When I was very young, my dad could always turn the cable back on..as well as the electricity in the harder months. Having a gifted electrician in my life was a wonderful thing 😁 He loved tinkering with his stereo components as well
I once lived in an apartment complex where I had an outdoor closet I could lock and inside it we 2 Air conditioning units and 2 50 amp outlets, one for me and one for my neighbor. When I was injured and out of work and my power was shut off I built a 50 amp jumper cable and powered my apt off the next door circuit. The building would leave the power on in vacant units so they could cool them to show them to new renters, about once a month the manager would come inspect the unit checking to see if the Air Condition or a oven was left on, because they could see the bill on a vacant unit. They never suspected me of stealing the power, One guy moved in and then back out in 2 months when he got his first electric bill
Putting into words, I'd say that bit depth represents possible slots for 1's and 0's. A slot (that is a bit) can have a 1 or a 0. This means that the more slots that you have (greater bit depth), the bigger the number, you can fit into those slots. Say you had only three slots (three bits), the lowest number that you could fit is 000, and the highest number is 111. The 111 in binary is equal to the number 7. If we increase the bit depth to 4 bits we could fit a bigger number into the available slots we could go from 0000 to 1111. The number 1111 in binary is equal to the number 15 in decimal. Looking at the three bits vs the four bits, we can see that with three bits, we could represent a half a sine wave with 7 possible numbers (or stair step voltage states.) With four bits, we could represent half a sine wave with 15 possible numbers, that would be shown as a stair step set of voltages. Having more slots, or as we say, greater bit depth, gives us more places to store numbers, and if each number represents a voltage level, with more slots we could have finer levels represented, because of the greater amount of unique numbers. Using our example, if we had three bits, we could have 7 stair steps in our half sine waves. If we had 4 bits, we wouldn't have any higher voltage represented than with the three bits, but we'd have smaller stair steps because we could divide the half sine wave into 15 different levels. So, the slots or bits represent possiblities for numbers, and the more slots that you have, the bigger the number you can have, and thus the finer the stair steps representing a half a sine wave. Where the sampling rate comes is in deciding just how many of these unique binary numbers we are going to use to represent a half a sine wave. Say that we have 8 bits or slots. If we only sample 5 times during a half a sine wave we'd have one sample at zero, one half way up the half sine wave, one at the top of the half sine wave, one half way down from the top of the half sine wave, and one back down at zero. Even though we had slots or bits enough for 256 possible stair steps, we didn't sample at a high enough sample rate to get that many stair step voltages to represent our half a sine wave. So because of our slow sample rate, our stair step voltages were huge, in comparison to what they could have been with a faster sample rate. In summary, slots are bits (bit depth.) Sampling rate is how many unique numbers you can stuff into these slots or bits. Now, I'm going by memory here, and that's always fallable. Someone correct me if I've screwed this up. This is a fun subject, Paul. Thanks !! (PS: for every bit that you add, that is for every potential slot that you add, you increase the resolution by 6 dB between bits. You don't get a higher maximum voltage by adding bits. You get finer stair steps by adding bits. This give you finer resolution, that is more unique number combinations to represent a half a sine wave, and thus clearer sound, up to a point. Also, because a half a sine wave is represented by stair steps, the more bits you have, and the greater the sampling rate to fill these bits with unique numbers, the stair step voltages can have more resolution at lower volume levels, that is voltage levels. This increased resolution gives you more dynamic range. More slots and unique samples to fill the slots, give you more dynamic range. Quantization errors are when the stair steps start to be "blockish" as you go lower and lower in volume, that is lower in voltage, to where there are few enough of these stair steps, that an analog signal can't be replicated accurately. This causes a type of distortion.)
I likely have this wrong, but I come from the visual image realm where I thought sample rate was similar to resolution, so it’s like the number of pixels in an image. Obviously the more pixels (dots) in a square inch makes for a more continuous looking image. The bit-depth pertains to how much information is held within each pixel - 8-bit can hold 256 colors, but 24-bit can do 16.7 million. So I thought similarly that sampling is the number of times a wave is sampled and bit-depth is the QUALITY of the sample, not just the dynamic range. If I have that totally wrong, I’d like to learn so I can hang my head in shame.
It's funny, some of the people teaching others about audio have the sound of the room they're in dominating the entire voiceover. Thanks for caring. What is your setup for recording the audio on these? Love the channel!
Ah sorry Paul. I called you Tom in my SoundBlaster AE 9 comment. Now to this vid. Interesting. I too wandered because there some people that say CDDA is just perfect and why go HD cause you can't hear it anyways. Well I sorta can. What do I mean by sorta? I do hear and apreciate the difference in CDDA 44.1 KHZ 16 bits audio and full Blu-Ray 96 or 192 KHZ 24 bits audio. I even did some tessts with Sound Forge just making different files with all the bit depth and sample rates I could choose. So what I noted doing this is that yes sample rate gives you more air more openness, but meh bit depth does something about clarity as well. Gosh I really wish I had at least a Zoom F6 and of course the Rode NT1-A to finally put that debate to bed. If I had the 833 Sound-Devices than that will really put that to rest once and for all. The 833 replaces the verry much awesome $4000 744-T. Man I love that recorder. Loved it every time my friend Neal Ewers demoed it with the Rode NT1-A mic. Those demos from Neal are what made me want boath things.
Great explanation. I can't like it though because it currently has 808 likes and of course the we can't mess the one of the greatest drum machines of all time :-)
Paul; I digitize my vinyl records using a Tascam DA3000 in 24 bit, 192khz. If I record too quietly (which happened to me a few times), will I lose a part of this 24 bit advantage. If so, should I worry about it or is it negligible? Also, DSD sounds great. The Tascam can record that too. It's way more airy and transparent than 24/192. Problem is; I record a vinyl record and then throw it in Adobe Audition to cut it into tracks, remove pops and clicks (manually) an get the levels correct between my various records. I then export it to 24/192 WAV and put it on my NAS to play it on my streamer. Is there a 'good' way of doing this with DSD recordings? I can of course convert them to 24/192 using Tascam's audio editor and I do notice that it'll still sound slightly better than 'normal' 24/192, but how much is left of this once I'm done editing?
There's no practical way to utilize the 120dB possible with 24 bits and, besides, the best dynamics you can get with vinyl is about 70dB, so you'd got say 40dB to lose if you need. I wouldn't worry about it. Just stay a few comfortable dB away from max and you should be good.
oversampeling would be an option. The Hans Beekhuizen Channel has some in depth pieces on that, Beekhuizen is very knowledgable and has an independent view which makes sense.
In a nutshell PCM uses groups of 0 & 1 to COMPOSE the height of wave form, DSD uses 0 & 1 to CONTROL the direction of wave form (like vinyl records use groove to control the direction of needle that generates wave form). Sample rate could be considered as frame rate of video; more sample rate means more smooth of the motion.
In a delta-sigma DAC, PCM is converted from a 16 or 24 bit keyword to a 1 bit delta modulation (high frequency single bit stream) - just like DSD. All DSD does is save you from doing the entire delta modulation process.
Many audiophiles conflate bits and sample rates used in the recording process with the bits and sample rates in the playback files. This is a mistake. High sample rates and more bits are used in the recording process to enable editing but are pointless for playback. If you actually had a 24 bit recording with a true 144dB dynamic range and the equipment able to handle it (which is impossible) and your quiet listening room had a 30dB background then a peak sound above background would be 174dB which would be lethal. No, I am not exaggerating, this sound level kills. Even a true 16 bit 96dB recording would be unplayable as the quiet passages would be inaudible or the loud passages screamingly loud. Most modern recordings can be delivered with 10 or 12 bits. A full symphony orchestra giving it all it's got (ƒƒƒƒ) peaks at about 104 dB SPL. Let's give the orchestra 105 dB, and 105 dB - 30 dB = only 75 dB real dynamic range. Most recordings are compressed dynamically to a sensible range so the listener is not constantly having to twiddle the volume control. Higher sample rates on playback produces ultrasonic noise and does not improve audio in the audible range.
@@toltecstrings1 The maximum SPL from a loudspeaker is not the same as the dynamic range of a recording. The dynamic range is the difference between the faintest sounds and the loudest peaks. 16 bit/96dB is a huge dynamic range and using this full range may be undesirable on playback as the faintest sounds may be inaudible or the loudest sounds may be ear-splitting.
Piggybacking on this segment, I am hoping that you will eventually address Blu-ray Audio Only as a format in relation to the areas that you are discussing here......
@@worm6820 This is a very impressive format, if anything, just for its storage capabilities. For example: you can put all of Beethoven's Symphonies on one Blu-ray disc. I find it interesting that Paul is pushing DSD, which is been around now for probably close to 20 years while, thus far to the best of my knowledge, ignoring the Blu-ray Audio format. I like the SACD, SHM, DSD formats etc, but if a given title is available in the Blu-ray Audio Only format, it is always my first choice.
For musicians/producers/engineers I highly recommend this video ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE--jCwIsT0X8M.html , for listeners/connoisseurs I suggest this: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-cIQ9IXSUzuM.html
Hello Paul, very interesting video...one (at least) question comes as conseguence: upsampling a 44 to192 bring real benefit in terms of audio quality (if done in proper way, for example with Roon Rock or Audirvana) or it is just another stupid exercise like many others famous in audiofile world? Another point comes then... why some (many) old records, 1950 -1960 and so on, 16/44 sound so much better then the update remastered/upsampled fake-high-res you can find, and every day more common, in Qobuz? I would like to see a short video from you about those points here too, should I send you email about or this post is already enough? 🙂
Though I love DSD, I consider the best answer on the question that sample rate and bit depth only become a factor if and when the quality of the DAC (taking out of consideration the rest of the gear) is good enough. A really good DAC sounds really amazing at 16 bit 44khz. But it comes at a price, such a DAC will well exceed the price of 5000$. On a lower priced DAC, higher sample rates (and bit depth) sound better and surely will DSD, but it is miles away from a really well executed DAC at 16 bit 44khz. I own a Cambridge DAC capable of 32 bit and DSD 256, but even at high rates my AudioNote DAC sounds so much better at 16 bit, which does 24 bit and oversamples 1x at max. I actually downsample DSD to CD quality and enjoy the recording advantages. Cheers. And yes, it sucks being an audiophile when there is little money to spend, but when you own a 100/k system a 250/k system is appealing.
A really top performing DAC is well engineered and doesn't actually need to cost US$5000. Highest performing DAC silicon e.g. AK4499 is the cost of a few burgers and putting it on a PCB with appropriate high quality discrete components and into a box for best possible performance can be done for well under US$200 on the BOM. But you don't pay the BOM cost but the BOM cost with a margin. So why do you end up paying US$5000? It's mostly the same way as some people sell custom made iPhones with diamonds glued on them for insane money (check Falcon Supernova iPhone 6 Pink Diamond, yes, iPhone 6). And personally, I would rather spend US$5000 on a DAC than US$2 million on a Jacob & Co wrist watch even the latter might be a better investment. I own a >US$100K system and don't desire any US$250K system. Rather I've stopped using vinyl discs and even CDs. The convenience of streaming on demand whatever music you want has taken a hold on me. Check my RU-vid channel with a video of the world's biggest audiophile gear collector and also vinyl/CD collector from my trip to his place (Southern China) a few years ago. He owns a famous knife company. His spending on audio is in the 10s of millions of Dollars.
@@ThinkingBetter Where do I buy a DAC for a few hundred of dollars that performs really good, at the level that I mean? I've heard 500$ Dacs sound very enjoyable and more than satisfactory. But thank you for disagreeing, the world would be such a boring place without that.
@@edmaster3147 I work in audio engineering. There is a cost of a DAC related to the chip itself. You want a better chip that usually will cost 10s of Dollars. Then you need a very clean very low noise power supply that doesn't yield noise into the analog stage. That is also several Dollars of components, perhaps 10. Then most importantly, you need an output stage that can provide highest audio fidelity and balanced output. That's easily another US$10 of BOM cost. Other components are needed incl. connectors and others. And of course you need to put it in a nice aluminum case, let's say that cost US$30 on the BOM. DACs are experiencing the semiconductor revolution like your PC does. A US$10 Apple DAC dongle from today is surely not audiophile, even it sounds better than DACs from the 90s called audiophile back then and many DACs of today. Check the table search "Master SINAD (distortion) Comparison Graph for DACs". Of course sound quality of a DAC is not just about simple measurements.
There are many great dacs that cost less than 200usd, toppings, or smsl, the RME ADI dacs that are brilliant costs some 900usd. So not need of 5000usd dacs.
@@GusdeHbg Yeah, there is no justification for a DAC to cost US$5000 from any intelligent component cost perspective. If 95% of the cost goes to the cost of the engineering and it involves some brilliant solution, perhaps it could be worth it, but I have higher confidence that larger teams of highly skilled people can put better solutions together on a modern DAC chip than some dude hacking something into a FPGA in a small company where the company sells a few pieces. And a modern DAC chip doesn't cost much to turn into a complete audiophile DAC.
Anything more than 20 bits makes no actual sense for media distribution. Sample rates above 44.1kHz can gain some better fidelity in the treble using DSD (what DSD is about) or PCM, but the gain is rapidly diminishing as you increase sample rate. 192kHz is like TVs with 8K resolution where your eyes can't actually distinguish the improvements anymore.
Very educational, when listening to music and I see 24 bit 96hz and then 24 bit 192 and then 16 but 44.1 I'm like whats the difference? To my ears 24 bit sounds better but that's just what I think. I always wondered just what the difference is.
For every bit you get a 6dB dynamic range increase. CD quality is 16 bits and 16x6 = 96dB of maximum dynamic range. A vinyl disc playback is only around 70dB or 26dB less than CD. That's a huge improvement (noise is practically gone on a CD when it's digitally mastered). 24 bits means 24x16 = 144dB. This is a meaningless dynamic range for any music and no actual audio system can support such range. An awesome pre + power amplifier might do 120dB which equates to 120/6 = 20 bits. Thus you waste at least 4 bits when streaming a 24 bit PCM file. More realistically you waste near all the bits from 16 bits to 24 bits in an actual system. Thus, what would make more sense is 96kHz 20 bits PCM, but we use 24 bits because it equates to 3 bytes of data and data is organized in bytes.
@@ThinkingBetter so the 96hz or 192hz is more important? It's that what I'm reading here and the 24 bit doesn't mean much or at all? So when I'm listening to Amazon music I should be looking at 44.1, 96hz and 192hz? That means more for sound quality?
@@JC-bl9bo Yes, I care more about higher sample rate, but it is not a simple topic. When you play Amazon HD Music you can get Ultra HD representing anything more than 44.1kHz 16 bits. I use this service but you ideally want to stream such that your DAC runs at the same clock as your audio stream. Unfortunately, if your DAC runs at a fixed clock, let's say 192kHz, you end up doing sample rate conversion for any music other than 192kHz PCM, which is lossy, unless you manually change the DAC sample rate to match the track played to feed the data bit-perfect to the DAC. For example, if you play 44.1kHz track, you really want your DAC to run also at 44.1kHz, but that might not even be possible in many devices (e.g. phones running at fixed 48kHz).
@Douglas Blake Sample rate conversion can involve more or less precise math and what is best for audio is most heavy for the CPU. For example Android AOSP provides this as an option and if you make an Android device you can build your Android OS to sound more or less good depending on choice of sample rate conversion algorithm. Standard options include linear, cubic and sinc with some choice of coefficients. The best is sinc but a device maker can make his own algorithm also. Sample rate conversion is the hidden issue with PCM playback that nobody seems to talk about.
@Douglas Blake Alright, almost nobody talks about this sample rate conversion topic. Possibly near 100% of consumers are listening to 44.1kHz music on 48kHz DACs on their smart phones and nobody whines about it even the algorithm is crappy for the reason of saving CPU cycles delivering some great gaming FPS scores or other performance scores.
Paul keeps pushing DSD music files but they are very few sites on the web selling DSD . I have tested them on my system and DSD is by far better than anything I have ever heard . The cost is prohibitive with many Albums costing around $40 to $50 dollars in the USA so im not sure why DSD is being pushed so much , particularly when streaming is taking over most peoples music.
24bit audio has virtually no noise floor so the sound is incredibly pure and dynamics don't feel constrained but free and more expressive. High frequency sampling rate has more to do with a more natural and precise reproduction due to moving out of the audible band the intervention of output filter. Mqa is even better in this regard, thanks of its reconstruction algorithm of the analog waveform, in fact its sound is incredibly real (when fully decoded on dedicated hardware).
Two topics: 1) The discussion skirts what I consider to be top priority for any music on any medium in any format: Did the studio do a great job, or a so-so job, or a lousy job? A crummy mp3 file done expertly in the studio will sound better than a 192 kHz, 24 bit file that the studio screwed up (and the studios, almost without exception, screw up). Would you rather watch a fantastically recorded show on an old fashioned tube-type TV? Or would you rather watch an out-of-focus show on a 4K TV? Ideally, you would want to have properly recorded song in a file with both a high sampling rate and high bit depth. For hit songs, such sonic gems do not exist. My best sounding digital hit songs are red book / CD quality (44.1 kHz / 16 bit), and I put forth a serious effort to hunt down the best sounding versions of hit songs. The record companies have countless badly mixed and badly mastered songs that they simply re-release at higher sampling rates and higher bit depths. All that does is allow you to hear more preciously how badly they screwed up the mixing and the mastering. ----- 2) @5:29 "...and it's very much like analog (referring to DSD)". I respectfully disagree. DSD is 100% digital. It is a file that is stored on a hard drive (or solid state drive, etc). It is no different than any other computer file, because 100% of computer files are comprised of bits, and those bits make up the "0's" and "1's" that comprise every file on your computer. A computer cannot store analog data on its hard drives. Let's put it this way: Any DSD file can be stored on your thumb drive (flash drive). That is because it is just another digital file. Paul can correct me if I am not accurately remembering him from one of his other videos... but he explained that on a scope, DSD looks just like analog. That does not make it analog. DSD files are still zeros and ones, as is every file on your computer. @5:41 "We stick very much to DSD, because it's so much better analog sounding than PCM" PCM goes toe-to-toe with DSD, when the PCM file is properly created. Paul is using an analog-to-DSD converter that is fantastic at taking sound, in real time, and converting it to DSD. Paul is *not* using an analog-to-PCM converter that is of the same quality in converting sound in real time to PCM. Perhaps such a high-end analog-to-PCM converter does not exist? Who knows? But the PCM format is not to blame for equipment that does not properly create a quality PCM file. Like having an Excel spreadsheet: One is created by an expert, and has all kinds of charts, macros, color coding, layouts, formulas, cross referencing, that clearly conveys the data it contains, and... one is created by a novice that has a mishmash of data entered that is hard to understand. Excel is not to blame for the latter. DSD might really be better sounding than PCM. But until someone can provide a list of real time, high-end analog-to-PCM converters that were tested against real time, high-end analog-to-DSD converters, we will not know the answer. Paul has concluded that DSD is the clear winner. Yet, he has not auditioned PCM converters as I described above. He has used only one, and that does not qualify as being representative of all analog-to-PCM, real-time converters. Like those claiming that interconnects make no difference in sound quality, when they never did listening tests with quality interconnects, Paul is guilty of doing the same with real-time analog-to-digital converters, by drawing a conclusion on equipment he has not heard. Paul one time believed that a direct connection from a CD player to amps would be better than going through a pre-amp. He changed his mind when someone convinced him to plug his CD player into the Aesthetix Calypso pre-amp. He changed his mind because he did a listening test. Yet, in the case of high-end, real-time, analog-to-PCM converters, Paul has not done proper listening tests (he did tests with only one real-time converter box), and still makes conclusions. His conclusion might be correct. But he is guessing, because he never tested with other real-time converter boxes. Not good. Cheers!
@@steven3557 I do not think that Paul set out to misinform anyone. He heard a live band record to both PCM and to DSD, and the DSD was the clear winner. He is telling the truth, based on his test with that specific equipment. The problem is that since Paul was so impressed with the DSD test, that he went all-in on DSD, and did not test other analog-to-digital converters. He heard one, that happened to be DSD, and he was sold. Had Paul listened to real-time, high-end digital-to-PCM converters, he might have been blown away by them, too. In fact, if he had done the latter, then he might have been sold on PCM, and not bothered with DSD. But we will never know, because Paul will not audition high-end, analog-to-PCM converters. By the way, your CD player has a DAC. All CD players have built-in DACs.
FIRST OFF 😂. CAPS 2 C, NOT SHOUTING… IM 65 EASIER TO SEE UPPER CASE. BIT RATE AND SAMPLE RATE ARE THE BASIS OF DIGITAL SOUND, BUT NOT THE END RESULT.. IF YOU HAVE 24 BIT 192 HTZ UR NOT GOING TO HEAR THE DIFFERENCE WITH AN IPNONE OR IPAD, OR CHEAP EAR PHONES… BUT, IF U HAVE A RECEIVER WITH GREAT DACS, ATMOS OR AURO-3D AND SPEAKERS TO MATCH, THEN U CAN HEAR THE DIFFERENCE…
Hmmm at least to my ears 24bit is alot more Grainular and sensative to fine transients then whatever sample rate increase does. Please let me know how changing sample rate of the recording affects sound.
OK, let's not "judge", just state the facts. Paul said when he was poor he got around utility lockouts. Just the facts, he stole services. Now the real question, when he became wealthy did he voluntarily contact the services and offer restitution? That would have been good. If not, does everyone condone what he did? Is everyone saying "It is OK to steal if you are in a tight spot?" If so there are lots of people who stole food or clothes and got busted for it. What is the right balance? Is Paul excused just because he is an audiophile? Or just because he is CEO of a company? Does everyone practice "conditional morality?"
Great video Paul. I hope you could answer these questions. I've heard that most of the expensive AV receivers actually work at 48kHz internally when processing room correction and bass management etc. and not 96kHz or 192kHz because that would demand much more processing power and DSP, and those rates are way beyond capable of doing these things. Then after processing it goes through 192 kHz DAC (the advertised manageable sample rates - if so chosen on say a AUDIO DVD or Blu-ray). You, as an engineer, is this true or false? How do we know that we actually are listening to 24-bit 96kHz/192kHz music, or if we are being fooled? Asking all the way from Scandinavia. I want to know if there really is a difference between a cheap AV receiver's pre-outs phono outs vs the more expensive pre-outs on a pre-pro balanced outs (say XLRs)? I have to ask all this to get the thrust from this community to know that I get what I pay for. Thanks!
