000 euros but they are lacking some detail and clarity
Paul, please make a video talking about the differences between DSD64 and 128 (or even 256). I'm perfectly happy with DSD64 from SACD's, but I was wondering if there's an audible difference in DSD128. Many thanks for your amazing content!
You only obtain two and a bit points on a sound wave cycle when you sample a 20kHz sound wave using 44kHz sampling rate. reproducing a sound frequency at 20kHz using 2 and bit data points is next to useless. The bass or lower frequencies are sampled very well at 44kHz sampling rare. Example: a 44 Hz sound frequency will produced 1000 data points when sampled at 44kHz. So the reproduction of the original frequency will be very accurate and near original. You may need at least 4 data points to reproduce a reasonable sample of the sound. You will also need to assume a sound wave shape like a sine wave or saw tooth wave etc in order to use these 4 data points. 4 data points at a sampling rate of 44hHz implies that the frequency is about 11kHz
Hi Paul, Nice that one of my countrymen has posed a question. You must visit Sri Lanka. Have Colombo in your bucket list. Far far safer security wise besides the pandemic risk. Love your videos, regrettably I dont own any PS Audio products as yet. Maybe Maybe later.
IMHO commercial music was recorded and engineered back then better by people who really had some great talent. The classic gear that was used is all copied today by software plugins and cheap knockoff outboard gear that doesn’t sound the same. Also, 2 inch tape sounds glorious. You can hear the layering on so many tunes and how incredible some of those albums were. The techniques, talent and especially budget is long gone.
Some of the best recordings were in the 50s and earlier. The From The Original Tapes of Bill Haley and Buddy Holly sound magnificent. Ellington stuff from the 30s and 40s are great too. But from the 60s on, more and more processing was done on most of the masterings. And, imho, SQ went downhill on a lot of what was released to the public.
Paul seems to be dazzled by the high DSD sample rates. PCM and dSD cannot be compared in this way, they are completely different digital formats. Bigger is not always better. PCM can deliver perfect waveforms without errors or stair steps. DSD DOES produce stair steps in the output waveform. This is why DSD must use much higher sample rates to the point where these errors become inaudible.
@3:15 Regarding up-sampling: "There's no more information there. That's a total farce. There's nothing more there" That is not true. To be blunt: That is flat out wrong. When you up-sample, you are having a computer calculate the difference between sample #1 and sample #2, and it creates sample #1.5. The created samples are contrived. But they are real. Like having two frames from a movie of a person walking, and having a computer generate an in-between frame that would have been there if the movie was recorded with the extra frame. It is not perfect. It uses math to estimate where every pixel would be, and many will be wrong. But it will be somewhat accurate. A similar mathematical formula is used to create extra audio frames, in-between existing frames, based on comparing frame #1 to frame #2. And this goes on for every frame (or every sample) in the digital audio file. Of course it will not be as good or as accurate as it would have been if recorded at the higher sampling rate to begin with. But it does a fairly good job, depending on the software used to generate the extra samples (and if done on the fly by the CD player, then the engineering of the player's hardware is a factor, too, in terms of processing power). So up-sampling does create real samples. It is based on the adjacent samples, and not on the original analog input. But the extra sample are real. In fact, Paul speaks about the steep filter needed for 44 kHz (he should have said 44.1 kHz), which up-sampling alleviates. Why? Because up-sampling creates more samples, and therefore the filter will be less steep. So the fact that the filter is less steep is evidence that there are more samples. I suspect that Paul is calling "no more information there", based on the extra samples not being created from the original analog sound source.
@@NoEgg4u Imagine 2 microphones in a studio with a band playing. One microphone is connected to a recorder sampling at 16 bit 44.1kHz and the other is sampling at 24 bit 88.2kHz. You cannot just upsample the recording from the 1st microphone and expect the same information as the 2nd. The 2nd microphone captures all the detail in the recording. The 1st microphone throws away 8 bits of resolution and half the time domain in comparison.
I don't think the "steep filter" argument is actually relevant. Even the 1st CD player (Philips Cd100) had 4 times over sampling so that a more "relaxed" analog filter could be used. Nowadays I don't think *any* DAC has a steep analog filter.
Many old recordings can and do sound good. What ruins good recordings is engineers who Brickwall old recordings. Many so-called remasters are ruined by the Loudness Wars. There are good transfers. I find Sony Legacy recordings well mastered.
and this is exactly why people are now desperate to get hold of initial and early CD releases of classic 70s and 80s albums where the master was little more than a 16/44.1 digital output from the analogue 2-track master. At the time, audiophiles scoffed at this and how carelessly record labels had just dumped a recording onto a CD as quickly as possible. Now they have by far the most dynamic range.
surely its almost impossible to capture the feel and emotion of actually being in front of a live band and any attempt will always be a reproduction governed by the equipment used and therefore will always fall short as surely no system yet can reproduce a marshall stack at full tilt, a 10 piece drum kit played with venom, a bass players rig at full chug all while a vocalist gives it his all ? sure you can capture the electronic signals but not the moment and feel it will always be a reproduction of the original and lose something along the way
Even the Live Performance of a band as you describe it, has to be live engineered and mixed to sound any good. Only acoustic music like a Jazz trio or an orchestra in a good hall will sound „raw & pure“ the best. But even there in musicproduction we‘re adding f.e. close-mics and artifical Reverb to make it sound „larger than life“. But in the end youre right - nothing comes to a passioned live perfomance you’re experiencing in a space together with the musicians. There are a lot more things happening / senses triggered, and it comes from human to human. (And probably a microphone, placement, cables, digital mixing desks, the decisions of a soundengineer, amps, PA-speakers, speaker-tuning, roomacoustics etc. 😉)
It's with transferring analog recordings in hi-res for the same reason it's worth transferring 35mm film to 4k. There is information there and it can be a more faithful final rendering.
Ahh, upsampling - probably the reason why the 200 USD universal disc player designed after 2012 ended up sounding like a 5,000 USD Wadia CD player from 1995 just by upgrading the analog output stage of the 21st Century era UDP. 🤔
All of the recording from the 70s and early 80s were recorded to tape, NOT CDs. The quality of professional studio master tape recordings from this era is much, much higher resolution than CDs. And higher than most digital recordings done today. Paul either missed the question... Or is promoting his own DSD DAC and recordings as better than CDs... which they are!
Once I tried saying DSD was cool in a synthesizer group on Facebook, and got totally slammed by a bunch of old farts giving me full dissertations as to why 44.1 16 bit is the end all for digital.
I know that Dark Side of the Moon by Pink Floyd was remixed and remastered for their SACD release. I have it and it is really fantastic. You can watch the TH-cam video they did at the time of the release and they explain what they did for the SACD toward the end. It is worth while IMHO.
The key here is to understand the term (never used of course...) of HI DEFINITION TRANSFER. Older recording i.e analogue tape masters are NOT high resolution because they were not recorded in hi resolution - in the vast majority of cases what are being marketed are hi definition transfers. Nothing wrong with that as it captures everything from the original recording but they are not high resolution. SACD came out in 1999 - I can remember that easily and I'm not even remotely an audiophile or hi-fi expert! DSD cannot be compared with PCM as it uses a different process so banging on and on and on about DSD64 being amazing as it is 64 times CD 44.1 'sample rate' is a totally pointless exercise. Yes, I consider DSD a superior format for audio capture but nothing wrong with 44.1 16 bit PCM if done correctly. Upsampling PCM or remodulating DSD to high bit rates can benefit playback as it allows more gentle filters to be used on playback - you are creating a better signal for the DAC to work with before conversion to analogue.
@@KR1275 Not sure how many download free files nowadays, but streaming audio legitimately is at least dominating over physical media. If you go back to the 70s-90s, you would find a lot of people recording music on cassette tapes...maybe piracy isn't actually worse today.
@@KR1275 most people don't bother if they can just listen everything on Spotify. Maybe there are Hi-res news groups where you can download stuff but streaming is a a legal rip off so most people use that.
@@D1N02 Nothing really beats streaming where you can pick from 60+ million music tracks in FLAC CD or better quality e.g. Amazon Music HD costing a cheap McDonalds meal per month.
@@ThinkingBetter The only physical media doing well is vinyl, which saw a 20% increase in 2020 (according to RIAA), and possibly cassette sales. Anything else is tanking.
It is a complete waste of time any file above 48kHz for two general listening you're not gaining anything there's nothing to gain just wasted space! So what's the point bitrate that's a different matter
As the recording bit rates and depth increase, the ability to capture transients also increases. I have seen several reports where the transients where photographed on an oscilloscope, sowing first the nondigitized trace of the impulse frequency (just a single blip) followed by 44.1/16, then each jump all the way up to DSD64. At 44.1, the fop of the blip was about half the height of the nondigitized, with noticeable ringing. With each jump in bit rate or depth, the amplitude of the blip got closer to the nondigitized, and had less ringing. You could even see visually the improvement jumping from 192/24 to DSD64.
I'm so glad CD's came out in the early 80's. Because we were able to get a lot of good music put on a digital file before they decided to remaster old records to raise the volume and brick them. When I buy older music, I always have to buy the first release they did on CD. Because the later remastered versions usually always sound like crap and distorted.
In the case of one of the first direct to digital recordings Donald Fagen’s The Nightfly.
Roger Nichols was quoted saying they made the decision to record the Nightfly on digital equipment. He said "We booked the Village Recorder in 1981 to cut tracks for Nightfly and decided to try the 3M digital machine. We ran a Studer A-80 24-track analog machine in parallel with the 3M for the test. After the band laid down a take we performed an a-b-c listening test. The analog and digital machines were played back in sync while the band played along live. We could compare the analog machine, the digital machine, and the live band. The closest sound to the live band was the 3M digital machine. We re-aligned the Studer and gave it one more chance. The 3M was the clear winner. We rolled the Studer out into the street, (just kidding) and did the rest of the recording on the 3M 32-track machine. When it came time to mix, we mixed to the 3M 4-track machine."
The original resolution of the 3M recording was 24/50 and best you can get is 24/48. Comparing the 24/48 version to the 16/44 CD version is a good A/B test and cool way to test the resolving power of your system.
Back in 1971 or 1972, I had an open reel tape that was cut from a digital master. It was digitally recorded by a company named, I think, SoundStream. I think they were the first company to digitally record music, at least on the commercial level. I got it from the old Windsor Tape Club back when I was a student at UNC (class of 1973).
A bit of trivia. The 0.1 KHz of the 44.1 KHz is only there because the 16 bit Digital audio frames at 44.1KHz fitted nicely on each scan of the recording heads on a Sony U-Matic video cassette which was at the time a convenient way of transporting the digital master to the CD mastering/pressing plant.
The CD story says it was 44056 initially to fit in video cassette, but Sony insisted later to change it to 44100.
Back in the 70's, I used to buy "half-speed" mastered LPs. As I understood it, the pressing master was cut by turning it at half speed (33 1/3 divided by 2), with the master tape playback running at half speed. These were about twice the price of a normal LP, and they definitely had more detail.
They did, but they often lost low end as bass started to fall off below the low frequency limits of the equipment.
For example, high speed duplicators would lose the high end because the high frequencies fell beyond the capabilities of the duplicator.
I think the thrust of the question was really more about older analog recordings transferred to high resolution digital vs Red Book, more than about upsampling earlier digital recordings.
SACD and DVD-Audio came out in around 2000. In the early days of SACD, probably 95% of the releases were from older analog masters.
But mostly remixed (better than before). That can explain the better sound quality.
@@KR1275 and some of it is also pure Placebo effect remixes and remasters are usually something to avoid in my experience for various reasons.
Sound has been the same for a long time, it all depends on the recording and care that goes into it.
It's interesting to say, however, that the older analog recordings were very high resolution. As far back as the 50s, the recording equipment and techniques were astonishingly good. If you could get uncompressed copies of the masters, you'd think they were among the best modern recordings.
Indeed! The quality of broadcast and recording technology even in the 50s was fantastic.
@@graxjpg Tape hiss from those days was the best ever ;-)
There are definitely a lot of great recordings from decades ago but I've found that the best of analog starts in the 1970s. Which recordings from pre 1970s are astonishingly good? Right now while writing this message I'm listening to a 96kHz 24 bit digital version of the master of Apr 30, 1960 Satchmo Plays King Oliver Louis Amstrong St James Infirmary from Amazon HD Music. It's not bad, but it's not really awesome either. But hey, on a good audio system, music from the 1960s sounds much better than modern music on the average Bluetooth player or earbuds that the majority of today's younger consumers are using.
@@graxjpg In the 70's WBUR had a program every Sunday afternoon called Adventures in Sound. Edgar Vilcher used play tape masters of classical albums he owned and send them to the transmitter using a microwave link instead of the telephone lines that were usually used between the studio and transmitter.
I used to record those on my Revox A77 two track deck and the sound was very good, better than vinyls of that same recording. Now mind you this was on FM which is limited to 15khz and the Revox, while pretty good, was not the best available by a long shot and yet they sounded better to me.
@@dell177 that sounds awesome? You wouldn’t happen to have digitized any of those recordings, have you?
0:44 - Paul, I believe the writer was talking about old 70s and 80s recordings reissued on SACD.
I actually bought quite a number of the very early Sony SACDs (mainly classical) and was amazed at just how good most of them sounded - and still do.
Jazz at the Pawnshop. An audiophile standard, recorded in 1976, by a professional recording engineer who dragged his equipment to a bar in Sweden to record a pretty decent jazz combo. The music is good, imaging is perfect, and the ambience is real with clinking glasses and quiet conversations in the background. The recording was made on analog tape. The vinyl record is a sonic delight. Naxos has re-released it in a 2x (128 bit) DSD format. The DSD makes the pure analog vinyl sound like there is a curtain between the listener and performers. If done properly, there is absolutely merit to upsampled releases, and DSD is the King of formats!
Yes, I will agree, it is pretty "real sounding".
I have it on SACD, and it sounds amazing. It's a good demo for when you want to show off the benefits of SACD/DSD. Without digging it out of my storage boxes, I don't think it was Naxos, and I certainly didn't think it seemed there was a curtain between me and the performers.
@@Kiwi_Col The curtain is there when playing the vinyl. The DSD is superior.
@@johnlebeau5471 Thanks John, I was half asleep when reading that...lol.
One of the best recordings ever. Dang, the klinking of the glasses is so real, you feel as though you are there.
Great explanation. I feel like I have spent half my life trying to explain digital audio to 'audiophiles'. Just one thing to add that once you have recorded in DSD it can be converted and compressed (so long as it is a lossless alogorithm) for storage or transmission but it should be converted back to DSD before conversion back to analog audio. It shoudl be said that the so called '1 bit' digital audio is just another way of encoding the analog signal: it is a system known as Pulse Density Modulation (used to be called 'pulse width' when I was a boy). CD used Pulse Code Modulation (PCM) which sends the bits to a DAC in chunks of 16 at a time. At the time CD came out the frequencies required for pulse density modulation were beyond the capability of the digital electronics (8 MHz was a very fast computer processor back then; now we use processors at 4 Ghz). The trick is you can convert PCM encoded audio to PDM (SACD 1 bit) before converting it to analog. The reason for doing this is to get the benefits of using the much simpler converter electronics. Since the analog signal is actually present in a PDM bit stream the converter is little more than a low pass filter with a very gentle roll off ! If anyone wants to understand how this works look up Fourier Analysis as it relates to square waves (It is actually how the human ear does it !).
Last thing if you record to PCM you do loose some benefit compared to DSD recording but at modern sample rates for PCM the difference is negligible; far greater harm is done if you try to do any processing on the digital data ; it must be converted back to analog for processing then stored as DSD again to preserve the benefits.
Sorry to rant a bit but I can't fit this on a T-Shirt (something to do with data density ?)
Trying to follow the thread of this answer is an exercise in futility... it's partly the fault of the question, which encompasses several topics that can easily sidetrack Paul - which is pretty easy to do. What the questioner really wants to know is whether a 192 kHz digital master taken from an analogue tape recording compared with a 44.1 digital master taken from the same analogue tape is better. There's an assumption that older tape recording techniques are inherently worse. Listen to a high res version of "Rumours" and tell me that that recording is worse than one done today. It sounds better than many recordings being done today on digital equipment.
Paul actually did answer that, stating that although the 192kHz won’t add musical information, yes it will sound better as it allows the filter to be more benign than the brick filter used for 44.1kHz.
@@doowopper1951 You mean a 192 upscale from a 44.1 source?
Analogue has a higher resolution in time as well as in signal value.[1] It also has imperfections like noise that can "hide" sound and ambience, like on these old cassette-tapes that were recorded without noise reduction. I was surprised at how much detail was present below the noise level, after I applied a program that removed that noise.[2] If I can still get this program to work, I wonder what this can do to old studio tape recordings!
[1]
However, it likely also has more dynamic issues like non-linear distortion, compression, slew-rate, tape saturation, speed-issues[ like wow and flutter, and noise or hum. These are audible imperfections, but not always bad. They can sound so good that digital studio workstations have tools to mimic them: "tube sound" or "tape sound". We like it!
[2]
It is rather technical, but if you remove the noise in the time-domain in a sensible way, the resulting sound is much better than removing it in the frequency domain(which most programs do). And yes, I first had to digitize the recording, but the noise level of a tape recording is probably about 12 of the 24 bits, so there is something to gain.
Older recordings and high resolution
ANALOG IS ANALOG
DIGITAL IS DIGITAL
Good Music is Good Music
Just Enjoy the Music !!!
This reminds me of movies as well. Some film restorationists don't believe that there isn't much benefit to 4k or 8k scans of old 35mm films because the optics and film available at the time weren't so high quality. You'll see heavier film grain and optical flaws. However there are some benefits such as medium format such as 70mm or VistaVision, restoration work and archival work. Though some will argue that 35mm film has a resolution equivalent to 8k.
SACDs derived from an analog master tape have the capability to sound far better than CDs provided that there is not an intervening PCM digital step in between.
This is correct. What we find though is that with most popular music recordings from the late-80s onwards is that the 2-track master was digital. Most likely a DAT at 44.1 or 48kHz. There were a few other solutions that might've given a few extra bits in the Digital word-length. What we need for a true HD re-master is for the analogue multitrack tapes to be digitised at say 24/96 or using DSD and a new 2-track master produced. When we get into the 90s, we have more and more multitrack recordings made in the digital domain. A new master can be produced, but there's no more detail to tape to be had from the source recordings. Some of the big 80s artists recorded in a mix of Digital and analogue and their re-masters have been elaborate affairs / labours of love.
I have a few SACD‘s ...but non of them sound really outstanding! Whereas most of my regular CD‘s (with and without HDCD) sound amazing!! It’s the recording not the medium
@@bikemike1118 It's true, they can show up a lot of the creaks and joins in recordings from the 1970s. I'm thinking specifically of the Queen back catalogue that was released in Japan on SACD.
@@bikemike1118 In your case, it might be that the master recordings may not be very good to begin with, or the mastering of the recordings for the SACD are poorly done. SACD is not necessarily about sonic spectaculars (though it can deliver that in spades), it's more about getting closer to the best sound that a particular recording can offer. If you don't mind sharing, which SACD titles were you not satisfied with, and does the CD layer of the SACDs you have (if applicable) sound much different in comparison to the SACD layer?
@@circattle And those Queen SACDs sound magnificent, imho.
Paul, unless I have misunderstood you, I think the brickwall fltter is placed at the upper end of the audio band not above the sampling frequency; that is, for RBCD the filtering has to take place between 20kHz and 22.05kHz and that's why it has to be so steep..As far as early SACDs are concerned, I bought a number of those very early Sony's and as I had many of them already on either CD or LP I was always amazed at how good those SACDs sounded - better, in fact, that many more recent ones!.
So is the DSD format our gateway to hearing original, uncompressed master tape quality? And is the original, uncompressed version embedded in a PCM? Because we aren't having to deal with a physical product anymore, it would just be about the limits of storage, which are relatively unlimited nowadays. It's almost like we finally have as close to a perfect way to record now. It's very exciting.
On the other hand, high resolution versions of analog masters? Wonderful.
If you still have the original master from which records were pressed it will make a difference
many companies in charge of these things dont anymore to them its just a fast cash grab since the majority of the listeners wont even notice the change anyway sadly
I think the point to be taken here is that even DSD - whether on SACD or otherwise - is only as good as what you put on it.
isnt that common knowledge? it can never sound better then when it was recorded so if it was recorded in a bad way or mixed in a bad way no matter if you have a 50 million sound system it will ultimately sound bad
Paul I have an Audio Lab 6000CDT Transfer and an Aune X8 Magic DAC. It sounds great
Again, Paul does not answer the question and bangs on about DSD. The answer is that copying an old recording to another digital version cannot recover detail lost in the original recording. It may be a good copy but it cannot be better.
With respect to Paul, he states this exactly at 3:16.
DSD is intriguing, but I listen to mostly rock. DSD recordings for sale online are jazz & classical.
HDTracks has Beach Boys, B-52s at 24/192 , but there is no info on how these were mastered.
I don't know if they are upsampled or legit remasters for hi res.
It would be nice if there was a site that rates the audio quality for HD files.
oh god yes totally agree
i recently upgraded my DAC and unintentionally got MQA support so i figured hey why the hell not to try Tidal and imported my spotify playlists and was casually listening then a few albums just sounded Terrible it was like my youngest son who is 4 years old had mastered them not sure why or where but something on that "Master / MQA" track was just off
Paul I have a question !! are amp selector switch affect badely the sound ? impedance , etc.
If the original file recorded or copied in the studio is 24 bits and you change it to 16 for cd, and changes it back to 24 bits there is no loss. it just restructures what was there.
"Let it be a signal, limited in band, and its time interval divided in equal parts, so that intervals are obtained such that, each subdivision comprises an interval with period {\ displaystyle T} T seconds, where {\ displaystyle T} T is less than {\ displaystyle f_ {m} / 2} {\ displaystyle f_ {m} / 2}, and if an instant sample is taken arbitrarily from each sub-interval, then the knowledge of the instant amplitude of each sample plus the knowledge of the moments when the sample of each subinterval is taken contains all the information of the original signal. "
The first 24-bit recorders began to be used in the late 1980s, I think. We have many albums that we can be improved to listen to in streamers.
But most (if not all) CD DACs supersample to, say, 8X 44KHz and have done since the 90's. Then the harsh filtering isn't necessary.
Great question that I think about all the time
This didn't actually answer the question I was hoping he would answer - my question is that we know old recordings weren't done digitally - they were done analog and presumably when they reissues a "remastered" SACD version of the same music that was previously released on CD format, they are digging out the old analog recordings and remastering them with modern digital equipment. The question is, how much better should we assume that this is going to be than the 16/44.1 version from 30 years ago? How good were the analog recordings? How much better is the modern digital conversion gear? My guess is that the answer will be "it depends" but I'd still very interested in Paul's thoughts here.
short answer is stay away from remasters. you have a good source one you like the sound of you can dabble in new masters and versions but there is no guarantee that it will be better
Long answer is there is no way to know if they even used the original recording so it wont really matter what they do or call it if they used the original recording and if that recording was good there is a high chance the new version will sound good. but there is many issues with this who did it? we all put our own flavour on things even unintentionally so really no matter what it will be different some for better some for worse.
The digital gear is better then the analogue that's for sure but you are 100% right it all depends on how they do it and who does it and how experienced he or she is
Does the old analog recordings rely on bitrates like digital nowadays?
No because there is no bit rate or sampling. Analog is one continuous wave(or infinite sampling as an audio engineer friend of mine would say).
What about by favorite album Brothers In Arms by the Dire Straits they recorded the album on PCM for CD, but in the 1980 what where they recording at 16 bit 20 bit 24bit what and if they recorded it at 16 bit or 20bit why use DSD?
It's like Boston's first album. Recorded on some of the best analog stuff out in the 70's. Sounds great on CD also. SACD not so much. If I ever wanted a SACD to shine it was that album but was a huge disapointment.
@Jingle Nuts Mine also. Just listened to it front to back a couple hours ago stereo set up testing.
yeah some re-releases on new tech just sounds bad i upgraded my dac a while ago and unintentionally got MQA support so i figured id give tidal a try to see what all the fuss was about and was casually listening and some of my favourite albums sounded like something my 4 year old son mixed together just killed the music and i went to the CD and compared it just to make sure i didn't remember wrong and nope something somewhere got messed up
i love this channel
If the Nyquist theorem specifies that you need a 2f sampling frequency then why does my 50MHz oscilloscope sample at 1GSps?
multiple samples per clock cycle. that has nothing to do with nyquist.
Because they are being generous :^)
Thanks easy street photo but It's everything to do with the Nyquist theorem. A sample rate of 2f is just the bare minimum and more is better. The 50MHz refers to the analog sinewave bandwidth. The newer version of my oscilloscope is still 50MHz analog sinewave bandwidth but with 2GSa/s. Paul is seemingly incorrect about the aliasing problems with higher sample rates - aliasing occurs with insufficient sample rates particularly with modulated signals where incorrect peak voltage readings can occur. (i.e. the sample rate is such that the peak voltages are never sampled.) The only disadvantage of having higher sample rates is the higher amount of storage memory required for a given period but that's no issue in this day and age.
@@PebblesChan rule of thumb with scopes is 5f, not 2f. but your sample rate isn't your bandwidth, you're getting 20 samples a cycle - that has nothing to do with nyquist, your bandwidth, any of that. you just get that many samples. that's not bandwidth.
@@easystreetphoto2401 Apologies for my poor language. English is a second language for me. Let me try to clarify for some practise. The marketed bandwidth specified for an oscilloscope is the bandwidth of the analog front end prior to and independent from the sampling section. Most likely the bandwidth specified is the -3dB for that front end. The same oscilloscope is sold with many different bandwidth analog front ends ranging from 20MHz through to 200MHz - all feeding the same 1 or 2GSa/s sampler. The Nyquist theorem specifies that the sample rate has to be at minimum of 2f or greater of the observed signal frequency. The rule of thumb of 5f for scopes for the display of more complex waveforms than just a constant sinewave where there are higher frequency components present such as but not limited to triangle, saw, pulse and square waves so that the displayed signals are acceptable. This can be demonstrated by displaying a 20MHz square wave of on a 20MHz bandwidth oscilloscope vs a 100MHz bandwidth oscilloscope - The 100 MHz scope will display squarish wave and the 20Mhz scope a distorted representation something akin to a sine wave.
Paul never answered the question , which is yes high resolution files from the 70 and 80's can sound better than red bool CD's especially if they remastered it from the original multitrack tape which had a higher dynamic range than vinyl records could handle . Some recordings even had frequencies above what CD's are capable of and above human hearing but research shows we can somehow perceive them anyway and some high resolution files of music from back then preserves them.
Magnetic tape masters from 40 or 50 years ago will probably be in very poor condition. The only useable masters would be an early digital archive taken when the original tapes were still playable.
@@geoff37s38 Tape stored in an environment with temperature and humidity fluctuations help degrade tape much faster than tape stored in a controlled environment .
@@leekumiega9268 at the very least a 40+ year old tape would have serious HF loss and print through.
@@geoff37s38 I have nearly 40 year old cassettes and if they lost their HF it's not apparent, but they were CrO2 tapes not Fe02 . Anadialog has compared a 50 year old Pink Floyd DSOM cassette to vinyl to CD to SACD and in the show more section he provides High resolution files to compare. There was no hint degradation of HF on the tape, check him out on TH-cam.
@@leekumiega9268 I used to have a Brenell reel to reel tape deck. A really good bit of gear. The version I chose had 3 speeds up to 7.5 IPS but there was a 15 IPS version also. Apparently each of the Beatles owned one. They also made studio recorders. I gave it away many years ago.
Nyquist Criteria... can’t wait for the Zuill Bailey recording to release.
What is the Nyquist Criteria, don't you mean theorem....
@@MrBiggmartin over 50 years ago when I was studying electronics engineering the term Nyquist Criteria AKA Nyquist Criterion is defined just as Paul said. He didn’t remember the second word so I filled it in on my comment..
www.analog.com/media/en/training-seminars/tutorials/MT-002.pdf
Pauls definition is correct, ni doubt there. It's the name that is incorrect. It is called the Nyquist Theorem and was from the thirties I Believe.
@@MrBiggmartin what is the difference between the link I sent you in my last reply and what you found. This is not a matter of who is correct. You can call it what you will they are all the same definitions. I was taught the term Nyquist criteria in engineering school. The link I provided you from Analog devices which is a very large company uses the term criterion. The word theorem can apply as well because what’s your name. I only tried to fill in what Paul could not remember. Nothing more
Recently i decided to build my first non powered speaker system and watched alot of videos.And still do at this moment.for speakers im in 300 euro max range.I look at specs and things like that(and they dont tell the whole story).And so far one brand im considering is Wharfedale specificaly Diamond 11.1 speakers.Found them on sale for 229euros suposedly usualy costing 350.I live in europe.So Any answers will be greatly apriciated.
Hey the only answer will be to find a store that has them on the Demo floor and have a listen to them i personally have bad experience with Wharfedale but some like them also 300 euro wont get you far for that money i would recommend headphones and save up money unless you can find something in the used market.
The reason i say go and listen to speakers with you're favourite music is because no one can tell you what is good or bad its you're ears not mine only you know what kind of sound you like
one cheap company that is Swedish is Audio Pro their old Ego 515 were the best under 200 euro few years ago and i got 4 of them for my kid for 80 euros on the used market and they sound very similar as Bowers & Wilkins 603 S2 for 2000 euros but they are lacking some detail and clarity
Love these videos!
Paul, please make a video talking about the differences between DSD64 and 128 (or even 256). I'm perfectly happy with DSD64 from SACD's, but I was wondering if there's an audible difference in DSD128. Many thanks for your amazing content!
You only obtain two and a bit points on a sound wave cycle when you sample a 20kHz sound wave using 44kHz sampling rate.
reproducing a sound frequency at 20kHz using 2 and bit data points is next to useless.
The bass or lower frequencies are sampled very well at 44kHz sampling rare.
Example: a 44 Hz sound frequency will produced 1000 data points when sampled at 44kHz. So the reproduction of the original frequency will be very accurate and near original.
You may need at least 4 data points to reproduce a reasonable sample of the sound. You will also need to assume a sound wave shape like a sine wave or saw tooth wave etc in order to use these 4 data points.
4 data points at a sampling rate of 44hHz implies that the frequency is about 11kHz
@Petra Kann : Actually we can reproduce 20 kHz frequency with 2 sample points just fine.
Hi Paul, Nice that one of my countrymen has posed a question. You must visit Sri Lanka. Have Colombo in your bucket list. Far far safer security wise besides the pandemic risk.
Love your videos, regrettably I dont own any PS Audio products as yet. Maybe Maybe later.
What country does Paul not want to visit?
Excellent question.
The shit hole country.
North Korea?
@@63MGB1 "Chyna"
@@63MGB1 lol... I bet you he'd say he love to go there LOL
IMHO commercial music was recorded and engineered back then better by people who really had some great talent. The classic gear that was used is all copied today by software plugins and cheap knockoff outboard gear that doesn’t sound the same. Also, 2 inch tape sounds glorious. You can hear the layering on so many tunes and how incredible some of those albums were. The techniques, talent and especially budget is long gone.
Some of the best recordings were in the 50s and earlier. The From The Original Tapes of Bill Haley and Buddy Holly sound magnificent. Ellington stuff from the 30s and 40s are great too. But from the 60s on, more and more processing was done on most of the masterings. And, imho, SQ went downhill on a lot of what was released to the public.
What about DVD-A v SACD?
Paul seems to be dazzled by the high DSD sample rates. PCM and dSD cannot be compared in this way, they are completely different digital formats. Bigger is not always better. PCM can deliver perfect waveforms without errors or stair steps. DSD DOES produce stair steps in the output waveform. This is why DSD must use much higher sample rates to the point where these errors become inaudible.
@3:15 Regarding up-sampling:
"There's no more information there. That's a total farce. There's nothing more there"
That is not true. To be blunt: That is flat out wrong.
When you up-sample, you are having a computer calculate the difference between sample #1 and sample #2, and it creates sample #1.5.
The created samples are contrived. But they are real.
Like having two frames from a movie of a person walking, and having a computer generate an in-between frame that would have been there if the movie was recorded with the extra frame. It is not perfect. It uses math to estimate where every pixel would be, and many will be wrong. But it will be somewhat accurate.
A similar mathematical formula is used to create extra audio frames, in-between existing frames, based on comparing frame #1 to frame #2. And this goes on for every frame (or every sample) in the digital audio file.
Of course it will not be as good or as accurate as it would have been if recorded at the higher sampling rate to begin with. But it does a fairly good job, depending on the software used to generate the extra samples (and if done on the fly by the CD player, then the engineering of the player's hardware is a factor, too, in terms of processing power).
So up-sampling does create real samples. It is based on the adjacent samples, and not on the original analog input. But the extra sample are real.
In fact, Paul speaks about the steep filter needed for 44 kHz (he should have said 44.1 kHz), which up-sampling alleviates. Why?
Because up-sampling creates more samples, and therefore the filter will be less steep. So the fact that the filter is less steep is evidence that there are more samples.
I suspect that Paul is calling "no more information there", based on the extra samples not being created from the original analog sound source.
Interpolation is not detail. As Paul states, it just allows for a simpler anti-aliasing filter.
@@circattle Please define detail.
@@NoEgg4u Imagine 2 microphones in a studio with a band playing. One microphone is connected to a recorder sampling at 16 bit 44.1kHz and the other is sampling at 24 bit 88.2kHz. You cannot just upsample the recording from the 1st microphone and expect the same information as the 2nd. The 2nd microphone captures all the detail in the recording. The 1st microphone throws away 8 bits of resolution and half the time domain in comparison.
@@circattle Everything that you wrote is correct.
Everything that I wrote is correct.
I'm learning SO much from watching PS Audio on TH-cam. Thank you so much for sharing your 'immense' knowledge.
We have an appropriate term in the UK - "You can't polish a turd".
"Putting lipstick on a pig"...........lol
or, "Trying to make a silk purse out of a sows' ear"
True, but you step aside a turd in a more fabulous way ;-)
"aliasing", not anti. Anti is the attempt to lessen the aliasing.
I don't think the "steep filter" argument is actually relevant. Even the 1st CD player (Philips Cd100) had 4 times over sampling so that a more "relaxed" analog filter could be used. Nowadays I don't think *any* DAC has a steep analog filter.
Many old recordings can and do sound good. What ruins good recordings is engineers who Brickwall old recordings. Many so-called remasters are ruined by the Loudness Wars. There are good transfers. I find Sony Legacy recordings well mastered.
and this is exactly why people are now desperate to get hold of initial and early CD releases of classic 70s and 80s albums where the master was little more than a 16/44.1 digital output from the analogue 2-track master. At the time, audiophiles scoffed at this and how carelessly record labels had just dumped a recording onto a CD as quickly as possible. Now they have by far the most dynamic range.
surely its almost impossible to capture the feel and emotion of actually being in front of a live band and any attempt will always be a reproduction governed by the equipment used and therefore will always fall short as surely no system yet can reproduce a marshall stack at full tilt, a 10 piece drum kit played with venom, a bass players rig at full chug all while a vocalist gives it his all ?
sure you can capture the electronic signals but not the moment and feel it will always be a reproduction of the original and lose something along the way
Even the Live Performance of a band as you describe it, has to be live engineered and mixed to sound any good. Only acoustic music like a Jazz trio or an orchestra in a good hall will sound „raw & pure“ the best. But even there in musicproduction we‘re adding f.e. close-mics and artifical Reverb to make it sound „larger than life“.
But in the end youre right - nothing comes to a passioned live perfomance you’re experiencing in a space together with the musicians. There are a lot more things happening / senses triggered, and it comes from human to human.
(And probably a microphone, placement, cables, digital mixing desks, the decisions of a soundengineer, amps, PA-speakers, speaker-tuning, roomacoustics etc. 😉)
Hello again.
It's with transferring analog recordings in hi-res for the same reason it's worth transferring 35mm film to 4k. There is information there and it can be a more faithful final rendering.
Ahh, upsampling - probably the reason why the 200 USD universal disc player designed after 2012 ended up sounding like a 5,000 USD Wadia CD player from 1995 just by upgrading the analog output stage of the 21st Century era UDP. 🤔
Yeah, no it didn't.
All of the recording from the 70s and early 80s were recorded to tape, NOT CDs. The quality of professional studio master tape recordings from this era is much, much higher resolution than CDs. And higher than most digital recordings done today. Paul either missed the question... Or is promoting his own DSD DAC and recordings as better than CDs... which they are!
Not true. Magnetic tape recordings are hugely compromised and nowhere near as good as a good digital recording.
@JoyFLIPS : There is no analog format out there that is higher resolution than CDs. :)
Analog is always better.Always.Period.
@The Normal: No.
Once I tried saying DSD was cool in a synthesizer group on Facebook, and got totally slammed by a bunch of old farts giving me full dissertations as to why 44.1 16 bit is the end all for digital.
Elac photobomb🙂
I know that Dark Side of the Moon by Pink Floyd was remixed and remastered for their SACD release. I have it and it is really fantastic. You can watch the TH-cam video they did at the time of the release and they explain what they did for the SACD toward the end. It is worth while IMHO.
The key here is to understand the term (never used of course...) of HI DEFINITION TRANSFER.
Older recording i.e analogue tape masters are NOT high resolution because they were not recorded in hi resolution - in the vast majority of cases what are being marketed are hi definition transfers. Nothing wrong with that as it captures everything from the original recording but they are not high resolution.
SACD came out in 1999 - I can remember that easily and I'm not even remotely an audiophile or hi-fi expert!
DSD cannot be compared with PCM as it uses a different process so banging on and on and on about DSD64 being amazing as it is 64 times CD 44.1 'sample rate' is a totally pointless exercise. Yes, I consider DSD a superior format for audio capture but nothing wrong with 44.1 16 bit PCM if done correctly.
Upsampling PCM or remodulating DSD to high bit rates can benefit playback as it allows more gentle filters to be used on playback - you are creating a better signal for the DAC to work with before conversion to analogue.
Paul is making old tech great again...SACD was great but the world has moved on to high definition streaming.
the world?? I guess 99,9% just download 'free' files from the net. Maybe you mean the same.
@@KR1275 Not sure how many download free files nowadays, but streaming audio legitimately is at least dominating over physical media. If you go back to the 70s-90s, you would find a lot of people recording music on cassette tapes...maybe piracy isn't actually worse today.
@@KR1275 most people don't bother if they can just listen everything on Spotify. Maybe there are Hi-res news groups where you can download stuff but streaming is a a legal rip off so most people use that.
@@D1N02 Nothing really beats streaming where you can pick from 60+ million music tracks in FLAC CD or better quality e.g. Amazon Music HD costing a cheap McDonalds meal per month.
@@ThinkingBetter The only physical media doing well is vinyl, which saw a 20% increase in 2020 (according to RIAA), and possibly cassette sales. Anything else is tanking.
Damn dirty bit bucketing. 🤨
Good explanation Pop
It is a complete waste of time any file above 48kHz for two general listening you're not gaining anything there's nothing to gain just wasted space! So what's the point bitrate that's a different matter
As the recording bit rates and depth increase, the ability to capture transients also increases. I have seen several reports where the transients where photographed on an oscilloscope, sowing first the nondigitized trace of the impulse frequency (just a single blip) followed by 44.1/16, then each jump all the way up to DSD64. At 44.1, the fop of the blip was about half the height of the nondigitized, with noticeable ringing. With each jump in bit rate or depth, the amplitude of the blip got closer to the nondigitized, and had less ringing. You could even see visually the improvement jumping from 192/24 to DSD64.
64 times is not a lot....don't be silly