Amir, I understand why people "hate" you in the audio world. Your ability to use SCIENCE to dispel audio myths is....wonderful!!! Your videos/reviews are very well done. As a retired teacher I very much appreciate your ability to describe things that are sometimes very complicated into something understandable. This video is a good example. Thank you for all your effort(s). Mr. Pete--------------> aging hippie
Much appreciated. Hopefully even the people who "hate" me realize that I am standing on shoulders of giants and much research has been made in these areas. On teaching, I got a lot of training teaching very difficult topics (e.g. Unit operating system internals) to people with little knowledge of topic. It is my number one focus in what I write or the videos I do: how to convey the topic in a way that people from every level of experience understand it.
The power of placebo in audio is incredible. Several studies have been made which prove that if a listener thinks the system is very expensive, he will consider it better than a cheaper comparison, even if the electronics would be identical. Also even 2db difference in gain causes the perception of a better sound for the louder sample... I have always wondered how people invest thousands and thousands to dacs and amps and then use cheap very flawed speakers, when they could get literally a thousand fold improvement by investing to the speakers and room acoustics instead of a new amplifier...
@Douglas Blake Amirn, when you done the test on the trinnov altitude can you please throw it in the cat litter box for the scam that trinnov , dolby labs atmos is thank you.
This really is like taking undergrad systems + signals and DSP classes again. I hold a BS and MS in computer engineering but my professors back then weren't as good as you Amir lol, thanks
@metal571 And that is the real issue...95% of the individuals (professors & teachers in this case) who really do understand the technology simply are not adept at TEACHING it to students who don't already have a very solid understanding of the principles used. I had a math/computer engineering professor like this. He worked on the Apollo mission's computer & trajectory systems, and could pretty much solve just about any formula/calculation in his head and spit out the answer. He obviously fully understood the math and science behind it, but couldn't teach it to others to save his life. He went "on leave" for a few months and a Physical Ed teacher/Football Coach took over the class. I went from a "D" in the class to a "B+" in just 2 months because he could actually TEACH the subject, haha. Good teachers are the absolute most valuable assets we have. Unfortunately, they are overworked and underpaid, at least here in the U.S. So huge "props" to Amir for the excellent and free education! ...still "underpaid", sorry!
@@bbfoto7248 yeah man. One of my buddies who is a math professor tells me he's basically just thrown into teaching. There's little to no training for professors, and especially for those who have technical degrees. They studied the source material, but they didn't major in how to teach it.
@Douglas Blake 100 percent. You should learn how to pick up anything quickly on the job in school. The info you need to know will come with the job itself over time. That's definitely been my expertise as well
I've watched this video 5 times over a span of one year, and each time I watch it, I'm learning new things. USB reclocker, USB filters, digital-to-digital converters are all the rage now, all claiming to fix the USB-induced jitter problems. Many of them cost several thousand dollars, which I can instead spend on room treatment for SQ improvement. Thank again for your lesson(s).
I worked for 10 years as a software developer for CD record test equipment. That people do not know we have pits and lands BUT that is not null or one.. the information is in the length of those pits and lands. So we are interested in WHEN we go from pit to land or vise versa. That lengths is counted in time from the shortest 3 time units to the longest that is 11 time units. So one thing that we looked at was each pit and lands lengths. The variation of the length 3T of a pit/land we called jitter in the industry of optical media testing/production. Another interesting thing were that we had a long 3T and a short 4T. Then we can and will get difficulty to determine if it is a 3T or 4T.. But that bad quality should not be on a disc. And error correction kicks in but only to a certain point. And if it is way off we get E32 errors that is not correctedable errors. But by then the the jitter also out of specifications (there is not allowed to be any E32 on a CD if it is then it is out of specifications). Anyway we have length variations on a optical format that has a jitter specification and then we have not yet taken the disc out from our shelf yet. In other words jitter can be a variation measurement on many different things.😍
Good explanation. Jitter is a fact of life in so many devices. Fortunately most of the time we are interested in a binary outcome as in the example you mention. With DACs, that becomes part of the analog make up of the output so can get more tricky.
“The art of digital audio” by John Watkins is a solid text on the subject ..... This is also why ripping your CD’s with something like EACD makes sense & playback from an SSD is superior to direct playback via a CD ( no matter how much you spend on that HiEnd CD transport... )
@ReaktorLeak I’m well aware of the “block diagram” there are differences under the hood in execution that are both measurable & audible - I have a 500 page reference book that delves a bit deeper...it’s a bit more than just FIFO & one of the reason Y those in the know, don’t spend $$ on fancy transport, but encode to SSD via EACD - the devil is in the details....
@ReaktorLeak Exact Audio CD - shareware that’s been around for a while, for those who don’t have professional CD authoring software - Digital sound was initially advertised as “perfect sound forever “ 40+ years ago... are we there yet? Look at the “vinyl revival” ? People have no idea how to capture, mix, master digital audio so now the same geniuses who wouldn’t know on which end of the tone arm to install a cartridge are proposing vinyl as the new medium for fidelity 🙃
I first recommend audiophiles to treat their rooms firstly. I have seen many top end systems with no room sound treatment, so are effectively listening to the room effects.
If a speaker pushes very little bass you will not have seriously problematic room modes. If your room is very reflective you will hear scatter and imaging will disappear. But in those cases and similar ones I would suggest headphones and enjoy the full experience. There is no magic in non problematic speakers. ;-)
I just sent a 1KHz sinewave through my room, and my measurement shows it is below human hearing ... therefore room treatment should not make a difference.
Thanks for this. I will need to watch several times to take it all in. Every audiophile should educate themselves and be able to spot snake oil from a mile away. Thanks.
Cutting through the BS with science and facts. Thanks for what you do. It's a monumental task which should 1)make us better consumers 2) make the industry start respecting facts and 3) start weeding out / identifying those that say / think they know but really don't. Thanks
Much appreciate the comment. For years I have been frustrated watching people provide wrong explanation of these topics. Waiting for someone else to do it didn't work and here I am :).
I tried to explain to my audiophile uncle once that even an infinitesimally small buffer could eliminate all jitter *transmission* issues, leaving it pretty much to the clock inside the DAC. Couldn't get through. Not only is the audio world full of cognitive biases, but it's also filled with that phenomenon where you can't use evidence and reason to change peoples' minds.
Louis Rossman put it best; if someone did not use evidence and logic to reach the conclusion they have, you cannot use evidence and logic to change that conclusion.
Thank you. After having watched your presentation I now not only have a better understanding of jitter but also regard it as largely a solved problem that I shouldn't worry about too much.
Amir, you focused on the DAC. Although you mentioned a poor source may cause problems as well but did not elaborate. I am interested in learning about how much quality differences are there among different CD transports' or DVD players' toslink digital audio outputs. Some say the coax outputs (relative to toslink outputs) of the same cheap source cause less jitter but more EM noise. What would you say?
Great. I was very worried I had gone over the top as far as technicalities. I actually re-recorded this video for the first time as the first one I felt was way too technical (and long).
I am still amazed by how may people choose to handle a scientific topic based on what they believe or feel. It seems a trend these days where people discard scientific results as just the opinion of scientists. If it looks very good and costs a lot of money it must be good right!? ;)
@@brydon10 One time I read a study about doctors (MDs). Their knowledge of nutrition was, on average, not different than the knowledge in the lay-public. This is because nutrition is one particular field in health, and many doctors will only ever take perhaps a single course during their education. Why bring this up? Signal processing is but one, small aspect of electrical engineering. Assuming some of the people in the audiophile industry even have formal engineering backgrounds. On top of that, the other part of audio is cognition and perception, things that are typically separated into completely separate schools. Perception will be taught in the humanities department, and engineers will be unlikely to cross over and take that course. And of course in the end, with such high profit margins on things widely accepted as "true" by audiophiles, why cut off your income stream? Was it not MIT Cables - a once-respected company in subjectivist circles - that was caught filling those "magic boxes" on their cables with glue, and claiming there was special stuff inside?
@@vinnytube1001 Paul believes in measurements to a certain level - a level that is competent and shows nothing is "broken". After that comes the "audiophile" tuning. This is what he has mentioned before, something along this line. I can't however say that none of PS Audio's electronics measure very very good though. I simply don't know.
Wow! Simply amazing explanation!! I’ve listened/read at least a dozen articles on jitter, and it never really sunk in until this comprehensive video. It’s great to understand jitter in the context of our hobby. Thank you.
Dear Amir, is the jitter a concern in the optical in of the SMSL Sanskrit 10th MKII? Would you advise the Topping E30 instead if Optical connection is important? Thank you.
Excellent video Amur! If a DAC uses the asynchronous mode of USB does this improve jitter tolerance? How is it that DACs are not able to buffer data and regenerate the clock from TOSLINK to attenuate jitter?
This is the third video I watch on the subject expecting to hear an example of what jitter actually sounds like. You'd think that for video dealing with understanding a phenomenon that affects your listening experience there'd be audio examples.
I remember my Signal Processing course and the book "Oppenheimer - Discrete-Time Signal Processing" ! People that don't have the knowledge will benefit from your very good and well explained video, thanks a lot for that. There is so much BS around audio, and unfortunately people are often wasting their money and get nothing in return, sometimes they even get less!
Excellent introductory presentation on jitter Amir. A subject that has often been misinterpreted and misunderstood over the last couple of decades by many journalists in the audio press. You mentionned that Julian Dunn's research papers on jitter may not be easy to find online. About a dozen years ago I downloaded many of them. I don't know however if they're still around.
Extremely educative Amir, thank you. So, with the jitter limits found in the paper you shared (~40 ns at best), am I right to say that any concern about jitter in modern audio is just non-sense? (Few exceptions like the SPL Phonitor X DAC, unfortunately)
Maybe do a wow and flutter distortion equivalency to digital’s jitter distortion and when you show that the best turntable's have jitter distortion that is 10k times greater than even a basic DAC, audiophile's will finally quiet down about digital jitter.
I used to subscribe to the Hans Beekhuyzen audiophile channel on here, thinking that he knew what he was talking about... then I started watching your content Amir... with the result that I've unsubscribed from him and I'm now watching EVERYTHING you post! Thank you for explaining the real science behind digital audio 😁
What is MQA actually? Nobody could ever explain it to me. Looks like another zipped format from a long distance and control over the master recording, but does anybody know what it really does?
The main case I've heard for jitter being introduced into an audio signal is via the source when running S/PDIF cables because the clock of the source is in control and is in many cases worse than that of the DAC, or is influenced by things around it (like being in a high performance PC). Is the jitter introduced in such an environment likely to be audible? I've seen people justify expensive sources and DDCs based off of this and the idea that USB introduces tons of distortion (which from what I understand is somewhat accurate, and hence the desire to use TOSLINK).
There are a few mistakes in this analysis. First audio codecs do not sample at 44100, instead they sample at a much higher rate at a 1 bit then downsample. It start with a CIC filter that takes the 1 bit signal either adds or subtracts one to an internal counter. The counter is read every Nth sample and subtracted from the previous reading. (the delta/sigma part of a delta sigma converter) This is equivalent to a flat top or averaging filter. From here you further reduce the rate though a series of filters ending with one of more halfbands. During this step, you are careful to shape the noise out of band so that the last halfband can remove most of the noise. When you hook up a codec to a PC using USB you are creating a clock domain crossing issue. The PC will ask the audio interface for N samples, but the audio interface may be a little bit ahead or behind because it is running on its own internal clock. This difference needs to be resolved by slowing down or speeding up the codec. This is often done by adding jitter to the MHz clock described above. I think they highpass the jitter to waveshape it above 20KHz. A DAC works the same but in reverse. Here are my measurements of the jitter on a focusrite scarlette 8i6 th-cam.com/video/lbjCtdp1TQw/w-d-xo.htmlsi=C6pkRhg-hFEB7mQP You will notice that all the jitter noise is above 30KHz, which you can not hear, but it will affect any op amp down stream causing other problems.
Great exposé on the nature of jitter, Amir. I particularly liked the segment on audibility. Am a physicist and long-time audiophile so wasn't phased (no pun intended) by the theory, but it was nicely illustrated with your analyzer. Video well worthwhile watching. Thanks for putting in the effort. All the best, Rob in Switzerland.
Thank you, enjoyed the video. I have often heard jitter talked about but never known what it is, now I know a little more. Still not sure I have ever heard jitter tho. ;)
I would appreciate it if you would discuss the PS Audio AC Regenerators, especially after it received a God-like assessment in the April Stereophile magazine in terms of its transcendent impact on sound. I was almost tempted to purchase their smallest unit and then I was grounded by what I have learned on the Audio Science Review. Would love your input.
This would in fact be MOST interesting - whether having one in a good chain (i.e., assuming that the equipment it is connected to filters its input voltage properly as part of the AC to DC rectification and regulation process) makes any difference whatsoever.
Pro tip: Borrow a pair of high end interconnects and put them through a roller. The snake oil will ooze out and then you can use it to enhance any device you want!
Good to see. That listening test should be made again, and with different listening groups included, from novice to “audiophile” and professional. That might give an interesting result. Sad to hear a respectful company as SPL using low quality DA converter. Maybe they should stick with doing only analogue gear. When doing a test do you share these with the developers of the product for a comment?
15:15 Maybe this is a matter of semantics, and a detail, but the intermodulation products is auditable for differing reasons than those in the time domain which rely on masking AFAIK. Obviously its not incorrect to say they are masked in the frequency domain, but it somewhat obscures why they are more audible. The IMD is typically not directly harmonically related to the signal being produced since there are sum and difference products rather than f2, f3, etc. The harmonics, being related, are less objectionable because they meet our expectations from musical instruments which do the same thing and produce similar harmonics. Intermodulation distortion, however, can produce a large splatter of frequencies that are not directly harmonically related to the fundamental, so they are much more audible, and give a raspy sort of sound to the audio when they are present. Good video, BTW. Edit: Although I will waffle a bit and say that the lower pairs are probably more audible since they can be interpreted as being their own fundamentals by the auditory center.
I don't know why people are relating jitter in audio with jitter in digital communication. In all digital stuff there is no clock, not in the real sense! All audio and video clocks are regenerated locally the only jitter left is the one within the reference crystal, witch is very low!
Note the date of the paper - 32 years ago. In that time, jitter has become so well understood that it has virtually ceased to be an issue, with manufacturers having engineered it out of their products, even at moderate cost.
Interesting video and a good refresher for me. My question is: How much of an effect does jitter have on a moderately priced modern DAC and how audible is it. Is it something we need to be concerned with?
This is exactly what I was searching for after doing some measurements on my DAC and wondering what I was seeing in the FFTs. One question I am still trying to understand: How does changing the timing (nanosecond) affect the FFT spectrum? I understand the jitter frequency changes the side-band spacing away from the base tone, but when talking about the audibility threshold, the results are given in nanoseconds.... so how does the FFT change as you make the jitter "more" audible vs less? Thank you for a great video! Clear explanations of a very complex subject.
Answering my own question (I think). It seems that the delay (nanoseconds) defines the amplitude of the side-band. Jitter Amplitude (dB) = 20 log(Jw/4) J is the jitter delay in nanoseconds w is the base frequency. so at 10kHz and 100(ns) Jitter, the amplitude is -72dB. at 10(ns) Jitter, the amplitude drops to -92dB.
Very interesting! I have a 29 year old Wadia 15 DAC I guess it was considered state of the art back in 1992! It would be so interesting to see one of these Decoding Computers measured by you as it's built like an absolute tank and they do do things a bit differently to most DACs! I would send you mine but I am in the UK so that won't work to well!! Try and get hold of one if you can!
Amir, great review!! A bit off topic but can someone please give me an explanation why certain dacs, especially r2r dacs, provide a wider and deaper soundstage. I presume, soundstage is our brain hearing certain frequency response. When looking at measurements, a lot of dacs seem to be the same, bit in practice after trying out 5 different dacs in my system, they all differ from one another
Hi Amir! Thank you for your work! Am I right in saying that according to how well many or most of present DACs measure, and that this is usually far beyond the human hearing threshold, they should all sound much similar? So how is it, in your opinion, that there are still clearly differences when hearing them, even when they both have incledibly low jitter figures? For instance: what measure gives a clue on the DAC's ability to perform good spatial 3D reconstruction? Time alignment and phase between channels is pften close to perfect, but still.... And another question: does it make sense to keep searching for the DAC with the best SINAD, when such SINAD is in the order of well over 100dB? After all, even with a modern class D ampliffier, nobody will be able to benefit from 100 dB+ dymanic range in a domestic environment, or will he?
The funniest thing was people who bought and installed rubber rings on the outer edges of their CDs. PS Audio's digital designer Ted Smith claims he can hear digital jitter of 2 parts in 10 billion. He knows he got things right by measuring it with his toe tapping meter. I'm not making this up. My grandmother had a toe tapping meter that went off the scale whenever Lawrence Welk came on. In the RBCD standard each track is scanned a minimum of 6 times. Typical noise and distortion in my experience for CD players is invariably well below the threshold of hearing. By comparison phonograph cartridges typically have distortion of 2 to 3 percent under the most optimal conditions. IMO the main difference between the sound of one DAC and another is due to differences in analog frequency response where even a small difference is audible.
Hi Amir. Very nice explanation of jitter. I think it would be great if you could explain to your viewers the different USB/Audio implementations e.g. why usb audio class devices are different from disk devices, the difference between class 1 and class 2 UAC drivers/devices and most importantly the differences between synchronous, adaptive, and asynchronous modes. This may help to explain why DAC implementation is more important than cable quality and I’d also be interested to know how you identify the class and implementation and include that in your test results. Thanks.
You don't need to know these. There are many more moving parts that can affect performance tremendously. You only need to see what it really measures. How it's implemented doesn't matter.
@@johnyang799 Sorry John but the USB implementation approach makes a massive difference to the performance of the DAC. You implement a UAC class 1 device in synchronous mode and you are asking for trouble. In fact I doubt any serious DAC designer would do that, but it’s important to understand and know what you are buying. Implementation is king with USB and DACs. Same for USB power filtering/isolation.
@@beer_goggler My original comment still stands. What you should really look at is the final performance. Don't buy anything without seeing 3rd party measurements.
@@beer_goggler Reference voltage, oscillator, pll, asrc, and more. Each of these have different specifications. All these can contribute to jitter in the final output. And these matter not less than the usb implementation. Again you really should read final measurements of jtest. Anything else don't matter for you.
It would be interesting to create an music file WITH "noise", and "jitter", and "interference" of a measurable amount, and that 'should' be in the audio spectrum. Then, have a clean file. Play them both on your system and then try to discern an audible difference, (have someone else switch the tracks randomly). I seriously wonder how many people could tell that they were hearing grunge and noise, and when they weren't. I've wanted to literally just toss in some random noise generator in friends "audiophile" systems without telling them, and see if they notice at all over a month.
@@AudioScienceReview And in this particular case, the jitter does in fact look like a jitter induced blur! But, the blur is not due to shaking so much as losing the ability to focus one's tired eyes.
I found the last part the most interesting. I've always wondered exactly how much jitter does it take to become noticeable? I hear 'audiophiles' obsessing over it and yet, I can't say I've ever heard it or found it to be problematic. Why is that?
The problem with jitter test is that you can't turn it on and off on demand. You have to use a specialized tool to generate them which hardly anyone has.
@@AudioScienceReview Is there a particular DAC that you've evaluated that clearly sounds bad due to jitter? It would seem that someone would have needlessly incorporated a PLL in the clock source to create such a problem.
Amir, this was a great video as usual. As a fellow electronics and communication engineer I appreciate your efforts to educate the audiophile community. However I don't understand the shadow effect that you mention in the FFT. The shadow effect makes sense in the time domain because our ear continues to hear the previous louder noise for a bit and sort of misses the next softer noise in close succession. In the frequency domain, the lower frequency spike, the main frequency and the higher frequency spike occur at the same instance in time. In this case, you will hear all three at the same time, and if the lower frequency spike is in a more sensitive audible frequency then it can still be more disturbing than the higher frequency spike. Care to explain more about this?
@Douglas Blake that doesn't answer the question. Amir claims that the higher frequency spike is causing problems while the lower frequency one is not but both have the same amplitude and occur at the same time. If what you say is true then we don't even need to worry about jitter because the sidebands will be of lower amplitude than the main band in almost all cases.
@Douglas Blake None of that is on topic. What does amir mean by the shadow effect? Why do higher frequency sidebands have greater impact than lower frequency sidebands?
Very interesting. I think jitter it’s the pedigree of an audio circuit. Without jitter probably some harmonics disappear, all audio circuits would sound equal.
Hi Amir, I found this video very usefull. Turns out I didn't understand jitter from a to z. Nice to know also what the audible difference between 24 bit and 16 bit can be and what it's not. I think you're coming from a balanced view, neither too objective nor too subjective, but using both and verifying. Your videos and website are like a fresh spring wind in a hifi world. Keep up the good work. By the way: Could you do a video on audible differences (if any) between high res files vs standard 44.1k?
Hi Marc. Glad to see you see the approach I take to these topics. On high res vs CD, that has been done so much that I prefer to not jump into that. In passing in other topics/videos, I may address it.
I recently went into my dealer to audition speakers. I thought that I'd be able to listen to music streamed from Tidal connect, off of my phone (Samsung S8 plus) to their system, since I know my playlists. The owner said no, that would introduce "Jitter", so he plugged his phone into the usb connection of each system that we listened to and played my requests. He was very adamant on this point and was a little condescending when I expressed doubt. Is he correct?
Since jitter can have any spectrum and level, it is impossible to answer the question of if I have heard jitter. But yes, someone challenged me once on ability to hear jitter at different levels and I did. Please see this video th-cam.com/video/0KX2yk-9ygk/w-d-xo.html
You are assuming that people's USB sources are clean, but in reality most peoples computers have very noisy USB outlets that produce a lot of jitter. Filters to isolate noise and reduce jitter are a huge help to me and others with very noisy computers. The problem with your conclusions is that you assume that if your pro equipment doesn't show something, that it doesn't exist in regular people's setups.
Is that right? The digital signal variations causes small analog variations after conversion that can be audible as noise and those analog variations are called jitter?
How do you deal with asynchronous cases? For instance, the DUT is ADC which is not synced to your AP analog-generator at all, you should see the sidebands products are breathing because of beat-tones. If you do AVG, you'll see much higher sidebands than they actually are.
Since our goal is to never have jitter spikes, measuring their amplitude correctly is not a goal. If it were yes, you would have to use a different window function and such.
When I listen to music piped through my TV then to my DAC via optical the sound isn't as good as when I go from my computer to DAC via USB. Would that be because of Jitter? That's what I've credited for the distortion I hear. Or is it something else.
The only thing that put 'jitter' on my radar was while looking for ways to playback Tidal MQA files. The consensus of what I found indicated that of the available connections types HDMI has the most 'jitter.' Followed by maybe Coaxial and Toslink - I can't remember which is supposed to be better. It has remained a question how much of a factor that is. Like, does it impact SACD DSD playback or only PCM? Does my relatively cheap Sony 4k Player that plays SACD suffer as a transport that supplies DSD over HDMI? If you believe everything you hear you would believe a $6,500 SACD player might have a memory buffer that resolves 'jitter' issues, etc. My Sony sounds good enough to me for $250.
Jitter impacts all playback, not just PCM although the impact is different depending on how the DAC works. HDMI by itself has high jitter but receivers with ESS DACs in them resample and with it, sharply reduce it. I measure this in all of my AVR and AV Processor reviews. See: www.audiosciencereview.com/forum/index.php?forums/home-theater-avr-and-processor-review.35/
Newsflash: Every CD player has a digital input buffer, to account for variations in the CD's spin speed and any errors that need to get corrected (or data re-read, because of scratched and/or dusty discs), without interrupting the audio stream (except of course for severe and/or uncorrectable issues).
@@SteelBlueVision I believe the premise of the 'audiophile' SACD player is that it has a 30 second buffer for SACD DSD disc reading - which is presumably much larger than your standard buffer. It was the PS Audio SACD player if that is of any consequence. $6,500 is too pricey for me so I'll have to suffer without that extra large buffer, etc.
Shouldn't even a $10 DAC have an internal buffer and reclock the samples to it's internal clock meaning that any jitter in the source (outside buffer overruns and underruns which would lead to actual bitloss but that would be extreme jitter) would be completely ignored?
Idea for a new video: Address the common argument that says tests are worthless because music is not a 1 kHz sine or even 32 tones. How can your tests directly translate into a device performance for playing music?
@ben and Amir, this is very easy to test, actually. All you need is a good DAC and a good ADC. You start by direct connecting the DAC/ADC to see the level of degradation in the digital information sent (level matched, of course). Then you insert a different DAC and/or ADC and compare the level of degradation, all over the course of whatever music you choose. How much do the bits change, on average, and perhaps with other statistical measurements, especially over interesting segments of a piece of music. The goal is to gauge which of two or more devices reproduces the music most faithfully at a bit level.
@@AudioScienceReview topics why measurements are important and how companies can improve without raising cost, honestly you two just chatting for any reason would be fun to listen to.
A new suggestion down the road, and to the list. Does having. digital music files with sample rates over 44.1K or 48K actually make sense. Do we need bit depths over 16. Do they actually make difference, of course if they are the same master. I feel that the industry wants to sell us the same music over and over again, tweaking here and there, so it sounds ever slightly different and get our money. There are records bing sold today that probably have been sold in Vinyl, 8 Track, Cassette, CD, Re-mastered CD, MP3, FlAC and now Vinyl LP over again. Same music just repackaged. So do we really need MQA or 24 bit 192 or 768. Can adults actually hear a difference.
@Douglas Blake Hello Douglas, thanks for the insight and your reply. My understanding is like you say. As we age we start losing hearing spectrum. Having sample rates over the CD standard of 44.1 or 48K should be more than enough for most adults. A 44.1k sample rate will reproduce frequencies up to 22.05K which is much more than a typical human, much less and adult. So why all the fuss with hight "high res" audio. Also 16 bit depth produce SNR off 96db or more. A 24 bit depth produces 142db (or so), question is do we need that. My only recommendation from years of reading is to make sure you use a lossless codec (FLAC, ALLC) and it seems that you are doing it already. Cheers.
@Douglas Blake When you say: "I typically set my systems to 24/96k for this reason" you mean your Vinyl backups? Those setting should be more than plenty for the response of the record player. However my original question to ASR is, do humans actually hear a difference or is it just marketing. I don't use vinyl anymore. As many I grew up on vinyl and still keep my Technics installed but as many other 80's kids I am done with the cleaning, the changing of the sides all that romanticism the new generations like. I rather have backed up Cds and digital music at 16/44.1 like it was originally published and I am more than happy. My Mac mini connected to the system and all the music is right there. Cheers.
@Douglas Blake I don't think your are limiting your system. The only limitation is the source material if anything. This is a great hobby hope you and all continue to enjoy it. Cheers
In a real world listening test with high quality headphones (e.g., HD580 or HD650 and good headphone amplifier), going beyond 10-bits is pretty much pointless. If you are using IEMs, you many need an additional bit or two. So, all these discussions of 16-bit vs 24-bit for music, when 12-bit is two bits too many, are really REALLY pointless. Don't believe me? Try it for yourself with a properly dithered ten bit audio conversion of the song of your choice. When you hear how practically inaudible the noise floor of 10-bit is, at listening levels that do not cause permanent damage to your hearing, you will realize that 16-bit is, already, way way in excess.
When I used a topping D90 I could definitely hear the transients were off. Once I got the Bifrost 2 things just fell into place. This is a highend hobby and you can’t simulate complex symphonies with simple test signals. You have to spend time with your music, get intimate with it, then you can pick up jitter easily.
Amirn, when you done the test on the trinnov altitude can you please throw it in the cat litter box for the scam that trinnov , dolby labs atmos is thank you.
22:35 Just want to point out that you contradicted yourself. Listener 8 went from under average listening to sine waves, to the best when listening to music, while listener 2 went from being the best at listening to sine waves to almost the worst listening to music. This supports the argument that actually listening to music is at least as important as listening to sine waves. Also, the variance between tracks is really huge for each listener. This could be caused by how similar each track is to what they listener normally listens to. As an example. perhaps listener 2 listens mostly to string quartets and so is more sensitive to sine wave distortion.
When using a sample set of exactly one to prove your case, and throwing some selection bias into that to further sweeten the mix, do realize that what you are saying is not at all statistically significant.
@@SteelBlueVision Do you actually understand any of the words you just used, and what they mean in the context of what you said? Keep in mind that dealing with sample size, bias, and significance is part of my day job. Please start with what you mean by "Sample size of exactly one."
One time I thought I heard artifacts while using a cheap DAC that I keep in my bedroom. However, I couldn't determine if it was caused by jitter or by my upstairs neighbor going to town on his extremely obese girlfriend.
wierd i think biologically we are more sensitive to timing rather than the quality of a tone. to judge distances is more important than the quality when a tiger or jaguar is hunting is in the night
Yes it sounds right, but also basically since we are conscious in the phisical world where everything is energy and frequency, the quality is already defined(sounds, white noise etc) so the precision is the only thing that is left and others are "taste"
@Douglas Blake Sorry, I do not have any sort of background in radio transmission so I might get confused a little bit. In this case the carrier is the music/test tone and the modulator is the jitter right? Don’t know if you’re familiar with Yamaha dx synthesizers but if not, the carrier normally is the frequency of key you press, or something relative to it and the modulators are set as ratios of the carrier. There are also ways to change the frequency of both the carrier and modulator with velocity, time envelopes and physical controllers so you can make either operators fixed or dynamic. I guess it must be a bit different from radio transmission.
Very correct. The only difference is that the modulation factor is *typically* so small with jitter, that it doesn't create any sidebands past the first pair. So in that regard, it looks like AM modulation. But true nature of it is FM.
Okay, 5 minutes in... I'm going to assume you don't actually tell us what audio jitter is. I mean, you could have just given an audio example of it and we would have known in the first 20 seconds.
Audiophiles have massive egos. If amp or dac cost is small they say it on their review: It still cant reach performance of a multitousand dacs or amps... They want to flex with their purchases.
Amir, I understand why people "hate" you in the audio world. Your ability to use SCIENCE to dispel audio myths is....wonderful!!! Your videos/reviews are very well done. As a retired teacher I very much appreciate your ability to describe things that are sometimes very complicated into something understandable. This video is a good example. Thank you for all your effort(s).
Mr. Pete-------------->
aging hippie
Much appreciated. Hopefully even the people who "hate" me realize that I am standing on shoulders of giants and much research has been made in these areas. On teaching, I got a lot of training teaching very difficult topics (e.g. Unit operating system internals) to people with little knowledge of topic. It is my number one focus in what I write or the videos I do: how to convey the topic in a way that people from every level of experience understand it.
He's like the James Randi of the audio world!
The power of placebo in audio is incredible. Several studies have been made which prove that if a listener thinks the system is very expensive, he will consider it better than a cheaper comparison, even if the electronics would be identical. Also even 2db difference in gain causes the perception of a better sound for the louder sample... I have always wondered how people invest thousands and thousands to dacs and amps and then use cheap very flawed speakers, when they could get literally a thousand fold improvement by investing to the speakers and room acoustics instead of a new amplifier...
Amir's taken to TH-cam like a fish to water.
@Douglas Blake Amirn, when you done the test on the trinnov altitude can you please throw it in the cat litter box for the scam that trinnov , dolby labs atmos is thank you.
This really is like taking undergrad systems + signals and DSP classes again. I hold a BS and MS in computer engineering but my professors back then weren't as good as you Amir lol, thanks
Thank you Metal. :) I wish I was taught all these concepts at University as well instead of having to learn them on the job!
@@AudioScienceReview most things are learnt on the job, University is for drinking , drugs and girls
@metal571
And that is the real issue...95% of the individuals (professors & teachers in this case) who really do understand the technology simply are not adept at TEACHING it to students who don't already have a very solid understanding of the principles used.
I had a math/computer engineering professor like this. He worked on the Apollo mission's computer & trajectory systems, and could pretty much solve just about any formula/calculation in his head and spit out the answer.
He obviously fully understood the math and science behind it, but couldn't teach it to others to save his life.
He went "on leave" for a few months and a Physical Ed teacher/Football Coach took over the class. I went from a "D" in the class to a "B+" in just 2 months because he could actually TEACH the subject, haha.
Good teachers are the absolute most valuable assets we have. Unfortunately, they are overworked and underpaid, at least here in the U.S.
So huge "props" to Amir for the excellent and free education! ...still "underpaid", sorry!
@@bbfoto7248 yeah man. One of my buddies who is a math professor tells me he's basically just thrown into teaching. There's little to no training for professors, and especially for those who have technical degrees. They studied the source material, but they didn't major in how to teach it.
@Douglas Blake 100 percent. You should learn how to pick up anything quickly on the job in school. The info you need to know will come with the job itself over time. That's definitely been my expertise as well
I've watched this video 5 times over a span of one year, and each time I watch it, I'm learning new things. USB reclocker, USB filters, digital-to-digital converters are all the rage now, all claiming to fix the USB-induced jitter problems. Many of them cost several thousand dollars, which I can instead spend on room treatment for SQ improvement. Thank again for your lesson(s).
I worked for 10 years as a software developer for CD record test equipment.
That people do not know we have pits and lands BUT that is not null or one.. the information is in the length of those pits and lands. So we are interested in WHEN we go from pit to land or vise versa.
That lengths is counted in time from the shortest 3 time units to the longest that is 11 time units.
So one thing that we looked at was each pit and lands lengths.
The variation of the length 3T of a pit/land we called jitter in the industry of optical media testing/production.
Another interesting thing were that we had a long 3T and a short 4T. Then we can and will get difficulty to determine if it is a 3T or 4T..
But that bad quality should not be on a disc. And error correction kicks in but only to a certain point.
And if it is way off we get E32 errors that is not correctedable errors. But by then the the jitter also out of specifications (there is not allowed to be any E32 on a CD if it is then it is out of specifications).
Anyway we have length variations on a optical format that has a jitter specification and then we have not yet taken the disc out from our shelf yet.
In other words jitter can be a variation measurement on many different things.😍
Good explanation. Jitter is a fact of life in so many devices. Fortunately most of the time we are interested in a binary outcome as in the example you mention. With DACs, that becomes part of the analog make up of the output so can get more tricky.
You never know who can pop up in the comments section with even more specific knowledge than the host.🤔🤫🤗😉😇 And that's saying a lot!
“The art of digital audio” by John Watkins is a solid text on the subject ..... This is also why ripping your CD’s with something like EACD makes sense & playback from an SSD is superior to direct playback via a CD ( no matter how much you spend on that HiEnd CD transport... )
@ReaktorLeak I’m well aware of the “block diagram” there are differences under the hood in execution that are both measurable & audible - I have a 500 page reference book that delves a bit deeper...it’s a bit more than just FIFO & one of the reason Y those in the know, don’t spend $$ on fancy transport, but encode to SSD via EACD - the devil is in the details....
@ReaktorLeak Exact Audio CD - shareware that’s been around for a while, for those who don’t have professional CD authoring software - Digital sound was initially advertised as “perfect sound forever “ 40+ years ago... are we there yet? Look at the “vinyl revival” ? People have no idea how to capture, mix, master digital audio so now the same geniuses who wouldn’t know on which end of the tone arm to install a cartridge are proposing vinyl as the new medium for fidelity 🙃
The value your videos bring are a great asset to the audio community. Thank you!
I first recommend audiophiles to treat their rooms firstly. I have seen many top end systems with no room sound treatment, so are effectively listening to the room effects.
An alternative is to use 'speakers that play well even in bad rooms, but that will probably exclude many high-end designs.
@@johnwright8814 speakers that play well in bad rooms still aren’t going to magically overcome poor acoustics
If a speaker pushes very little bass you will not have seriously problematic room modes. If your room is very reflective you will hear scatter and imaging will disappear.
But in those cases and similar ones I would suggest headphones and enjoy the full experience. There is no magic in non problematic speakers. ;-)
I just sent a 1KHz sinewave through my room, and my measurement shows it is below human hearing ... therefore room treatment should not make a difference.
@@andyvu8312 Its not as simple as that.
Thanks for this. I will need to watch several times to take it all in. Every audiophile should educate themselves and be able to spot snake oil from a mile away. Thanks.
Much appreciated. Yes, some of these videos need repeat watching as there is a lot to take in the first time.
Good luck with that. Because being an audiophile these days is being a believer.
I prefer „every HiFi enthusiast“
Thanks Amir. You are gifted how you explain this and show how it works. I learn from all your videos. Keep up the good work, cheers!
You are very kind. Thank you very much.
Cutting through the BS with science and facts. Thanks for what you do. It's a monumental task which should 1)make us better consumers 2) make the industry start respecting facts and 3) start weeding out / identifying those that say / think they know but really don't. Thanks
Much appreciate the comment. For years I have been frustrated watching people provide wrong explanation of these topics. Waiting for someone else to do it didn't work and here I am :).
@@AudioScienceReview We will be here too.
I tried to explain to my audiophile uncle once that even an infinitesimally small buffer could eliminate all jitter *transmission* issues, leaving it pretty much to the clock inside the DAC. Couldn't get through. Not only is the audio world full of cognitive biases, but it's also filled with that phenomenon where you can't use evidence and reason to change peoples' minds.
Louis Rossman put it best; if someone did not use evidence and logic to reach the conclusion they have, you cannot use evidence and logic to change that conclusion.
@@deadassadam Well said.
This is the best demo and breakdown of types of jitter i have wver heard in 20 years.
Thank you. After having watched your presentation I now not only have a better understanding of jitter but also regard it as largely a solved problem that I shouldn't worry about too much.
I'm all wowed and fluttered, as the song goes 'The wheel in the sky keeps turnin'
Amir, you focused on the DAC. Although you mentioned a poor source may cause problems as well but did not elaborate. I am interested in learning about how much quality differences are there among different CD transports' or DVD players' toslink digital audio outputs. Some say the coax outputs (relative to toslink outputs) of the same cheap source cause less jitter but more EM noise. What would you say?
Thank you Amir for making us more clever idiophiles :-) such a pleasure to learn through your experience.
Fascinating. Thank you. I really did learn from this. The presentation was also warm, friendly and, as a consequence, accessable.
Great. I was very worried I had gone over the top as far as technicalities. I actually re-recorded this video for the first time as the first one I felt was way too technical (and long).
Amir shall shutdown all the snake oil business!
I am still amazed by how may people choose to handle a scientific topic based on what they believe or feel. It seems a trend these days where people discard scientific results as just the opinion of scientists. If it looks very good and costs a lot of money it must be good right!? ;)
No chance. Audiophile self esteem depends on the thought that they can hear special things no instrument can capture
Even audiophile design engineers like Paul at PS Audio believe in "snake oil". I mean they're certainly convinced about all these things.
@@brydon10 One time I read a study about doctors (MDs). Their knowledge of nutrition was, on average, not different than the knowledge in the lay-public. This is because nutrition is one particular field in health, and many doctors will only ever take perhaps a single course during their education.
Why bring this up? Signal processing is but one, small aspect of electrical engineering. Assuming some of the people in the audiophile industry even have formal engineering backgrounds. On top of that, the other part of audio is cognition and perception, things that are typically separated into completely separate schools. Perception will be taught in the humanities department, and engineers will be unlikely to cross over and take that course.
And of course in the end, with such high profit margins on things widely accepted as "true" by audiophiles, why cut off your income stream? Was it not MIT Cables - a once-respected company in subjectivist circles - that was caught filling those "magic boxes" on their cables with glue, and claiming there was special stuff inside?
@@vinnytube1001 Paul believes in measurements to a certain level - a level that is competent and shows nothing is "broken". After that comes the "audiophile" tuning. This is what he has mentioned before, something along this line. I can't however say that none of PS Audio's electronics measure very very good though. I simply don't know.
Wow! Simply amazing explanation!! I’ve listened/read at least a dozen articles on jitter, and it never really sunk in until this comprehensive video. It’s great to understand jitter in the context of our hobby. Thank you.
Many thanks, Amir. You definitely helped me to understand the problem in a way I hadn't appreciated before.
Dear Amir, is the jitter a concern in the optical in of the SMSL Sanskrit 10th MKII? Would you advise the Topping E30 instead if Optical connection is important? Thank you.
Excellent video Amur! If a DAC uses the asynchronous mode of USB does this improve jitter tolerance? How is it that DACs are not able to buffer data and regenerate the clock from TOSLINK to attenuate jitter?
This is the third video I watch on the subject expecting to hear an example of what jitter actually sounds like. You'd think that for video dealing with understanding a phenomenon that affects your listening experience there'd be audio examples.
I remember my Signal Processing course and the book "Oppenheimer - Discrete-Time Signal Processing" ! People that don't have the knowledge will benefit from your very good and well explained video, thanks a lot for that. There is so much BS around audio, and unfortunately people are often wasting their money and get nothing in return, sometimes they even get less!
Brilliant!
Wish my school teachers had been as knowledgeable and informative at Amir!
Excellent introductory presentation on jitter Amir. A subject that has often been misinterpreted and misunderstood over the last couple of decades by many journalists in the audio press. You mentionned that Julian Dunn's research papers on jitter may not be easy to find online. About a dozen years ago I downloaded many of them. I don't know however if they're still around.
Thanks for another clear and informative video.
Extremely educative Amir, thank you. So, with the jitter limits found in the paper you shared (~40 ns at best), am I right to say that any concern about jitter in modern audio is just non-sense? (Few exceptions like the SPL Phonitor X DAC, unfortunately)
I wish I would have studied more math in school. I’m a biochem graduate but dabble in music production. The nitty gritty is so fascinating!
at 15:12 I would say that it spikes to -75 because your 12kHz signal is at -10 not at 0, right?
Yes. Sorry I forgot to zero that. I realized it in the middle of the video. Thanks for catching that.
Maybe do a wow and flutter distortion equivalency to digital’s jitter distortion and when you show that the best turntable's have jitter distortion that is 10k times greater than even a basic DAC, audiophile's will finally quiet down about digital jitter.
Dropping science. Thank you Amir.
I used to subscribe to the Hans Beekhuyzen audiophile channel on here, thinking that he knew what he was talking about... then I started watching your content Amir... with the result that I've unsubscribed from him and I'm now watching EVERYTHING you post! Thank you for explaining the real science behind digital audio 😁
I used to watch the Hans videos until he supported MQA, whitch is a fraud.
@@geoff37s38 amir also supported MQA :) i dont hate it tho.
@@nabildanial00 I recently bought an MQA DAC to try for myself. Not impressed.
What is MQA actually? Nobody could ever explain it to me. Looks like another zipped format from a long distance and control over the master recording, but does anybody know what it really does?
Hans is a salesman.
Excellent explanation of a complex topic like jitter, Amir.
Jitter is a problem in digital systems - so we need cryogenic cables with quantum nanoparticles. Obviously.
Ohh no! they only work at full moon. 👍
@@cryptout The moon's modulation of jitter - I have just the cable for you, sir!
@@johnwright8814 I had really bad jitter in my system, so I used these $1200 isolation feet under my DAC and the problem is gone. (This is a joke.)
And they must be expensive!
@@dennisbohner6876 Of course they are!
Well done for trying to inform the public on this topic. ;)
The main case I've heard for jitter being introduced into an audio signal is via the source when running S/PDIF cables because the clock of the source is in control and is in many cases worse than that of the DAC, or is influenced by things around it (like being in a high performance PC). Is the jitter introduced in such an environment likely to be audible? I've seen people justify expensive sources and DDCs based off of this and the idea that USB introduces tons of distortion (which from what I understand is somewhat accurate, and hence the desire to use TOSLINK).
Your videos are pure gold -- what a great explainer!
Terrific explanation. Love the detail.
Thank you for your explanation, very clear.
There are a few mistakes in this analysis. First audio codecs do not sample at 44100, instead they sample at a much higher rate at a 1 bit then downsample. It start with a CIC filter that takes the 1 bit signal either adds or subtracts one to an internal counter. The counter is read every Nth sample and subtracted from the previous reading. (the delta/sigma part of a delta sigma converter) This is equivalent to a flat top or averaging filter. From here you further reduce the rate though a series of filters ending with one of more halfbands. During this step, you are careful to shape the noise out of band so that the last halfband can remove most of the noise.
When you hook up a codec to a PC using USB you are creating a clock domain crossing issue. The PC will ask the audio interface for N samples, but the audio interface may be a little bit ahead or behind because it is running on its own internal clock. This difference needs to be resolved by slowing down or speeding up the codec. This is often done by adding jitter to the MHz clock described above. I think they highpass the jitter to waveshape it above 20KHz.
A DAC works the same but in reverse. Here are my measurements of the jitter on a focusrite scarlette 8i6
th-cam.com/video/lbjCtdp1TQw/w-d-xo.htmlsi=C6pkRhg-hFEB7mQP
You will notice that all the jitter noise is above 30KHz, which you can not hear, but it will affect any op amp down stream causing other problems.
Great exposé on the nature of jitter, Amir. I particularly liked the segment on audibility. Am a physicist and long-time audiophile so wasn't phased (no pun intended) by the theory, but it was nicely illustrated with your analyzer. Video well worthwhile watching. Thanks for putting in the effort. All the best, Rob in Switzerland.
Thank you, enjoyed the video. I have often heard jitter talked about but never known what it is, now I know a little more. Still not sure I have ever heard
jitter tho. ;)
I am glad. Thanks for watching Dave.
I would appreciate it if you would discuss the PS Audio AC Regenerators, especially after it received a God-like assessment in the April Stereophile magazine in terms of its transcendent impact on sound. I was almost tempted to purchase their smallest unit and then I was grounded by what I have learned on the Audio Science Review. Would love your input.
I have one and plan to test and review it.
This would in fact be MOST interesting - whether having one in a good chain (i.e., assuming that the equipment it is connected to filters its input voltage properly as part of the AC to DC rectification and regulation process) makes any difference whatsoever.
Thank you! Keep these videos coming, this channel is a goldmine when it comes to educating people on audio and exposing snake oil 🐍 🛢
Pro tip: Borrow a pair of high end interconnects and put them through a roller. The snake oil will ooze out and then you can use it to enhance any device you want!
Most of this went over my head, but I understood the debunking part
Good to see.
That listening test should be made again, and with different listening groups included, from novice to “audiophile” and professional. That might give an interesting result.
Sad to hear a respectful company as SPL using low quality DA converter. Maybe they should stick with doing only analogue gear.
When doing a test do you share these with the developers of the product for a comment?
15:15 Maybe this is a matter of semantics, and a detail, but the intermodulation products is auditable for differing reasons than those in the time domain which rely on masking AFAIK. Obviously its not incorrect to say they are masked in the frequency domain, but it somewhat obscures why they are more audible. The IMD is typically not directly harmonically related to the signal being produced since there are sum and difference products rather than f2, f3, etc. The harmonics, being related, are less objectionable because they meet our expectations from musical instruments which do the same thing and produce similar harmonics. Intermodulation distortion, however, can produce a large splatter of frequencies that are not directly harmonically related to the fundamental, so they are much more audible, and give a raspy sort of sound to the audio when they are present. Good video, BTW. Edit: Although I will waffle a bit and say that the lower pairs are probably more audible since they can be interpreted as being their own fundamentals by the auditory center.
I don't know why people are relating jitter in audio with jitter in digital communication. In all digital stuff there is no clock, not in the real sense! All audio and video clocks are regenerated locally the only jitter left is the one within the reference crystal, witch is very low!
Note the date of the paper - 32 years ago. In that time, jitter has become so well understood that it has virtually ceased to be an issue, with manufacturers having engineered it out of their products, even at moderate cost.
Interesting video and a good refresher for me. My question is: How much of an effect does jitter have on a moderately priced modern DAC and how audible is it. Is it something we need to be concerned with?
Dear Amir can you please do a video about a group delay impact in the headphones?
i thought i was smart until i watched you !
:)
An excellent explanation of the science behind the technology!!! Thank you! I am now a subscriber!
This is exactly what I was searching for after doing some measurements on my DAC and wondering what I was seeing in the FFTs. One question I am still trying to understand: How does changing the timing (nanosecond) affect the FFT spectrum? I understand the jitter frequency changes the side-band spacing away from the base tone, but when talking about the audibility threshold, the results are given in nanoseconds.... so how does the FFT change as you make the jitter "more" audible vs less?
Thank you for a great video! Clear explanations of a very complex subject.
Answering my own question (I think). It seems that the delay (nanoseconds) defines the amplitude of the side-band.
Jitter Amplitude (dB) = 20 log(Jw/4)
J is the jitter delay in nanoseconds
w is the base frequency.
so at 10kHz and 100(ns) Jitter, the amplitude is -72dB. at 10(ns) Jitter, the amplitude drops to -92dB.
Very interesting! I have a 29 year old Wadia 15 DAC I guess it was considered state of the art back in 1992! It would be so interesting to see one of these Decoding Computers measured by you as it's built like an absolute tank and they do do things a bit differently to most DACs! I would send you mine but I am in the UK so that won't work to well!! Try and get hold of one if you can!
Amir, great review!!
A bit off topic but can someone please give me an explanation why certain dacs, especially r2r dacs, provide a wider and deaper soundstage. I presume, soundstage is our brain hearing certain frequency response. When looking at measurements, a lot of dacs seem to be the same, bit in practice after trying out 5 different dacs in my system, they all differ from one another
All I wanted is to make the sound quality of my beat productions more stellar and here is where I’ve ended up so far. 😅
Hi Amir! Thank you for your work! Am I right in saying that according to how well many or most of present DACs measure, and that this is usually far beyond the human hearing threshold, they should all sound much similar? So how is it, in your opinion, that there are still clearly differences when hearing them, even when they both have incledibly low jitter figures? For instance: what measure gives a clue on the DAC's ability to perform good spatial 3D reconstruction? Time alignment and phase between channels is pften close to perfect, but still.... And another question: does it make sense to keep searching for the DAC with the best SINAD, when such SINAD is in the order of well over 100dB? After all, even with a modern class D ampliffier, nobody will be able to benefit from 100 dB+ dymanic range in a domestic environment, or will he?
👍 (this deserves a double like)
I really enjoyed you lesson on jitter! Thank you Amir!
Thanks Gino. Appreciate the feedback.
The funniest thing was people who bought and installed rubber rings on the outer edges of their CDs.
PS Audio's digital designer Ted Smith claims he can hear digital jitter of 2 parts in 10 billion. He knows he got things right by measuring it with his toe tapping meter. I'm not making this up. My grandmother had a toe tapping meter that went off the scale whenever Lawrence Welk came on.
In the RBCD standard each track is scanned a minimum of 6 times. Typical noise and distortion in my experience for CD players is invariably well below the threshold of hearing. By comparison phonograph cartridges typically have distortion of 2 to 3 percent under the most optimal conditions. IMO the main difference between the sound of one DAC and another is due to differences in analog frequency response where even a small difference is audible.
Hi Amir. Very nice explanation of jitter. I think it would be great if you could explain to your viewers the different USB/Audio implementations e.g. why usb audio class devices are different from disk devices, the difference between class 1 and class 2 UAC drivers/devices and most importantly the differences between synchronous, adaptive, and asynchronous modes. This may help to explain why DAC implementation is more important than cable quality and I’d also be interested to know how you identify the class and implementation and include that in your test results. Thanks.
You don't need to know these. There are many more moving parts that can affect performance tremendously. You only need to see what it really measures. How it's implemented doesn't matter.
@@johnyang799 Sorry John but the USB implementation approach makes a massive difference to the performance of the DAC. You implement a UAC class 1 device in synchronous mode and you are asking for trouble. In fact I doubt any serious DAC designer would do that, but it’s important to understand and know what you are buying. Implementation is king with USB and DACs. Same for USB power filtering/isolation.
@@beer_goggler My original comment still stands. What you should really look at is the final performance. Don't buy anything without seeing 3rd party measurements.
@@johnyang799 You comments are irrelevant to the point I’m making about USB implementation.
@@beer_goggler Reference voltage, oscillator, pll, asrc, and more. Each of these have different specifications. All these can contribute to jitter in the final output. And these matter not less than the usb implementation.
Again you really should read final measurements of jtest. Anything else don't matter for you.
It would be interesting to create an music file WITH "noise", and "jitter", and "interference" of a measurable amount, and that 'should' be in the audio spectrum. Then, have a clean file. Play them both on your system and then try to discern an audible difference, (have someone else switch the tracks randomly). I seriously wonder how many people could tell that they were hearing grunge and noise, and when they weren't. I've wanted to literally just toss in some random noise generator in friends "audiophile" systems without telling them, and see if they notice at all over a month.
Excellent review man. Keep the videos comin
When I don't have coffee, I get Jitter.
Now, there is a good line! :)
@@AudioScienceReview And in this particular case, the jitter does in fact look like a jitter induced blur! But, the blur is not due to shaking so much as losing the ability to focus one's tired eyes.
I honestly felt like I was back in one of my engineering lectures. lol.
I have seen my DAC produce the analog soundtrack longer or shorter than the time provided by the publisher. Why?
I found the last part the most interesting. I've always wondered exactly how much jitter does it take to become noticeable? I hear 'audiophiles' obsessing over it and yet, I can't say I've ever heard it or found it to be problematic. Why is that?
The problem with jitter test is that you can't turn it on and off on demand. You have to use a specialized tool to generate them which hardly anyone has.
@@AudioScienceReview Is there a particular DAC that you've evaluated that clearly sounds bad due to jitter? It would seem that someone would have needlessly incorporated a PLL in the clock source to create such a problem.
Amir, this was a great video as usual. As a fellow electronics and communication engineer I appreciate your efforts to educate the audiophile community. However I don't understand the shadow effect that you mention in the FFT. The shadow effect makes sense in the time domain because our ear continues to hear the previous louder noise for a bit and sort of misses the next softer noise in close succession. In the frequency domain, the lower frequency spike, the main frequency and the higher frequency spike occur at the same instance in time. In this case, you will hear all three at the same time, and if the lower frequency spike is in a more sensitive audible frequency then it can still be more disturbing than the higher frequency spike. Care to explain more about this?
@Douglas Blake that doesn't answer the question. Amir claims that the higher frequency spike is causing problems while the lower frequency one is not but both have the same amplitude and occur at the same time.
If what you say is true then we don't even need to worry about jitter because the sidebands will be of lower amplitude than the main band in almost all cases.
@Douglas Blake None of that is on topic. What does amir mean by the shadow effect? Why do higher frequency sidebands have greater impact than lower frequency sidebands?
Very interesting. I think jitter it’s the pedigree of an audio circuit. Without jitter probably some harmonics disappear, all audio circuits would sound equal.
Hi Amir, I found this video very usefull. Turns out I didn't understand jitter from a to z. Nice to know also what the audible difference between 24 bit and 16 bit can be and what it's not. I think you're coming from a balanced view, neither too objective nor too subjective, but using both and verifying. Your videos and website are like a fresh spring wind in a hifi world. Keep up the good work. By the way: Could you do a video on audible differences (if any) between high res files vs standard 44.1k?
Hi Marc. Glad to see you see the approach I take to these topics. On high res vs CD, that has been done so much that I prefer to not jump into that. In passing in other topics/videos, I may address it.
I recently went into my dealer to audition speakers. I thought that I'd be able to listen to music streamed from Tidal connect, off of my phone (Samsung S8 plus) to their system, since I know my playlists. The owner said no, that would introduce "Jitter", so he plugged his phone into the usb connection of each system that we listened to and played my requests. He was very adamant on this point and was a little condescending when I expressed doubt. Is he correct?
He is definitely wrong. He is running with stuff he doesn't understand.
Amir, have you tested yourself what level of jitter is audible to you?
Since jitter can have any spectrum and level, it is impossible to answer the question of if I have heard jitter. But yes, someone challenged me once on ability to hear jitter at different levels and I did. Please see this video th-cam.com/video/0KX2yk-9ygk/w-d-xo.html
You are assuming that people's USB sources are clean, but in reality most peoples computers have very noisy USB outlets that produce a lot of jitter. Filters to isolate noise and reduce jitter are a huge help to me and others with very noisy computers.
The problem with your conclusions is that you assume that if your pro equipment doesn't show something, that it doesn't exist in regular people's setups.
Is that right? The digital signal variations causes small analog variations after conversion that can be audible as noise and those analog variations are called jitter?
How do you deal with asynchronous cases? For instance, the DUT is ADC which is not synced to your AP analog-generator at all, you should see the sidebands products are breathing because of beat-tones. If you do AVG, you'll see much higher sidebands than they actually are.
Since our goal is to never have jitter spikes, measuring their amplitude correctly is not a goal. If it were yes, you would have to use a different window function and such.
When I listen to music piped through my TV then to my DAC via optical the sound isn't as good as when I go from my computer to DAC via USB. Would that be because of Jitter? That's what I've credited for the distortion I hear. Or is it something else.
It is possible that the path through the TV changes the signal or volume.
Some people claim that optical cables can induce jitter when bending. Is that true? I mean the effect must be minuscule.
Amir, your backlog with Alerts is piling ;-)
I am hoping for overtime pay. :D
The only thing that put 'jitter' on my radar was while looking for ways to playback Tidal MQA files. The consensus of what I found indicated that of the available connections types HDMI has the most 'jitter.' Followed by maybe Coaxial and Toslink - I can't remember which is supposed to be better. It has remained a question how much of a factor that is. Like, does it impact SACD DSD playback or only PCM? Does my relatively cheap Sony 4k Player that plays SACD suffer as a transport that supplies DSD over HDMI? If you believe everything you hear you would believe a $6,500 SACD player might have a memory buffer that resolves 'jitter' issues, etc. My Sony sounds good enough to me for $250.
Jitter impacts all playback, not just PCM although the impact is different depending on how the DAC works. HDMI by itself has high jitter but receivers with ESS DACs in them resample and with it, sharply reduce it. I measure this in all of my AVR and AV Processor reviews. See: www.audiosciencereview.com/forum/index.php?forums/home-theater-avr-and-processor-review.35/
Newsflash: Every CD player has a digital input buffer, to account for variations in the CD's spin speed and any errors that need to get corrected (or data re-read, because of scratched and/or dusty discs), without interrupting the audio stream (except of course for severe and/or uncorrectable issues).
@@SteelBlueVision I believe the premise of the 'audiophile' SACD player is that it has a 30 second buffer for SACD DSD disc reading - which is presumably much larger than your standard buffer. It was the PS Audio SACD player if that is of any consequence. $6,500 is too pricey for me so I'll have to suffer without that extra large buffer, etc.
looking at the formula and wondering if jitter is a kind of FM modulation 🤔
Shouldn't even a $10 DAC have an internal buffer and reclock the samples to it's internal clock meaning that any jitter in the source (outside buffer overruns and underruns which would lead to actual bitloss but that would be extreme jitter) would be completely ignored?
Idea for a new video:
Address the common argument that says tests are worthless because music is not a 1 kHz sine or even 32 tones.
How can your tests directly translate into a device performance for playing music?
I will have to figure out how to demonstrate that. I can explain it but the people who don't believe it, continue to not believe it.
@ben and Amir, this is very easy to test, actually. All you need is a good DAC and a good ADC. You start by direct connecting the DAC/ADC to see the level of degradation in the digital information sent (level matched, of course). Then you insert a different DAC and/or ADC and compare the level of degradation, all over the course of whatever music you choose. How much do the bits change, on average, and perhaps with other statistical measurements, especially over interesting segments of a piece of music. The goal is to gauge which of two or more devices reproduces the music most faithfully at a bit level.
Marvelous!
Do a podcast style TH-cam video with audioholics
Gene has reached out to me to do one. We have to figure out what the topics will be and timing.
@@AudioScienceReview topics why measurements are important and how companies can improve without raising cost, honestly you two just chatting for any reason would be fun to listen to.
great info.
A new suggestion down the road, and to the list. Does having. digital music files with sample rates over 44.1K or 48K actually make sense. Do we need bit depths over 16. Do they actually make difference, of course if they are the same master. I feel that the industry wants to sell us the same music over and over again, tweaking here and there, so it sounds ever slightly different and get our money. There are records bing sold today that probably have been sold in Vinyl, 8 Track, Cassette, CD, Re-mastered CD, MP3, FlAC and now Vinyl LP over again. Same music just repackaged. So do we really need MQA or 24 bit 192 or 768. Can adults actually hear a difference.
@Douglas Blake Hello Douglas, thanks for the insight and your reply. My understanding is like you say. As we age we start losing hearing spectrum. Having sample rates over the CD standard of 44.1 or 48K should be more than enough for most adults. A 44.1k sample rate will reproduce frequencies up to 22.05K which is much more than a typical human, much less and adult. So why all the fuss with hight "high res" audio. Also 16 bit depth produce SNR off 96db or more. A 24 bit depth produces 142db (or so), question is do we need that. My only recommendation from years of reading is to make sure you use a lossless codec (FLAC, ALLC) and it seems that you are doing it already. Cheers.
@Douglas Blake When you say: "I typically set my systems to 24/96k for this reason" you mean your Vinyl backups? Those setting should be more than plenty for the response of the record player. However my original question to ASR is, do humans actually hear a difference or is it just marketing. I don't use vinyl anymore. As many I grew up on vinyl and still keep my Technics installed but as many other 80's kids I am done with the cleaning, the changing of the sides all that romanticism the new generations like. I rather have backed up Cds and digital music at 16/44.1 like it was originally published and I am more than happy. My Mac mini connected to the system and all the music is right there. Cheers.
@Douglas Blake I don't think your are limiting your system. The only limitation is the source material if anything. This is a great hobby hope you and all continue to enjoy it. Cheers
In a real world listening test with high quality headphones (e.g., HD580 or HD650 and good headphone amplifier), going beyond 10-bits is pretty much pointless. If you are using IEMs, you many need an additional bit or two. So, all these discussions of 16-bit vs 24-bit for music, when 12-bit is two bits too many, are really REALLY pointless. Don't believe me? Try it for yourself with a properly dithered ten bit audio conversion of the song of your choice. When you hear how practically inaudible the noise floor of 10-bit is, at listening levels that do not cause permanent damage to your hearing, you will realize that 16-bit is, already, way way in excess.
When I used a topping D90 I could definitely hear the transients were off. Once I got the Bifrost 2 things just fell into place. This is a highend hobby and you can’t simulate complex symphonies with simple test signals. You have to spend time with your music, get intimate with it, then you can pick up jitter easily.
I draw the line at getting intimate with my audio devices. :)
@@AudioScienceReview LOL!
Excellent
I find jitter interferes with phase and so changes the stereo image, moreso than general tonality.
Amirn, when you done the test on the trinnov altitude can you please throw it in the cat litter box for the scam that trinnov , dolby labs atmos is thank you.
22:35 Just want to point out that you contradicted yourself.
Listener 8 went from under average listening to sine waves, to the best when listening to music, while listener 2 went from being the best at listening to sine waves to almost the worst listening to music.
This supports the argument that actually listening to music is at least as important as listening to sine waves.
Also, the variance between tracks is really huge for each listener. This could be caused by how similar each track is to what they listener normally listens to. As an example. perhaps listener 2 listens mostly to string quartets and so is more sensitive to sine wave distortion.
When using a sample set of exactly one to prove your case, and throwing some selection bias into that to further sweeten the mix, do realize that what you are saying is not at all statistically significant.
@@SteelBlueVision Do you actually understand any of the words you just used, and what they mean in the context of what you said?
Keep in mind that dealing with sample size, bias, and significance is part of my day job.
Please start with what you mean by "Sample size of exactly one."
One time I thought I heard artifacts while using a cheap DAC that I keep in my bedroom. However, I couldn't determine if it was caused by jitter or by my upstairs neighbor going to town on his extremely obese girlfriend.
What is j j j j ii ii ii ter ?
Ttthhaaaatt isss skeeewww aaannddd nnooottt, wait for it, jit?tr
wierd i think biologically we are more sensitive to timing rather than the quality of a tone. to judge distances is more important than the quality when a tiger or jaguar is hunting is in the night
Yes it sounds right, but also basically since we are conscious in the phisical world where everything is energy and frequency, the quality is already defined(sounds, white noise etc) so the precision is the only thing that is left and others are "taste"
Just like fm synthesis!
@Douglas Blake Sorry, I do not have any sort of background in radio transmission so I might get confused a little bit. In this case the carrier is the music/test tone and the modulator is the jitter right? Don’t know if you’re familiar with Yamaha dx synthesizers but if not, the carrier normally is the frequency of key you press, or something relative to it and the modulators are set as ratios of the carrier. There are also ways to change the frequency of both the carrier and modulator with velocity, time envelopes and physical controllers so you can make either operators fixed or dynamic. I guess it must be a bit different from radio transmission.
Very correct. The only difference is that the modulation factor is *typically* so small with jitter, that it doesn't create any sidebands past the first pair. So in that regard, it looks like AM modulation. But true nature of it is FM.
Okay, 5 minutes in... I'm going to assume you don't actually tell us what audio jitter is. I mean, you could have just given an audio example of it and we would have known in the first 20 seconds.
Audiophiles have massive egos.
If amp or dac cost is small they say it on their review: It still cant reach performance of a multitousand dacs or amps...
They want to flex with their purchases.
@@magog6852 youre not adding to conversation dummy
I grok.