USB Audio vs. I2S

USB Audio is the easiest and most common method of connecting a server or computer to an outboard DAC. What about this "new" format, I2S?
Have a question you want to ask Paul? Go to www.psaudio.com/ask-paul/

[@FrankyRedEyes]

Sometimes I watch these videos not knowing what Paul is talking about. This is one of those videos.

[@BlankBrain]

I2S (Inter-IC Sound) is an inter-chip standard that has been co-opted to transport audio data between devices. In fact, LVDS (Low-voltage differential signaling) is commonly used to actually transport the signals. HDMI cables provide enough twisted pairs and shielding to pass the I2S signals. One of the really cool aspects of using I2S is that the master clock can be transmitted or received to synchronize everything.

[@rightslot1]

Hello Paul, This comment is really more related to yesterday's thoughts and explanations than anything. I really like how you are still enchanted about the MAGIC that audio music can bring. 'Cause that is what it really is all about. Sometimes we don't need to know every stinking tidbit. {sometimes...} I had some friends back in college----Engineering (Colorado-Boulder). Sometimes these guys got so wound in the technical and scientific side, they didn't enjoy just a good long night of listening. YOU seem to have not lost that. Great Channel!

[@Paulmcgowanpsaudio]

Thanks so much. This comment made my day.

[@BlackFighter_95]

USB = No Audio Quality influence of the USB source as the transmission is digital and block wise (e.g. every 1ms all data gets transmitted for the next 1ms). I2S = Digital data Transmission every sample. The Clock Jitter of the I2S source has an effect on the audio quality. Think of a digital voltage, but the time is "analog". (Even at 192kHz, inter modulation products can be audible) => The more devices are in series, the worse a I2S signal gets. USB does not degrade the signal. Both I2S and USB conduct digital or mains interference from the source to the receiver. A well designed receiver (rarely the case) does not suffer from such interference. So far USB looks better, BUT as I2S is only used for audio and USB is mostly designed as cost efficient as possible, the interference over USB is often far worse. Chances are high that the supply voltage of the I2S source is more stable than the supply voltage of the USB. As a result: With a good receiver you can use both, but preferably USB. On a bad or average (non-HIFI) receiver, you should try I2S. NEVER feed I2S to a HIFI DAC if you can use USB.

[@andydelle4509]

Good point on the grounding issue. OK so I2S could be transformer or optically coupled. But so can USB or SPDIF. In fact any and electrical noise isolation technology could be applied equally to all three interfaces. Electrical current is electrical current, which all three utilize. Let's fix the real problem here. Design good USB and SPDIF interfaces.

[@adaboy4z]

If my DAC has I2S should I use that to connect to my amplifier if it has I2S instead of XLR? I feed my DAC via USB from my laptop.

[@isaacsykes3]

Hey Paul can you please provide a link to the CD cases...thank you!

[@QoraxAudio]

The funny thing is, even cheap CD players use IsquaredS as interface between the transport and internal DAC.

[@ThinkingBetter]

Qorax I2S is from the 1980s done by Philips and has been used endlessly to move audio between parts inside CD players, TVs, receivers and other audio gear. There is nothing special about it other than it’s a synchronous serial data bus. Chip makers (in this case Philips Semiconductors), like to follow standards to make chips interoperable. I2S, I2C, SPI, MIPI and so on are all about this objective.

[@QoraxAudio]

@@ThinkingBetter Yes I have used I2C for some home projects in the past to communicate between a microcontroller and things like little LCD displays. However, embedded systems isn't really my cuppa tea tbh. I didn't know that I2S was specifically a thing by Philips though.

[@ThinkingBetter]

@@QoraxAudio I managed engineering of audio products since the early 90s. Factually, approx. 20 millions of products have shipped that I managed R&D of involving I2S. I have never for a moment considered I2S to be anything more exiting than I2C, SPI, UART or any other such low-level data communication interfaces. It works for what it is intended (move data from A to B) with very little complexity. Currently I'm professionally working on a system involving I2S btw. (via Analog Devices A2B). This topic is a bit tough cause I don't want to speak against Paul, and he obviously promotes I2S as interconnect (a feature of his products) and yes, I2S isn't bad for audio. But, to say I2S sounds better than USB (when used properly) is not a valid statement however.

[@mikelm222]

Hola May DAC ( The best ever measured by Stereophile ) have both HDMI I2S & USB inputs - everyone agrees that USB sounds best - just saying :-)

[@skipgordon5382]

How to connect a MacMini if my DAC has a i2S input? Is it advisable to go USB to i2S or use the HDMI output?

[@andydelle4509]

Paul, you fail to mention the major engineering pitfall of high speed parallel interfaces. Timing Skew. I2S is a parallel interface. You have four signals that all have to stay within a relative timing window. The fact that the data signal in I2S is serial does not relieve this requirement. The master clock is typically 24mhz and requires good rise and fall times to meet the timing skew window. Then you have the issue of switching. I2S needs four layers of crosspoints with a minimum 50mhz bandwidth to carry the master clock. That's expensive if say you want to offer three I2S Inputs on a DAC. There is no free lunch and I2S as an external interconnect interface is not the panacea many audio reviewers claim. Yes, you do reduce the jitter inherent in the SPDIF/AES Manchester encoding, but you now induce new and different problems which can have equally detrimental sonic effects. Telecommunication engineers understand this well.

[@Real_Tim_S]

@Danger Zone It doesn't matter. ASIO only helps latency, not jitter - and it's a software stack not a data transport. Latency is only a factor if you are doing mixing+effects with a tight Analog-to-Digital-to-Mixing/Processing-to-Digital-to-Analog path IO time. If you ARE doing mixing and effects on a PC, then what you need is an ADC/DAC that time-stamps samples and sends that clocking data to the PC (and yes, this can be done easily with USB!!!), so that the input and output streams occur in a deterministic manner. As long as the mixing software and the ADC/DAC understand each other's data pipes and the buffers are used minimalistically and never starved for sample data and always aligned, you won't ever have an audible problem.

[@dillonsaudio]

I'm not sticking up for the guy because I take anything personally, but I believe his explanation was sound! get it? 24 bits of dynamic range. You're looking for 64 bits I suppose. Too much.

[@1over137]

@@dillonsaudio It's interesting when you do the maths on audio word width and what it really means when connected to the analogue domain. The bit step frequency component is already well beyond what the vast majority of analouge front ends will handle. In fact most of it will be lost in the capacitance of the amplifier input lines. Meaning in emperical terms there is zero detectable different on an oscilloscope between samples 12, 13, 14 and a slower lower resolution step of 12, ..., 14. The capacitance and sub frequency components will just not exist in the output. You might argue that over large steps at higher frequencies the output will hit the high resolution point more accurately, but that's also non-sense as the digital domain can go from 0 to MAX_INT instantly, in a single sample, but no analogue audio system in history can move a speaker from one throw to the other in a time scale of microseconds, let alone nano seconds. Not even the speaker red shifting at the speed of light will match the digital frequency domain. A 1 nano second rise time in digital domain has a frequency component of 1Ghz. So. If you want to stop wasting money, you need to look at ALL your path points, from analogue in to analogue out and ask where you slowest freqency domain is, the weakest link and THAT is your fastest, highest resolution audio. Any more will just be averaged out in the analogue domain. There ARE other reasons to go for 24bit or even 32bit, which is my intended internal word size in a project, these only really exist in the digital domain though. Word alignment for memory, CPU/DSP registers etc. At various times you have to pick your poison in your word length. "Clocking" 8 bits bytes is simple from a programming point of view, but is inefficient on a 32bit fixed word legnth processor like an ARM core. It's also confusing having to keep check on your apparent versus actual bit width in buffering alignments, frame stuff/drops etc. Fixing to one internal bus format simplifies most of the coding. It does as you point out increase cost. However if you can solder and read a datasheet, you will find that the different between the BOM (Bill of materials) a PCB print run and some soldering, your audio box might cost you £100 to make. Such a product would sell on the market for £1000+. There is much merit in this approach. If you can code, do electronics etc.

[@1over137]

You may also note, that while I am using an internal 32bit architecture for SPI/I2S almost all of my analouge and USB front ends will be at 16bit 48K. My ears are too old and too badly damaged by heavy rock and heavy techno that I will not notice any more. Besides. USB front ends that do more than 48K16bit are costly and usually involve dedicated audio processor ICs and drivers. Chips like the PCM2707 if you can find them, simplify that process incredibly, including the reclocking off teh async USB interface.

[@8-Bit-Retro]

Off topic, love the artwork - the one on the right is reminiscent of the rear cover of the SLF (Stiff Little Fingers) album 'Nobody's Heroes'!

[@NeilDSouza7]

Luigi : I2S = Spongebob Square Pants ........ USB = Squidward

[@stevefick3919]

"SQUIDWARD!" LOL! Love that show!

[@NeilDSouza7]

I saw a Freddie Hubbard Poster in the Cupboard

[@PebblesChan]

I2S is a bidirectional digital signalling interface between a microprocessor or equivalent like an FPGA to an audio CODEC integrated circuit. In its standard form it is for very short distances like the iPhone or Raspberry Pi to its analog headphone CODEC. PS Audio use it for inter audio component communications. However in many cases, the data is clocked to CODECs as numeric values like PCM. Data across I2S can be specified and interpreted anyway one wants. You can use the I2S interface can be used to send PDM (pulse code modulation) but to ensure accuracy the master clock must have low jitter. USB is multi peripheral interface protocol and associated physical media that is like a small network. Data is encapsulated with metadata which has to be stripped off to read it.

[@net_news]

I2S is more similar to the old RS232 serial connections... otoh USB is like ethernet (with Power over Ethernet and the kitchen sink included haha). Audiophiles advocating for USB is like libertarians advocating for more taxes. Completely non sense... USB is the opposite of simplicity and cleanness.

[@ThinkingBetter]

@@net_news Audiophiles advocating for USB not being inferior to I2S understand how things work. Just avoid using your computer's power via a USB cable to power your DAC. That 5V in a USB cable is usually not very clean. And also, make sure your components are properly grounded so common mode noise is not migrating into your analog audio circuits. And don't use a DIY large PC with a ton of noisy fans. Fan noise is one of the worst enemies for a good system. Personally I like Mac Mini to drive my systems. And frankly, Mac OS is way more stable and unproblematic than Windows. Complexity is much higher with USB than I2S, but your DAC doesn't care if it can get data bit-perfect and doesn't need to dance around with a PLL compensating for some external clock master. One beauty of USB is that it is asynchronous allowing the DAC to precisely control the clock (no audible jitter).

[@ThinkingBetter]

ThinkLearnSolve Right, I2S is one-way. Good to see there are others who aren’t just imagining things here. Of course USB can do excellent audio and with a ton more capabilities than I2S. This discussion is utter nonsense and it’s not rocket science to isolate an analog output stage of a DAC to avoid whatever noise USB might carry.

[@PebblesChan]

Contrary to what you believe, I2S is bidirectional using dedicated data signals for transmission and reception. It is used in situations where the CODEC provides both ADC and DAC functions such as a microphone and a headphone jack. The signal names are BCLK, LRCLK, SDOUT and SDIN. As for I2S being similar to RS232 connections, it really isn't because most RS232 implementations are asynchronous and do not use a synchronizing clock signal. I2S is similar to the SPI (Serial Peripheral Interface) without the Chip Select signals and the MOSI and MISO equivalent to the I2S's SDOUT and SDIN signals. USB's communication rates vary depending upon the version being used. 1.0 is 1.5Mbps (half-duplex) 1.1 and 2.0 is 12 Mbps (half duplex) except 2.0 can negotiate to enter 480Mbps (High-Speed) with compatible devices. Since half duplex communications is used, the effective data transfer rate is slower due to the datagram, protocol, and data direction switching overheads. It wasn't until 3.0/SuperSpeed was introduced that full duplex communications (using its additional pins) at higher speeds such as 5, 10 and 20 Gbps was accomplished to eliminate the overhead of switching the data direction flow.

[@ThinkingBetter]

PebblesChan I2S is one way in its core specification. You can of course add as many channels of audio data and in whatever direction you want just clocked in parallel. Actually I work on a project doing exactly this. I2S is nothing special and I’ve done projects with it since early 90s shipping in millions.

[@ThinkingBetter]

USB is a flexible TWO WAY data connection that can do everything data and allows a DAC to PULL data ASYNCHRONOUSLY with its clock in full control (no audible jitter and bit-perfect audio is possible). The computer can identify a device capabilities through USB. An I2S bus is very low level and is SYNCHRONOUS involving a defined clock master. One doesn't sound better than the other one if the audio specs are the same because you can assume that every 1 and 0 come through reliably. My usual concern with USB is any power or digital noise from a PC leaking through to an analog signal. But this is an issue of good design and not USB itself. Think of it like this, your DAC is happy if it gets every bit correctly, it can run the timing in full control (avoid jitter) and there is no EMI or power rail carried noise disturbing the analog output stage. USB is not a limiting factor.

[@ford1546]

Of experience USB audio can cause a lot of noise! I've experienced it! Remember that USB has 5V and that a desktop computer has a lot of noise! The best thing is to use digital coax or optical from your computer to DAC! Have experienced noise with Cambridge dacmagic plus!

[@arthurott4561]

@@ford1546 The best thing is not to have a computer in your audio system.

[@ford1546]

@@arthurott4561 Yes I know! But running music from a computer to DAC does not produce as much noise as many audio lovers think. If you use DIGITAL coax or opical from your computer you get good sound! Only high end people who apsoutt should have the best who notice something microscopic!

[@Finn-McCool]

HAPZ1ES and forget it. 😉

[@net_news]

Yes, USB is "flexible" and that's why is so complicated, it tries to do everything (and they keep adding features to it!!!). IMHO for audiophile purposes, simpler protocols like I2S are much better, no drivers, no complex controllers, no power, no fuss.

[@larsv6144]

What about that hyped USB implementation in the Holo Audio May and Spring KTE? Is that close to i2s?

[@nl_2652]

I hadn’t heard of I2S before today! Thanks for this video. Hope you all are still healthy and jammin to some good tunes!

[@beslemeto]

How are you using it?From PC to converter and then to a DAC?

[@bigjt37]

Awesome, Paul. Really look up to you. Leaned a bunch from you. Thanks for the info everyday. 💯

[@dandonna852]

How many PS. audio engineers does it take to put a cd rack together?

[@Spock105]

Where can you get these Blue Note posters ?

[@noahstonemusic]

Liked for the Freddie Hubbard poster

[@jay13200]

I do have a question Paul, between I2s or Fiber optic using SFP Port or Dante is the best or lowest noise ?

[@wilcalint]

Paul needs to go back into his lab and do some new testing. Released in 1986, 34 years ago, an I2S data stream can carry one or two channels of data with a typical bit clock rate between 512 kHz, for an 8 kHz sampling rate, and 12.288 MHz, for a 192 kHz sampling rate. USB 3.1 Gen 1 (released in July 2013) is capable of data transfer speeds up to 5Gbps. You can't buy a computer these days without a USB 3.1 interface on it. Gen 2 will be capable of 10Gbps. So that makes USB 3.1 Gen 1 at least 500x faster and 27 years newer then I2S. I think Paul, in this video is refering to USB 2.x or older. There is no specific i2S connecton on a Raspberry Pi, but using software, and the pin header, you can create what is called a "HiFiBerry". Paul mentioning this makes me wonder if PS Audio has a project using the Raspberry Pi. I use a Pi running KODI for my Multimedia player including HiDef Audio. On my Fedora boxes I use VLC. www.tomshardware.com/news/difference-between-usb-3.1-gen1-usb-3.1-gen2,36941.html kodi.tv/

[@disnaess]

Try the I2S - LVDS (HDMI) board by Iancanada, or similar. The HDMI cable is by nature capable of transporting the I2S signal in a fully balanced way from one device to another, they've the needed amount of connectors and the needed amount of shielding (even more, if they're transmitting video signals). It doesn't need to be modified, multiplexed (like SPDIF) or whatsoever. An excellent and for now the best possible way to transfer these low voltage digital signals from a transport, or a source, to a highly capable modern DAC. Nobody cares about USB speeds, I2S + HDMI in audio is a solution avoiding USB from the design itself completely, regardless of speed or whatsoever. The main issues with usb are conversions and high jitter, it's simply inferior to true direct I2S.

[@andydelle4509]

@@disnaess And that's fine for an electronics hobbyist or an audiophile that doesn't mind his expensive devices modified and the warranty and possibly future service of any kind voided. But this is not an open public standard to interconnect products.

[@wilcalint]

@@disnaess Jitter in a USB data stream is a non-issue as virtually all receiving end circuitry have some kind of buffer. Sometimes huge buffers. The data is reclocked out of the buffer so whatever jitter is on the incoming stream is reclocked out of it. And all of that is built into the USB chip. The Raspberry Pi features four USB 3.0 ports

[@wilcalint]

www.ftdichip.com/Support/Knowledgebase/index.html?an232beffectbuffsizeandlatency.htm

[@ThinkingBetter]

@@wilcalint USB is asynchronous and can by definition not be the source of jitter.

[@damianox9]

I have a smartphone with dsp and dac with i2s interface. This phone sounds better than many expensive desk audio. I think the world is crazy choose USB.

[@1over137]

It's interesting to hear this from an audio product producer and see just how much BS they come out with. Granted most of that BS comes from what other product manufacturers are doing rather than what "can be done". The I2S bus speeds and clocks are completely configurable. You can run I2S at 8k baud and send phone conversations over a very, very long cable. If that's not long enough just re-encode the low rate I2S/SPI into UART or whatever line encoding you want. NOT covered here is the main problems with using USB. Clock skew / SOF feedback, dropouts etc. The USB Audio interface does not provide a clock. USB is poll/pull only. These is no push. An audio stream from USB cannot be asked to hold/pause, as it might be coming from a live source. You can ask for it to be reclocked periodically to rollback clock skews. But it's a pain in the a** trying to run a synchronous protocol like I2S interfacing with an asyncrhonous interface like USB. There be dragons. The other issue with USB is that 99% of USB hosts will have a SMPS and a real Earth/Ground reference (in the UK anyway). This means that ALL of that PC processor/clock/HDMI/PCIe noise that the PC is dumping onto the ground/sheild is coming to your device via that sheild. Best practice is to not connect it on the device.... if it's single purpose and does not require a shield. However the USB digital Ground is no fun either. I often see 1VPP of common mode noise (the ground referencing swinging around by 1 volt.) Getting a reliable USB connection and leaving ALL of that noise behind is very, very difficult, usually requiring high speed USB opto-isolators and ground filtering/common mode chokes etc. etc. Or using a common source of clean ground. The "basic digital cable" referred to here, is, I assume SPDIF? It's not much better. The control signals, clocks, frames and everything else is ANALOUGE encoded on the single conductor of the coaxial cable. There are many "pinch points" where the modulation of these clocks and controls can interfere with the bit stream. SPI, SPDIF, I2S are all proprieatry "standards" from Phillips and Sony from various era. The reason they aren't really standard is because, technically, they are proprietary protocol interfaces made by those companies, who allowed others to implment them for compatibility with Phillips and Sony products. It's just when it got out there everyone started using it and it spread. Without an actual published standard in like IEEE though, it grew horns, extra arms and there are dozens of different, mostly compatible sub-standards now.

[@maaxsxzone2914]

Can you recommend one USB to I2S interface? I want one device between my PC and your DAC.

[@DBTAudio]

You can research the Singxer SU2. KitsuneHiFi offers their upgraded KTE version.

[@beslemeto]

@@DBTAudio What about Sonore ultraDigital USB to SPDIF and LVDS i2s?Are they good?

[@mtabernig]

However, Ethernet cable....CAT 7 ..... Can be run for over hundreds of feet at high speeds. What gives? So i can run cat cables to transport audio over a long distance?

[@js4135]

After using I2S, I'll never use usb again! It's so much more detailed! This is from the inexpensive $56 Douk Audio converter that reclocks the USB signal to I2S and a standard HDMI cable. I'm totally sold; this is superior after hearing it. I'm hearing things I've never heard before.

[@pokrog]

i got one of those too and it blew my mind. so much so that i ordered the upgraded $150 one to try out too.

[@beslemeto]

@@pokrog Are you sure?I guess signal is just signal and converting it will decrease the sound quality?Is signal coming from your PC to that converter?

[@pokrog]

@@beslemeto oh i'm positive. i've used it on about 20 different DACs under $1000 and all of them had a massive increase in quality. clocks in entry level DACs are awful and it makes for a lot of jitter and lack of control. signal is signal, you're right, but how it get's from point a to point b is important and how accurate the clock is matters a lot. if your headphones aren't very resolving you probably wont get as much benefit but if you have a set of planars or highly resolving dynamic driver headphones then it's impossible not to notice the drastic leap in quality and detail.

[@beslemeto]

@@pokrog Thanks for responding!I am using speakers actually.Can you tell me which converter are you using exactly?I don't want to spend too much.

[@pokrog]

@@beslemeto i use a cmd-17 with the femtosecond clock upgrade but the douk u2 pro is great too. i use it for gaming when i want to use atmos and it's almost as good as the cmd-17 but it's 1/3 of the price. if you're using speakers the biggest difference you'll notice is a wider and more precise soundstage and massively better bass precision. what DAC are you using?

[@tlhuffman]

You know, I have been hearing about all of the terrible problem people have with USB digital connections, especially with Windows, for years. In the 3 years since I have been using on my Windows laptop and 3 or 4 different DACs, I have not had a bit of trouble. Never installed a driver. No problems whatsoever. Just lucky, I guess.

[@ThinkingBetter]

Tom Huffman Yes, USB could be quite horrible ages ago. I have used 8 different USB DACs in the last 6 years on my Macs with no problems except one time I had a USB DAC powered by my iMac with a poor USB cable leaking digital noise into the 5V wire. A couple of general rules for audiophile USB DAC setups are not to use your computer to power your DAC and make sure your computer and audio gear are all grounded.

[@1over137]

I think it took quite a while for the USB Device Class standards to get traction. Once those become standard in most OSes and manufacturers realised the benefit in just implmenting to a common standard, most compatibility problems went away. Not all though. You will also find DACs/ADCs which will work "driverless" in standard USB Class compliance for USB Audio standard, but... if you install the drivers will enable a virtual audio device or devices which supports much, much more.

[@marcsmithsonian9773]

I2s very short length a meter tops, USB in theory can be used to transfer anything also it allows for error corections protocol to be implemented. What system is used to stream data ? On PC you can use windows standard direct X or ASIO drivers for studio qulaity... problem is very fer people including oldtimer here have heard properly builid circuits, Meridian uses XMOS decoder to dac provides much better quality then i2s dac..

[@azzinny]

In some magazine tests, USB "asynchronous" achieves quite low jitter.

[@net_news]

Yeah but... at what cost? Complex controller, drivers installation and even a sophisticated cable!! You can get low jiiter with I2S without any of that complexity.

[@ThinkingBetter]

With asynchronous data transfer, USB is by definition not the cause of jitter issues. USB is not a negative concerning audio fidelity at all when properly used (e.g. don't power your DAC via the USB data cable 5V).

[@ThinkingBetter]

ThinkLearnSolve The point about audio running asynchronously is that USB doesn’t force a clock mastering scheme between the nodes like SPDIF and I2S do. Ideally there is only one clock managing the audio data flow and it’s the DAC’s own clock not needing PLL to try to match anyone else running a clock potentially yielding jitter issues. Ideally sufficient audio data is buffered in the DAC not yielding buffer under-runs and it is in full control of the precise timing tick of sample data being processed in it. Likewise, audio streaming from the internet is asynchronous and thus not a source of jitter. Digital studio gear needs digital synchronous audio interconnects for good reasons e.g. AES3.

[@aweidenhammer]

Cool. What dacs have I²S inputs?

[@doctorstu6011]

Aside from ours?

[@aweidenhammer]

@@doctorstu6011 well, yes. Not that I don't want yours, I just like to know stuff;)

[@adelkharisov]

Andrew Weidenhammer maybe Mytek?

[@divertiti]

Rockna

[@angelwars3176]

How about a single box that contains storage for audio/visual files (a few Tbytes at least) a user friendly GUI, processing to 'prep' audio files to different sample rates etc etc if you wish to do so and then a multi channel capable DAC and video processing? Then we can forget about tedious bloody cables that do/don't do whatever because it's all hard wired and configured from scratch.

[@angelwars3176]

@@mikebagwell8229 yeah seen it and its bigger/smaller brother. Basically a blu-ray player that has storage rather than an optical transport (loads of space!!!) and GUI on it would be great IMHO as it's an audio/visual processor multi channel capable. I get hi-fi separates but it seems really unnecessary on the digital side - i'm totally for digital audio/video but i hate hooking computers up to DAC's. I don't understand why consumers wouldn't welcome a one box solution :(

[@josefbuckland]

Woo hoo! Paul you complete my day

[@clothyardshafts]

The I2S advantage over usb is patently obvious after listening to both. That is true even for snake oil smellers like me. What I’d like to be able to do is to take avantage of I2S in moving music from my Aurender N100H to the DirectStream DAC. I know that the rendu gear will do the trick but it doesn’t have the Swiss Army knife things that I want to do - i2s connection, dsp, Roon and native DSD pass through. Nothing does all of that stuff and more’s the shame.

[@ThinkingBetter]

Let me smell some snake oil for you...in terms of USB vs. I2S, can you agree that your DAC only cares about getting the right data (no data errors), clocking the data with the right precise clock (no audible jitter) and not getting noise induced into the analog circuits (output stage)? If yes, where should your nose smell a difference between USB and I2S? Well, the analog output stage is where there is a risk. If you do an A/B test and your DAC using I2S is powered by a premium power supply and your DAC using USB is powered via the USB 5V from a computer power supply, you are right. But then your brain should ask yourself...why not use an external power supply for the DAC? Also, common mode cable noise can be induced via USB, but it's not the fault of USB itself, but more likely about poor grounding, poor PC power supply and others. I had the most horrible audio experience a few years ago using a USB powered DAC with a bad USB cable that yielded a lot of 5V noise into the analog output stage of the DAC. Just avoid that and you can be a happy USB camper.

[@clothyardshafts]

ThinkingBetter thank you.

[@AngryChineseWoman]

You say I2S > USB, but have you taken into account audiophile USB cards, reclockers and galvanic isolators ? Cause they make USB sound so much better. It sure is more expensive, but I'm curious to know how cleaned USB sounds compared to I2S.

[@HareDeLune]

Which do you prefer: Good quality from much work and great expense, or good quality for little work and a fair price?

[@AngryChineseWoman]

@@HareDeLune It depends, if optimized USB is better than I2S, then maybe it's worth it.

[@chipsnmydip]

I have a PS Audio Direcstream DAC. If I use an Innuos Zenith MKIII server, an iFi iGalvanic galvanic isolator, and an Innuos Phoenix, all connected by expensive USB cables and with high high end power cables to a Shunyata Denali V2, there is still noticeably more grunge to the sound than the same setup with a Matrix X-SPDIF2 USB/I2S converter feeding the DAC. My USB line is about as polished as you can get, but it still benefits noticeable by using the DAC I2S HDMI input. If USB can go farther, it's on the DAC side, but I haven't heard anything as good as I2S yet.

[@allansh828]

Can we assume faster rated cables better? e.g. HDMI 2.1 cable is certified 2.7x as fast HDMI 2.0 cable, Thunderbolt 3 cable is 4x as fast as USB 3.2 gen 2 cable.

[@othonmeira5608]

No. "Faster" in data cables industry refers to the speed of data transmission per second (bytes per second), not how quick the data moves between cables. There is only a limited amount of data to go between the source and the DAC per second, in that parameter it just works or don't. The discussed and controversial point in digital cables/audio is jitter.

[@allansh828]

@@othonmeira5608 but I assume faster certified cable has less jitter and better shielding?

[@othonmeira5608]

@@allansh828 not as a rule, don't use this as a guide. USB cables usually are a minor concern in the system, good clocks can minimize their impacts and solve jitter issues. Shielding is important for reducing electrical noise and EM interference, but mostly decent built usbs are well shielded.

[@allansh828]

Othon Official Account I’ve bought a few audiophile USB cables. They all sound vastly different.

[@andydelle4509]

@@allansh828 Cables don't have jitter. Nor do they cause jitter by themselves. Jitter increased by a cable is a product of the driver circuit, cable characteristics, and receiver circuit. For example a very high quality cable can result in higher jitter than a dirt cheap cable depending on the driving and receiving circuits. If the cheaper cable is better matched to the electrical characteristics of the driving and receiving circuits, it will induce less jitter. And as consumer electrical interfaces are very loosely defined, it's a crap shoot buying cables on the basis of the least jitter.

[@Real_Tim_S]

IIS = Inter-IC-Sound, USB = Universal Serial Bus. These are both data transports - one is synchronous (clock aligned), the other is asynchronous (free-running). None of this matters. At all. What matters between these two is: "Is the sample data getting to the DAC before the next sample is needed? And: how stable is the clock driving the analog conversion?" The I2S transport has signal integrity limitations, just like ethernet, or RS-232 serial (or its "big brother" USB). Data alignment between all of those lanes of control and data can be an issue whereas with USB it would not be (due to packetized data containing both the sample and the control data in a deterministic fashion). But... You can put just about anything over twisted-pair copper or fiber and get 1000's of feet out of it as long as the wire and driver are designed for the data rate. You could squeeze I2S over CAT-6 with 50MHz RS-485 drivers beyond a 1000-feet. I2S is not "new" either - it's been around since the mid-'80's, just like I2C. What's new is that it is fashionable to have something new, and most engineers wouldn't have previously dared take a parallel line bus interface outside of a chassis, but now marketing is pushing Engineers to throw caution to the wind IMHO. Much like DSD was just tapping the ADC process, and requiring more filtering on the back-end.

[@CodeAsm]

I prefer Quad SPI on my amp and audio builds :P

[@ThinkingBetter]

CodeAsm HackBird LOL, Yes or Octa 400kHz I2C 😂

[@CodeAsm]

@@ThinkingBetter 5 Mbit/s Ultra Fast-mode ? hehe, or SPI. but maybe just some homegrown protocol can be faster

[@sirsmarty274]

I just use the optical from the sound card to a dac then to the amp, its never sounded better

[@mw9558]

I2S does however not have any error detection. Not even a parity bit.

[@AudireHiEnd]

Usb audio, yes?

[@frontier9]

If its digital data - and it gets to the DAC then one doesn't sound better than the other UNLESS its getting a better data rate ... but sure USB can be a pain.

[@ford1546]

Of experience USB audio can cause a lot of noise! I've experienced it! Remember that USB has 5V and that a desktop computer has a lot of noise! The best thing is to use digital coax or optical from your computer to DAC! Have experienced noise with Cambridge dacmagic plus!

[@trog69]

Sorry you had that experience. Perhaps using a gaming motherboard has been the difference as my PC sound has been absolutely silent. No noise at all while using USB.

[@trog69]

I would add that I did not have to install anything to use USB with the Emotiva preamp/DAC.

[@ford1546]

@@trog69 Okay. I have the Asus Sabertooth 990fx and AMD FX-8350. older computer yes but no cheap computer! It is not the data signal in usb that is the problem but 5v in the usb cable that is the problem i think. Put a noise meter inside a computer and the meter goes full immediately! Measure 5v in the usb cable with the oscilloscope and you will see that there is noise

[@ford1546]

@@trog69 Almost all usb sound cards or usb dac use the standard win audio driver and therefore there are no driver problems.

[@ford1546]

@@FooBar89 you are wrong. Paul has talked about noise inside a computer several times. remember that usb carries 5V. if you are unlucky then you can get noise from some computers and others not so much noise on the same DAC. Put a noise meter inside your PC and the meter goes full! some DACs also use 5v which is inside the usb cable!

[@philipcooper8297]

Can you actually hear the difference between USB, HDMI, COAX, TOSLINK....and other wired (optical) digital connections?

[@alianlammers9025]

Yes, if one is broken;)

[@janinapalmer8368]

Philip Cooper not unless you have Golden Hearing ...

[@chrisvinicombe9947]

Never had a device that outputs all of them so never been able to test it.

[@Finn-McCool]

Yes of course... Then again, I can hear the directional change of gossamer as it bathes in an ethereal stream of pure heat within the anechoic atmosphere of a sonic vacuum. 😉 I thought every audiophile was like that. 😂

[@chipsnmydip]

You can hear a difference in all of them. Probably the most dramatically different are USB vs Toslink because you are going from the most electrical noise to electrical isolation (but with jitter).

[@anonamouse5917]

I have a USB DAC made by Stereo-Link (now defunct). I plugged it in and it configured itself in Windows. No drivers were needed. Just for S&G I plugged in 5 6ft USB extensions to see if I could hear any degradation. My ears detected none.

[@ford1546]

Many if not all usb DAC uses the same universal audio driver! and therefore we have no driver problems. Of experience USB audio can cause a lot of noise! I've experienced it! Remember that USB has 5V and that a desktop computer has a lot of noise! The best thing is to use digital coax or optical from your computer to DAC! I have experienced noise with Cambridge dacmagic plus!

[@anonamouse5917]

@@ford1546 I guess the Cambridge dacmagic plus is one to stay away from then. I'd say that the PSU in your computer must be crap, but if the 5 volt rail is THAT noisy your computer wouldn't boot. There was no noise on my Stereo-Link 1200 even with 5 6ft USB extensions.

[@treborreissul]

Better asynchronous USB. And if you want trouble take iq2s

[@chipsnmydip]

I wish it wasn't true because I2S compatibility really narrows the field of potential DACs. Over the years I put together a very elaborate and expensive multiple box USB setup where I finally don't hear any grey background/lack of contrast. And yet, throw a USB/I2S converter into the chain and an HDMI cable to my DAC and it sounds much more smooth, and refined, removing that last bit of graininess. Maybe some extremely high end DACs can salvage even more from a good USB connection, but so far I'm inclined to think no matter what you do with USB, I2S will always be a little better. I2S also sounds comparatively better than a high quality ethernet setup, which is somewhat surprising, but then with ethernet you basically have to build a noisy little computer into the DAC to unpack everything.

[@sebsoud]

If usb is so bad for audio, why so many studios used usb ("pro") soundcards to produce many albums? This usb bashing is an "hifi" fashion... Usb was not so reliable 25 years ago, and at that moment it was better to use firewire. But things have really changed since that moment, and studio, or even home-studio (not to expensive) soundcards do use usb connexion in a much more stressful way that hifi dacs do, because it's multiple-channels transmission, and in bidirectionnal way, and also with much lower latency that what is required for just hifi listening. Sorry to say that, but this is just BS to pretend that usb is not reliable for audio...

[@net_news]

Not true!! Studios used Firewire for audio!!!! But USB was backed by Intel and when Apple switched to intel Firewire was abandoned and fall into disgrace. But USB was always sh*t, everybody knew that.

[@sebsoud]

​@@net_news It was quite some time ago. For Example, latest Motu soundcards can achieve less than 2ms latency on usb...

[@net_news]

@@sebsoud sure, USB can do today what Firewire did 25 years ago.

[@sebsoud]

@@net_news No, it's much better, regarding the system interrupts stress on the os. Or firewire would still exists :-D In win 10, the audio layers have completely been rewritten, and many people acknowleged great improvements.

[@ilovecops6255]

Panasonics and others use Firewires because it goes farthers and higerhs continuous data instead of USB burst.. Dont matter for me because my panasonics camcorder with Firewires ony streams at 320x240. I DONT LIKE BEST BUY NO MORE! thankes yoe and thumbbs ups!

[@ACM1000000PT]

maybe just PS AUDIO doesn't know how to do a decent USB conection in his product.

[@str1xt]

The actual USB male part is the worst invention ever!!!! its a 50/50 chance of sticking it in the right way.

[@Baerchenization]

USB is not a bus, despite the name ;)

[@Baerchenization]

@DIngbat donedon If you think this is funny, then you probably have the wrong kind of bus in mind ;)

[@StewartMarkley]

In electronics it's spelled buss not bus.

[@mark22732]

@@StewartMarkley What? No its not.

[@StewartMarkley]

@@mark22732 You need to look it up.

[@Baerchenization]

@@StewartMarkley Also in electronics, it is called "Universal Serial Bus". The USB Wiki spells bus with single S throughout the document.

[@ethan8023]

ps audio directstream dac‘s usb design is problematic，it is not immune to usb cable quality and noise. newer design dac even much cheaper models have much better usb design

[@ilovecops6255]

Thankes you mein freiunds! Is thiss the articles yue was lookin fo... Applications that benefit from Firewire (IEEE 1394) include nonlinear (digital) video presentation and editing, desktop and commercial publishing, document imaging, home multimedia and personall computing. The low overhead and high data rates of Firewire, the ability to mix real-time and asynchronous data on a single connection, and the ability to mix low spped and high speed devices on the same network provides a truly universal connection for almost any consumer, computer, or peripheral application. THAKES YOUE AND THMUMBBS IUPS!

[@ThinkingBetter]

Yes, USB in itself is absolutely not a limiting factor for excellent audio. Your DAC is happy if it gets every bit correctly, it can run the timing in full control (avoid jitter) and there is no EMI or power rail carried noise disturbing the analog output stage.

[@mikaeljonsson99]

Being audio guys you should be able to opererate a HP filter for your microphone. Throw in a de-esser too while your at it. Remind me not to buy this stuff. :D

[@net_news]

USB is a complicated mess. I don't like USB for audio (and for anything really, USB always was a mediocre technology, Firewire was much better). The only "good" thing about USB is its ubiquity because intel was behind it from the get go but that's not a techinical merit. For audio simple "stream-like" protocols like I2S and even SPDIF are way better.

[@birgerolofsson2347]

I try to stay away from USB as much as I can.

[@birgerolofsson2347]

@@nopinionocomnt The 2 main reasons is because it's something I don't need (for music listening) and because I've had "problems" with it when I've disconnected my hard drives the wrong way so my hard drives have broken down for me.

[@PaulChabot]

Can you explain why it is fundamentally better than USB? You say its better, but not actually state any facts...

[@johnsweda2999]

Why don't you just go 5 g would be a lot easier and better plenty of data there why have cables at all?

[@chrisvinicombe9947]

For a small company that would probably be very expensive to implement. A dumb cable has got to be way cheaper.

[@ThinkingBetter]

For a fixed setup, Wi-Fi is more than adequate for audio, if you want wireless data in the home.