DSD vs DoP

[@Twinhouse1950]

Hi, my name is Tom and I'm the guy who asked Paul these questions. I don't use internet streaming, but play from a local music server (Mac Mini). Almost my entire music library is music files in PCM format (in addition to a handful of native DSD recordings) and I have been doing a lot of experimentation to find the best way to play them. To my taste, direct playback of PCM sounds a bit "in your face", while conversion to DSD (over the DoP method) reproduces the details of the music in a much more refined way - without losing dynamic range. On the contrary, the music is presented more naturally - even modern recordings where the "loudness" method has been used during mastering. To achieve this I had to modify my Mac Mini, where I have replaced its internal SMPS power supply with an external Linear PS, as well as removing the internal hard drive and its SATA interface. Now I run the Mini from an external SSD (with OS, Audirvana and music library) connected to the Mini via a SATA > USB adapter. In addition, I upgraded the entire chain of network cables to CAT8. The upsampling process from PCM to DSD512 (via DoP) is at the limit of what my Mac Mini model is capable of, as the temporary files to be transported take much longer and the files are huge in size - but it works. I am not good at either the theory or the mathematics behind the PCM or DSD concepts, I have exclusively used my ears while experimenting. My music network is far from the most expensive, but I've made it sound like it was. Mac Mini > 2 pcs EtherREGEN audiophile switches in series > opticalModule > opticalRendu > DDC to DAC via i2S > tube preamp > power amp > electrostatic speakers + REL sub. All connected with very good power cables an interconnects. Voila! Cheers ;-)

[@Douglas_Blake_579]

I'm not much for streaming either. My system is a Windows HTPC, feeding a small DAC and A07 amplifier with Pioneer tower speakers. The HTPC is pretty much an out of the box NUC (Intel) with an external hard disk on USB3. My music collection consists of a mix of transcoded vinyl and downloaded pop music. I also tried DSD and found the added complexity, along with the delays it caused to be abit tedious and I eventually when back to my PCM files stored as high rate MP3 Files. From the purely technical aspect, if you are upsampling PCM to DSD you won't be getting any new information that is not in the original PCM file. You can't copy what isn't there... But have you considered that the major difference between the two is not the format or the transport? It might be in the DAC you are using or even in the DOP encoding process...

[@Twinhouse1950]

@@Douglas_Blake_579 Thank you for your comment! Yes, DoP conversion is a slower and more demanding way of listening to music, especially if you choose DSD512, which is my preference. According to the Denafrips distributor, their DACs process the received audio signal from PCM and DSD in two different ways - which affects the presentation of the music: "DSD is processed with 6bit 32 steps FIR. It's processed differently compared to PCM. In my opinion, DSD soundstage is wider, easy to follow. PCM sounds more focused, more vivid." In addition, I use in Audirvana a high-quality lowpass filter that eliminates all artifacts above 20 kHz that are generated in the DoP conversion. This gives the DAC an easier task in the conversion from digital to analogue. By choosing DoP in Audirvana, I can also choose to limit the transport of information in the audio signal by up to -6dB. This means I can increase the volume on my trusty Conrad Johnson tube preamp to a level where the music sounds better. This is not possible when playing PCM.

[@Douglas_Blake_579]

@@Twinhouse1950 Thanks you for the explanation. Much appreciated. So it would appear, from your answer, it's the two different paths in the DAC that is making at least some of the difference. I suppose that is to be expected and it seems a good thing in your case.

[@davebutler3905]

You still don't get what Paul is saying. DoP is just a way of snuggling DSD out of a computer that rejects thaat format. It shouldn't change the data in any way. It's still uncompressed DSD!

[@Stimpy77777]

The audiophile edict, “because we can & because were crazy...”

[@davebutler3905]

Thanks for explaing. It's like having a conveyor belt that only handles bottles but you want to send tubes of candy. So you put the candy into empty bottles.

[@tonyd9998]

Hi, I have a question. I have a WA 33 tube amp, I have a standalone DAC and I use Roon with their nucleus for my files. Do I still need a streamer component or will Roon nuc handle that?

[@alternativemusicandbeyond6519]

Whats the difference between Sealed Enclosures vs Ported ? This is question I haven't seen on PS Audio. I actually love Sealed Enclosures, one might say to do that correctly the Speaker keeps a 15' Driver, I agree.

[@paulstubbs7678]

Not sure why they bother with this DOP format, it's just adding a lot of baggage to the DSD stream, and extra baggage increases the data size, meaning a bigger load on the computer. It's being said DOP is so the OS can kind of understand it, not sure why it needs to, it's (DSD) is just another binary file that needs the appropriate software loaded to handle it. Like Microsoft Publisher files, no OS understands them, you just have to load Publisher to handle it, same with DSD, load a good player, like JRiver etc. All it needs is a small, one off header, that says how many channels and what the bit rate is, and with probably a touch of meta-data, and that's it, from there on, pure DSD, no 16 bit packets etc.

[@Paulmcgowanpsaudio]

You're correct, Paul but the reason the the OS cannot understand it is for lack of a driver to tell it what to do.

[@JonAnderhub]

You answered your own question. The DSD needs a header, information on the bit rate, and metadata, AKA DOP.

[@paulstubbs7678]

@@JonAnderhub No, I'm talking about having it once at the start/end of the whole file, Paul M is saying it's in every 16bit 'packet' so way more data than I'm talking.

[@paulstubbs7678]

@@Paulmcgowanpsaudio True, however there are heaps of files Windows etc does not understand natively, you just use the right software, rather than adding heaps of extra data to the file trying to make the OS happy, why?, it your serious about your sound your not going to get the OS to play it, so who cares if it's not recognised.

[@sidesup8286]

Lots of initials. There's a room in the courthouse where you can changeactually change your name to initials. Or a one word name like Madonna. Lots of people sitting there in the chairs waiting to be called concerning their name change. It was a long wait. IWhile waiting, I met a Mr. Dumas, a Mr. Trash, a Mr. Phart, a Mr. Dunadoggie, a Mr. Ratfoot, a Mr. Bozoh, a Miss Beaver, a Mr. Hitleer (first initial A is all he would say), a Mr. Trumpe, a Miss Hazzard, and a Mr. SOB. I wished them all well, and I hoped my number would be called calledsoon; but I mistakenly was in the wrong room room. I thought they issued new ID cards.

[@davidfromamerica1871]

Does your keyboard still have a bug or is it something else going on..?? Do a Google search on the problem.

[@deetgeluid]

So isn’t DoP basically a Delta Sigma converter?

[@artyfhartie2269]

Tom could be hearing DPD, PGh, CVp, KFC, DCs or DPZ2.

[@akenney1234]

I much prefer C3PO

[@artyfhartie2269]

@@akenney1234 Help me Obi wan Kenobi, you're my only hope!

[@ikemi1]

That's DoPe man😋

[@NoEgg4u]

@2:26 "Our computers, at Octave Records, are the fastest, biggest, that money can buy, today." Paul, you are allowing your staff to blow smoke up ... You sound like a Bose owner bragging about having the best speakers money can buy. Just the fact that Tom in Bergen Norway has computer gear that can process double your computer's limit should raise a red flag. Are you using AMD EPYC Geona or Bergamo processors? Or, Intel Xeon Platinum processors? Paul, please ask your staff to spell out which CPUs they are using, which GPUs they are using, and which motherboard chipsets they are using. If you do not know the parts that they purchased, then you do not know what they have purchased. It is akin to stating that you purchased amps that have the best parts money can buy, and not know what those parts are. I assure you that there are far, far, far more powerful computers than what your staff purchased, and such computers would yawn at your 32 channels of DSD 256 data . So please find out what they purchased.

[@Douglas_Blake_579]

Tom's machine is processing 1 (one) DSD512 stream ... Paul is talking about simultaneously processing 32 DSD256 streams ...

[@davidfromamerica1871]

@@Douglas_Blake_579 LOL, other people’s computers are faster than your computer Nah Nah Nah Nah. Kindergarten mentality. 🙄

[@housepianist]

Have you thought that perhaps those computers could be the best at this time? And even if they aren't, I'm sure they can accomplish exactly what the engineers want when producing music. I see similar complaints here when Paul talks about how great the recordings are at Octave Records or how really good his products are. He's speaking his own truth about that and his company as a whole, regardless of what others may think so why should that bother you? Your speculations make you sound like you're just looking for discourse just for the heck of it. Blowing smoke? Um, OK, I guess. 🤷‍♂

[@LeonFleisherFan]

Basically you're saying one is processing 16 times the data in real time, but think the one processing only 1/16 real-time data is faster?

[@NoEgg4u]

@@housepianist"Have you thought that perhaps those computers could be the best at this time?" What computers? Do you have information on Paul's computers that the rest of us do not have? "And even if they aren't, I'm sure they can accomplish exactly what the engineers want when producing music." That is not germane to Paul's comments or my comments. Paul claimed that his computers are the fastest, biggest, that money can buy, today. That is Paul spewing BS, while perhaps not realizing that he is spewing BS. Paul knows next to nothing about computers, and yet he made that definitive statement. Paul prides himself as teaching and helping people. But his computer statement is flat out nonsense. There are people that take the word of a channel's host as the gospel. It is proper for people to point out when a host spews BS, whether intentionally or accidentally. In this case, I am assuming the latter, because Paul knows next to nothing about computers, and yet he made such a bold statement as fact. "He's speaking his own truth..." That is garbage language. Would you be supportive of courts swearing in witnesses with "Do you swear to tell your truth, your whole truth, and nothing but your truth"? As soon as you say that it is "your truth", it is BS. Only "the" truth it the truth. I am not interested in your truth, my truth, or anyone else's truth. The only truth that is the truth is "the" truth. "...so why should that bother you?" You came to an imaginary conclusion. Try on this one: Why did my comment trigger you? I have no evidence that my comment triggered you. But I'll just ask that question, to plant it in the mind of viewers. "Your speculations make you sound like you're just looking for discourse just for the heck of it." I am looking for truth and honesty. Your accusation that I am looking for discourse is actually you looking for discourse. "Blowing smoke? Um, OK, I guess." Paul trusts his staff, implicitly. No one should be trusted to that extent. Yes, trust your staff, but validate their claims. Do not micromanage them. But do not let them run wild, either. Find the right balance. Due to Paul's trusting nature, he has someone on his staff that blows computer smoke up ... ... It is Paul's trust that his staff member is abusing, and Paul does not recognize it. Yes, Paul's trust is being abused, because his computers are not the best that money can buy. Not even close. Someone that Paul trusts is lying to Paul. And you are taking issue with having that person exposed. You would rather let that person continue to feed Paul BS computer information, and allow Paul to look bad when he repeats that BS on his channel. People that are IT hardware professionals know that what Paul said is nonsense. Do you think that Paul should continue say that he has the fastest, biggest, computers that money can buy, today? He has said it in other videos. Should he continue saying it, when it is not so? Paul's credibility is on the line. He can have a spotless record, when it comes to audio. But if he then starts spewing nonsense about computers (or anything else), it taints his otherwise spotless reputation as someone that you can trust for truthful, helpful information. My comment was to help Paul. It is better for him to dive in to the computer equipment he claims is the fastest, biggest, that money can buy, today. Paul is wealthy. He has the resources to hire an IT hardware professional consultant (or firm), pay them their fee, and have them evaluate the computer equipment that his staff tells him is the fastest, biggest, that money can buy, today. Then, when Paul learns that there are computers that are 10x more powerful, Paul should hold a meeting with his staff, and let them know that he made regrettable statements on his youtube channel, and that his staff must never feed him such BS, again. If nothing else, his staff will have a renewed respect for Paul, and will not see him as being someone that they can pull the wool over his eyes. Or is it better to have feelings rule the day, instead of facts?

[@Douglas_Blake_579]

In all due respect to Paul and his many accomplishments, I've never understood the fascination with DSD. Even he admits that computers don't recognize it as valid data and they have to wrangle it constantly to use it... Why not just go to high rate PCM and work the whole process in a single format? PCM192 is utterly indistinguishable from analog after you run it through a DAC and it's something computers can handle natively... I'm sorry but I just don't get the fascination...

[@user-tp4hw5oi6e]

Both DSD and PCM converge on the same signal as sample rates increase, but DSD priorities different information first. Increasing the sample rate of PCM enough to cover the same information would work, but data rates would become impractical. Subjectively many listeners like Paul and myself prefer DSD to PCM. This is because different priorities lead to different sound. DSD is not superior because it sounds better than all PCM. Its superior because DSD is more efficient than PCM at a given bitrate. DSD trades off sample depth for sample rate. at high-fi bit-rates this is more effective. DSD having bad support in OS is what causes the need for wrangling. PCM refers both to a data format "pulse code modulation" and to a protocol for sending and receiving PCM data. DSD over PCM uses the PCM protocol but does not use "pulse code modulation". DSD over PCM is just a away to send DSD from an application to a DAC without needing to define a new protocol. This process has no effect on the sound quality. At the begining of the video the flashing lights on the network switch are mentioned. this just has to do with the way DoP uses the PCM protocol, sending lots of smaller packets, while higher bitrate PCM sends the same number but larger packets. (I am not and expert on these protocols but I have observed this behavior in Roon over USB) I think that Paul talking about the "continuous stream" is misleading. PCM and DSD are both continuous streams of numbers. PCM is generally 16 bit numbers while DSD is 1 bit numbers.

[@InsideOfMyOwnMind]

I think he just means that DSD doesn't natively require packetization with packet headers and other overhead and those overhead quality as interruptions in the stream.

[@user-tp4hw5oi6e]

@@CJANegrcTtpX8HwbR2SBDd I believe when Paul says they record in DSD he means that they record at DSD sample rate AKA 64 * 4 * 44.1khz. but still at 32bit sample depth. After selecting a volume level and dithering each 32bit sample it is converted to a 1bit sample. This can be done in real time on the dac chip itself or in post. The single conversion from 32bit to 1bit is "lossy" but the resulting DSD is bit perfect DSD. The advantage of distributing in DSD vs PCM is that DSD files are 16x smaller that PCM files at the same sample rate. The mastering process in DSD really just means that the DSD is always handled at the DSD sample rate.

[@user-tp4hw5oi6e]

@@InsideOfMyOwnMind both PCM and DSD don't require packetization at the native level. PCM confusingly also refers to a protocol for sending "pulse code modulation" data. To send DSD or PCM data the PCM protocol is used which divides the data into packets. DoP describes using PCM protocol with DSD data.

[@Douglas_Blake_579]

@@user-tp4hw5oi6e Okay ... you cannot convert a 32 bit sample to a 1 bit sample no matter how hard you try. In PCM the data is in the bits... each 8, 16, 24, 32 bit sample is a binary number representing the voltage of the signal at the moment the sample was taken. In DSD the data is not in the bits... it is in the number of bits. As amplitude increases so does the density, or number, of 1 bits in the stream. This is not an organized data byte or word... it is an analog encoding scheme that is more akin to FM radio than to computer data. So no they don't converge to the same data, you can't change one directly into the other and, most importantly... computers do not natively know what to do with DSD, because *it is not computer data*

[@Oystein87]

🇧🇻🇧🇻

[@madmeister407]

What the fck, just use vyl, via, tt, pno, pre, and amp, used through spkr. Job done.

[@stimpy1226]

You're still very confusing Paul. TMI in your explanations.