Paul makes me laugh; "I'm an analogue guy'! And yet he just happens to sell one of the best DACs ever brought to the market and is about to replace it with an improved version. I guess that's why he converts digital to analogue so well (with a lot of help from Ted).
Good explanation for normal users. In my book, DSD (plus marketing) is, first and foremost, a strong instrument for record companies to stay in the business and release "improved, remastered" versions. However, many of the old, but properly done 16-bit CD masters from the sane days of mastering, with the now-old and specs-wise subpar Sony PCM-1630, actually sound better than modern DSD remasters and have audiophile value. Why? Because a good number of DSD/SACD remasters in fact clip and have lower dynamic range. These old CDs can sound better not because of the technology, but because of the (sane, superior) way it has been used in the 80s and the (insane, ridiculous) way it is used now-that takes sonic precedence. On the other hand, I view PCM as the sensible way to store files, and audio processing, mixing, mastering is done in PCM, with final PCM-DSD conversions (since computational operations are done on words). On the other hand, with modern quality DACs and implementations of filters and oversampling, there is nothing to worry about-no audible difference between PCM and DSD is demonstrable in double-blind tests. With that in mind, there also should be no audible difference when playing std-res vs. hi-res in the vast majority of cases for the vast majority of listeners, except in mastering environments, where there can be small differences in transient response and tonality, particularly with different dithering and noise-shaping algorithms. Finally, we also know that people can hear things that are not there at all, but are only in their mind. Thus, IMO, it is not justified to waste space and bandwidth for DSD files. Record companies know that and they INTENTIONALLY master standard 16-bit files/CDs to sound worse than 24-bit PCM or DSD/SACD versions (more limiting applied, slightly different EQing, etc.). Then many people can pick up that difference and say, "High-res is so much better, I'll buy that from now on, I'll buy high-res versions of all the CD albums I love". And so it goes. Very few try resampling the hi-res source with a good SRC, such as Weiss or maybe MBIT+, and ABX the two with their DACs-many would get a revelation. The vast majority of CD remasters of 70s and 80s albums I have heard are inferior to the original masters done in the 80s and maybe in the early 90s. The irony? In our days, we talk about bit depths and sample rates, DSD and PCM, we have expensive digital workstations, DA/AD, but we use all that technology to compress and limit music to ridiculous levels, within 6-7 dB of dynamic range, wasting all the technological achievements. Take, for example, the band Europe. 14 bits are also fine to play the albums from the last 10 years. Best regards,
I'm so very close to DSD goodness finally! Got a Sony x800 for SACD/DSD transport, just need the DAC & an HDMI 2 IIS/DoP converter box on my X-mas list now. Cant wait : D
It’s still difficult for me to take in all of the information that you try to squeeze in to 5 to 6 minutes. I think that’s why I’m not getting the hang of all of these new acronyms and ‘you are’ going off into into the weeds with your explanations and it starts to get a little muddy doesn’t absorb all in at once. If I were back in engineering school topics like this take at least several days of classes and assignments and consistently work on a course for months before being proficient in understanding everything that’s going on in these complex circuits. Someone should write a book on digital signal processing for dummies. Perhaps you should watch your own video today and see how many different terms you’ve used in this short period of time where your explanation should take hours for a true understanding of this topic.
If you are familiar with computer programming languages and coding languages, that is what Paul is talking about. Software and hardware drivers need language coding from a computer programmer to execute their functions. Just about everything you do with computers, phones, tablets, your audio equipment, even your microwave oven and your car requires some form of language processing code to function. Microprocessors will not work without firmware and drivers. Paul probably has no idea he is sitting on top of a flammable BOMB with that Tesla he drives…LOL THE Only reason Paul safely gets from point A to point B is the microprocessor’s in that Tesla. It also keeps his house from burning to the ground if it is parked in the garage charging..😎👍😀
Mr. Scratch, I'm guessing that Paul's last 2 or 3 sentences might have been sort of what you were looking for. Expounding on that analogy: Let's say that DSD data is oil, and streaming pure DSD is oil flowing through a pipeline. But your computer (or other hardware) doesn't have an oil pipeline port! But it does have the ability to unload oil from oil tanker trucks or train cars [PCM]! So we can instead put the oil (DSD data) into a bunch of oil tanker trucks or oil tanker train cars (PCM frames) to transport the oil that way instead of pipeline. When the oil (DSD stream) comes out, it's the same either way. Maybe that helps.
Thanks for explaining Paul! I didn’t understand the title. But because I like your video’s so much, I clicked on it anyway while feeding my fish. And I actually learned a lot!
If I understand correctly, DoP is a transport protocol for DSD. Similar to Internet Protocol TCP/IP, which stands for Transmission Control Protocol/Internet Protocol. Thank you very much for the detailed explanation.
Can we take the DSD file and index it via a separate track so that way with a special algorithm we can fool the computer into thinking the 16 indexes per second are bits so we can directly manipulate a DSD file? I’m thinking out my butt again but it’s an idea.
just use a dedicated streamer, which is able to put through DSD. That might be the best idea, as computers are bad sounding by nature, let alone the connectors.
Thank you Paul for taking my question. David in Kokomo, Indiana 50 miles north of Indianapolis where Elwood Haynes invented the automobile not Henery Ford. Kokomo also produced the first car radio at Delco Radio.
It's best to urge Intel to renew its HD Audio standard from PCM-based to DSD-based.. To AMD, it's debranded Azalia. So PC can upgrade internally to support native DSD. doing so also upgrads all smart phones which share audio chips with PC. But still processing DSD64 is not as efficient as doing 44.1khz/16bit or 48khz/24bit PCM....I like efficient practices like simple MP3..
DSD is a great method for playback, DAC's sound better with DSD. But to my best of knowledge and belief, the claims that it sounds better are not true. 44khz/16 bit sounds way better than DSD on an average DAC but only on a very good DAC. The dynamic range of CD quality is far exceeding the need and human hearing. As far as standard, the general public couldn't care less. Audiophiles are way better off investing in a really good DAC and enjoy the music, as the availability of DSD 'mainstream' is nihil and will never grow in the near future, as there is way to much data and cost. Which is not needed at all with a good DAC. If PS audio needs DSD, then their DAC's are very poor. (no insult intended)...but as a business model, it's very wise. Consumers pay more for the tracks so their spending is additional to the sales of audio gear. And the claim of better sound even is true.
@Douglas Blake a new standard will gradually introduce native PDM/DSD to PC...just like what the AC97 standard did to internal PCM transfer and sound cards...a gradual replacement to all private audio solutions on PC..
@Douglas Blake No it can fit into memory... 1bit data can be grouped,padded, or trimmed into 8bit unit, aka byte! Intel HDA standard already work at 192khz 32bit(larger scale than DSD64) since 2004 if you adjust PC's transfer setting. Persist store on memory or harddrives do not need to be standardized and wont be issue, that's why we've got dozens of software audio encodings.
@Douglas Blake people have been storing DSD music in .AIFF and .WAV formats for years.. just treat the PDM/DSD data as a 2822kbps x2 = 5645kbps stream...currently PC can handle 192khz/32bit stereo at 12288kbps within HDA public standard. while 44.1khz/16bit stereo are at familiar 1411kbps...
@@edmaster3147 yes. Mr.Paul terms it diminishing return.. I think DSD is a perfect way of documenting the sound waves...for the audio professionals, or recording artists.. but for me one of the music audience 44.1khz/16bit/stereo sounds great..even 48khz/24b is great for music product and production. the nature that DSD documents perfectly could mean it suits more for sonic science than for music art form.. the mixing processes always lose something to the original sounds.
Well, choo choo this was a real train wreck! 2:06"... pulse density modulation is the basis of all a to d converters and all modern DACs, all of them run DSD and they convert it to PCM." This is absolutely not true and even Sigma Delta converters don't convert PCM to DSD. Multi-bit Delta-sigma modulators first integrate a multi-bit input signal which is first passed through an integrator The output multilevel signal from integrator output is divided into 1s and 0s DSF (not all 1s or lack of 1s as in a DSD signal) For decreasing errors in the output signal, it converted back to multibit PCM for comparison to the incoming signal. 2:23 "...at the output so everything runs PDM. In this world of ours, this modern world. Didn't used to be that way it used to be when we had ladder DACs and Ladder A to D converters we went directly to PCM..." This is absolutely not true either! At the output, everything runs analog NOT PDM, and Ladder A to D and D to A converters still exist and there really is no real reason to have to deal with DSD or add in additional conversions from DSD to PCM to edit, back to DSD then to DoP (PCM) back to inferior DSD. The more conversions are done the greater the chances for errors in the conversion processes. 2:48 "so that we're you know routinely running things at 176 kilohertz and 32 bits or 24 bits (PCM) you really need to use Sigma Delta modulation..." This is completely untrue as well because ladder DACs can be and are used with these sampling rates. 3:23 ...basically DSD is a constant stream of ones and zeros never stops it just starts going when we turn the recorder on it starts going and as things get louder the density of those bits nothing the timing never changes if it's running at 2.7 megahertz or 11 megahertz like dsd256. Once again this is incorrect on multiple levels. 1. DSD Just like PCM takes samples so the stream of bits is only recorded at each sample, there is no data collected between samples so the data stream is not constant, 2. While the "sample rate" is constant the timing of each sample both in recording AND playback is dependent on a clock. The clock in turn is dependent on the power supplied to it and IS NOT constant! Differences in power supplied to the recording or the playback clock cause distortion of the signal whether it is DSD or PCM. 4:09 "...quiet okay but the point is it is a continuous stream just like analog, analog as I'm talking is a continuous stream background noise my blabbering blah blah blah it's a continuous stream if you look at it on a scope so is PDM or DSD continuous stream PCM is not a continuous stream, PCM is not a continuous stream... Okay, do I really have to point out that if you are watching this UA-cam video and listening to Paul's continuous analog stream of blabbering blah blah blahs that you ARE NOT listening to it in DSD, but you are listening to as AAC-LC converted to PCM. Go ahead and hook a PCM signal to a scope and you'll get an analog wave just like a DSD signal because both have to be converted to analog. However go ahead and google DSD, PDM, and PCM and take a look at the graphic representation of each. Notice how DSD is represented as groups of squares or rectangles whereas PCM is represented as an analog signal. DSD\PDM IS NOT more analog-like than PCM and is in fact less analog like than PCM.
Paul makes me laugh; "I'm an analogue guy'! And yet he just happens to sell one of the best DACs ever brought to the market and is about to replace it with an improved version. I guess that's why he converts digital to analogue so well (with a lot of help from Ted).
Good explanation for normal users.
In my book, DSD (plus marketing) is, first and foremost, a strong instrument for record companies to stay in the business and release "improved, remastered" versions. However, many of the old, but properly done 16-bit CD masters from the sane days of mastering, with the now-old and specs-wise subpar Sony PCM-1630, actually sound better than modern DSD remasters and have audiophile value. Why? Because a good number of DSD/SACD remasters in fact clip and have lower dynamic range. These old CDs can sound better not because of the technology, but because of the (sane, superior) way it has been used in the 80s and the (insane, ridiculous) way it is used now-that takes sonic precedence. On the other hand, I view PCM as the sensible way to store files, and audio processing, mixing, mastering is done in PCM, with final PCM-DSD conversions (since computational operations are done on words). On the other hand, with modern quality DACs and implementations of filters and oversampling, there is nothing to worry about-no audible difference between PCM and DSD is demonstrable in double-blind tests. With that in mind, there also should be no audible difference when playing std-res vs. hi-res in the vast majority of cases for the vast majority of listeners, except in mastering environments, where there can be small differences in transient response and tonality, particularly with different dithering and noise-shaping algorithms. Finally, we also know that people can hear things that are not there at all, but are only in their mind. Thus, IMO, it is not justified to waste space and bandwidth for DSD files. Record companies know that and they INTENTIONALLY master standard 16-bit files/CDs to sound worse than 24-bit PCM or DSD/SACD versions (more limiting applied, slightly different EQing, etc.). Then many people can pick up that difference and say, "High-res is so much better, I'll buy that from now on, I'll buy high-res versions of all the CD albums I love". And so it goes. Very few try resampling the hi-res source with a good SRC, such as Weiss or maybe MBIT+, and ABX the two with their DACs-many would get a revelation. The vast majority of CD remasters of 70s and 80s albums I have heard are inferior to the original masters done in the 80s and maybe in the early 90s.
The irony? In our days, we talk about bit depths and sample rates, DSD and PCM, we have expensive digital workstations, DA/AD, but we use all that technology to compress and limit music to ridiculous levels, within 6-7 dB of dynamic range, wasting all the technological achievements. Take, for example, the band Europe. 14 bits are also fine to play the albums from the last 10 years.
Best regards,
🤗THANKS PAUL,MUCH MORE PROOF THAT IT IS THE BEST TIME EVER TO BE AN AUDIOPHILE 😁💚💚💚
I'm so very close to DSD goodness finally! Got a Sony x800 for SACD/DSD transport, just need the DAC & an HDMI 2 IIS/DoP converter box on my X-mas list now. Cant wait : D
It’s still difficult for me to take in all of the information that you try to squeeze in to 5 to 6 minutes. I think that’s why I’m not getting the hang of all of these new acronyms and ‘you are’ going off into into the weeds with your explanations and it starts to get a little muddy doesn’t absorb all in at once. If I were back in engineering school topics like this take at least several days of classes and assignments and consistently work on a course for months before being proficient in understanding everything that’s going on in these complex circuits. Someone should write a book on digital signal processing for dummies. Perhaps you should watch your own video today and see how many different terms you’ve used in this short period of time where your explanation should take hours for a true understanding of this topic.
If you are familiar with computer programming languages and coding languages, that is what Paul is talking about. Software and hardware drivers need language coding from a computer programmer to execute their functions.
Just about everything you do with computers, phones, tablets, your audio equipment, even your microwave oven and your car requires some form of language processing code to function. Microprocessors will not work without firmware and drivers.
Paul probably has no idea he is sitting on top of a flammable BOMB with that Tesla he drives…LOL THE Only reason Paul safely gets from point A to point B
is the microprocessor’s in that Tesla.
It also keeps his house from burning to the ground if it is parked in the garage charging..😎👍😀
Mr. Scratch, I'm guessing that Paul's last 2 or 3 sentences might have been sort of what you were looking for. Expounding on that analogy: Let's say that DSD data is oil, and streaming pure DSD is oil flowing through a pipeline. But your computer (or other hardware) doesn't have an oil pipeline port! But it does have the ability to unload oil from oil tanker trucks or train cars [PCM]! So we can instead put the oil (DSD data) into a bunch of oil tanker trucks or oil tanker train cars (PCM frames) to transport the oil that way instead of pipeline. When the oil (DSD stream) comes out, it's the same either way. Maybe that helps.
Digital audio processing is like booze. Makes you feel good for a while but ok in small amount. Analogue audio is like spring water. Pure and simple.
I thought "dop audio" meant when someone is listening to music while they high.
It’s spelled “Dope”. You’re high!
🤨
Thanks for explaining Paul! I didn’t understand the title. But because I like your video’s so much, I clicked on it anyway while feeding my fish. And I actually learned a lot!
If I understand correctly, DoP is a transport protocol for DSD. Similar to Internet Protocol TCP/IP, which stands for Transmission Control Protocol/Internet Protocol. Thank you very much for the detailed explanation.
@Douglas Blake Yes, you are right, but only by transmisson over the network via ethernet. When transmiting from Source to DAC via I2S not.
Can we take the DSD file and index it via a separate track so that way with a special algorithm we can fool the computer into thinking the 16 indexes per second are bits so we can directly manipulate a DSD file? I’m thinking out my butt again but it’s an idea.
just use a dedicated streamer, which is able to put through DSD. That might be the best idea, as computers are bad sounding by nature, let alone the connectors.
Don't think dsd will ever become mainstream at this rate
I think it shouldn't get mainstream at all, and that it is a temporary thing that will eventually turn into a historical artifact.
from what I understand running 5.1 dolby digital over a spdif/toslink cable is similar to this
Cocobamnnna
Bingo! I got it! 😉💡
Thank you Paul for taking my question. David in Kokomo, Indiana 50 miles north of Indianapolis where Elwood Haynes invented the automobile not Henery Ford. Kokomo also produced the first car radio at Delco Radio.
First practical motorcar: Carl Benz, Mannheim, Germany, 1885.
Closest I got to pure DSD was via dop in car Audio
clear as mud.... but it sure sounds good
great explaination
Bing Bada Boom!👍
It's best to urge Intel to renew its HD Audio standard from PCM-based to DSD-based.. To AMD, it's debranded Azalia.
So PC can upgrade internally to support native DSD. doing so also upgrads all smart phones which share audio chips with PC.
But still processing DSD64 is not as efficient as doing 44.1khz/16bit or 48khz/24bit PCM....I like efficient practices like simple MP3..
DSD is a great method for playback, DAC's sound better with DSD. But to my best of knowledge and belief, the claims that it sounds better are not true. 44khz/16 bit sounds way better than DSD on an average DAC but only on a very good DAC. The dynamic range of CD quality is far exceeding the need and human hearing. As far as standard, the general public couldn't care less. Audiophiles are way better off investing in a really good DAC and enjoy the music, as the availability of DSD 'mainstream' is nihil and will never grow in the near future, as there is way to much data and cost. Which is not needed at all with a good DAC. If PS audio needs DSD, then their DAC's are very poor. (no insult intended)...but as a business model, it's very wise. Consumers pay more for the tracks so their spending is additional to the sales of audio gear. And the claim of better sound even is true.
@Douglas Blake a new standard will gradually introduce native PDM/DSD to PC...just like what the AC97 standard did to internal PCM transfer and sound cards...a gradual replacement to all private audio solutions on PC..
@Douglas Blake No it can fit into memory... 1bit data can be grouped,padded, or trimmed into 8bit unit, aka byte! Intel HDA standard already work at 192khz 32bit(larger scale than DSD64) since 2004 if you adjust PC's transfer setting. Persist store on memory or harddrives do not need to be standardized and wont be issue, that's why we've got dozens of software audio encodings.
@Douglas Blake people have been storing DSD music in .AIFF and .WAV formats for years..
just treat the PDM/DSD data as a 2822kbps x2 = 5645kbps stream...currently PC can handle 192khz/32bit stereo at 12288kbps within HDA public standard. while 44.1khz/16bit stereo are at familiar 1411kbps...
@@edmaster3147 yes. Mr.Paul terms it diminishing return.. I think DSD is a perfect way of documenting the sound waves...for the audio professionals, or recording artists.. but for me one of the music audience 44.1khz/16bit/stereo sounds great..even 48khz/24b is great for music product and production. the nature that DSD documents perfectly could mean it suits more for sonic science than for music art form.. the mixing processes always lose something to the original sounds.
More paper....
Do you live in structure made of wood?
Sir, It seems you getting some weight, Need watching out…
Yes, I have gained a few pounds. Thanks. Hard to keep it off.
Can someone tell me the difference between watts and db?
Watts is a specification for power - amps x volts = watts. Db is a specification for sound pressure or level.
Well, choo choo this was a real train wreck!
2:06"... pulse density modulation is the basis of all a to d converters and all modern DACs, all of them run DSD and they convert it to PCM."
This is absolutely not true and even Sigma Delta converters don't convert PCM to DSD.
Multi-bit Delta-sigma modulators first integrate a multi-bit input signal which is first passed through an integrator
The output multilevel signal from integrator output is divided into 1s and 0s DSF (not all 1s or lack of 1s as in a DSD signal)
For decreasing errors in the output signal, it converted back to multibit PCM for comparison to the incoming signal.
2:23 "...at the output so everything runs PDM. In this world of ours, this modern world. Didn't used to be that way it used to be
when we had ladder DACs and Ladder A to D converters we went directly to PCM..."
This is absolutely not true either!
At the output, everything runs analog NOT PDM, and Ladder A to D and D to A converters still exist and there really is no real reason to have to deal with DSD or add in additional conversions from DSD to PCM to edit, back to DSD then to DoP (PCM) back to inferior DSD.
The more conversions are done the greater the chances for errors in the conversion processes.
2:48 "so that we're you know routinely running things at 176 kilohertz and 32 bits or 24 bits (PCM) you really need to use Sigma Delta modulation..."
This is completely untrue as well because ladder DACs can be and are used with these sampling rates.
3:23 ...basically DSD is a constant stream of ones and zeros never stops it just starts going when we turn the recorder on it starts going and as things get louder the density of those bits nothing the timing never changes if it's running at 2.7 megahertz or 11 megahertz like dsd256.
Once again this is incorrect on multiple levels.
1. DSD Just like PCM takes samples so the stream of bits is only recorded at each sample, there is no data collected between samples so the data stream is not constant,
2. While the "sample rate" is constant the timing of each sample both in recording AND playback is dependent on a clock.
The clock in turn is dependent on the power supplied to it and IS NOT constant!
Differences in power supplied to the recording or the playback clock cause distortion of the signal whether it is DSD or PCM.
4:09 "...quiet okay but the point is it is a continuous stream just like analog, analog as I'm talking is a continuous stream background noise my blabbering blah blah blah it's a continuous stream if you look at it on a scope so is PDM or DSD continuous stream PCM is not a continuous stream, PCM is not a continuous stream...
Okay, do I really have to point out that if you are watching this UA-cam video and listening to Paul's continuous analog stream of blabbering blah blah blahs that you ARE NOT listening to it in DSD, but you are listening to as AAC-LC converted to PCM.
Go ahead and hook a PCM signal to a scope and you'll get an analog wave just like a DSD signal because both have to be converted to analog.
However go ahead and google DSD, PDM, and PCM and take a look at the graphic representation of each.
Notice how DSD is represented as groups of squares or rectangles whereas PCM is represented as an analog signal.
DSD\PDM IS NOT more analog-like than PCM and is in fact less analog like than PCM.
wow
@@geddylee501 There are even more inaccuracies as he goes along but I stopped because of the length of the response.
Geddy.... You obviously know your digital, that's for sure. Paul needs to take a class from you.