Jitter is RUINING your audio? The TRUTH about jitter! 

Thanks!! Agreed, Jim has some really cool videos.

Jim is amazing. It's fun how many forum myths of old fall apart when you start focusing on what actually makes a difference and the science behind it

I built my first DAW in 1999/2000. You had to pay quite a bit to get good convertors, and you had to pay more to get converters that were located in the breakout box (I drooled over the MOTU 1824 iirc), because converters close to the PCI bridge would pick up noise. And the budget converters weren’t guaranteed to be shielded. Anything available today in the recording space is orders of magnitude better than hardware 25 years ago. The playing field is more level than it has ever been, and it’s just going to get better.

Very well said!

Yeah, everything is so good nowadays. But... some are chasing some past times characters. And putting multiple so-so devices in chain might lead to audible problems sooner.

You're very much welcome!

Thank you! 😃

Try and produce techno and see if groove killing jitter will let you draw the same conclusion.

Amen

Ironically focusrite interfaces are pretty bad in jitter and driver stability

@@vadimmartynyuk Julian showed Focusrite performance data. Present your claims.

​@@vadimmartynyuk Maybe more than some other interfaces, but not enough to actually cause a problem for literally anyone making music. If you are deciding between interfaces based on amount of jitter, you're *looking* for a problem and picking based on random numbers on a random graph rather than using your ears.

isn't that dither? a way to mask quantization noises. Try oversampling instead

​@@miquelmarti6537 this is just in context of using an Ableton Live effect over a single channel to add some grit, say for a bass track. Sorry I wasn't being clear. Oversampling wouldn't really add much in this case. I mean, the idea is intentionally to reduce samplerate.

Audio Science Review testing found that the Jitterbug actually INCREASED jitter. Not just snake oil, but actively counterproductive.

​@@scorpioassmodeusgtx1811 *shrug* I'm relieved to hear that I didn't get a watered down dash of snake oil. Also, don't really care, per this video. EDIT: Skimmed the ASR review. The conclusion appears to be that it does nothing audible, and barely anything even confidently measurable, as is the conclusion of several other reviews linked around the ASR forums. I didn't see any mention of increasing jitter in the conclusions. Leaving my prized snake oil in the signal chain.

​@@attaboyBrad I believe the increase in jitter was only observed when combined with one DAC, namely the the Schiit Modi 2. But even then the differences were well below what could be considered audible. It's safe to say it has no practical positive or negative effect.

That noise was really audible actually.

@@StephanBuchin It can be. When dbx first came out, people were being told to way under-record, so that there would be less distortion. The dbx would take care of the hiss that way. dbx put in their instructions that you should record with peaks at or near 0 dB; that way, the noise control would be the best. Recording that way, you could still hear a little noise, especially on classical music. But things like jazz groups, rock, electronic Pop, the music covered the hiss pretty well.

Without people like you, all those brands would be bankrupt.

​@@StephanBuchin I studied and work in digital A/V and telecom electronics and thought just like you. It justified my inability to acquire high-end equipment, let alone test and compare it (not implying this is your situation). After decades, I finally had the chance to A/B systems on my own house (used gear from a friend's trade-in hi-fi store). Differences can be huge and obvious. Sometimes they are not. I appreciate your honesty, however you may want to personally test several high-end devices. My personal experience changed my mind significantly. Snake oil and overpriced gear exist, but not everything expensive is snake oil.

Nah you get to worry about bit errors way way way before you have to worry about jitter.

Well he is German!

He skipped over everything that mattered.

@@Pete.across.the.street What do you have in mind that was skipped? Honestly curious.

@@CharlesBurnsPrime he over simplified it. It's much more complex than he made it out to be. That - 120db gets amplified. If you do actual a/b listening test you can tell the differences. Not so much that b you hear the jitter, but you can tell when there's a lack of jitter.

@@Pete.across.the.street Interesting. I would ask that you suggest a test for Julian to consider. He seems to be the kind of person that can share corrections or exceptions to claims without ego problems. I would love to understand this topic better as it is often discussed with regards to digital audio.

The periodic jitter is a type of phase distortion, but aperiodic ends up more as random noise (effectively its sort of random sub-phase distortion but random is random).

This wrong idea of worrying about jitter these days comes from the peddlers who want to sell you their products... There is so much wasted money by poor audiophiles that believe those unscrupulous snake oil sellers...

you should try my plugin, NEL. It's a vibrato plugin, but on high octave settings its perlin noise modulator would essentially produce jitter

I was thinking the same, that metallic harmonic on the periodic jitter has a really ring mod bite to it.

@@mycosys yeah it's kinda pleasant to the ears to listen too aswell

already hunting for a max4live device to rip a jitter-controlled resampler from lol @@thebloodminister988

Yeah, there’s also a plug-in called decimort by D16 that has a jitter control

The side effects of jitter - even if allowed to persist until the DAC and not cancelled - end at the DAC reconstruction filter and at that point it's just sound, distorted sound but without any suspiciously bad characteristics, unlike say clipping.

@@SianaGearz thanks for the intel. I assumed there would be some effective filtering at some point, but wondered if all the extra 'hash' of jitter compromised components along the way.

Not too sure, distortion in audio components was also quite a bit higher back then. While jitter might add a hint og grunge, I think the most distortion is simply due to the older audio equipment that was used.

The quality of ADCs in general was variable in the early 80s. I've heard some very unpleasantly edgy, 'sharp' digital recordings made through early converters (different from 'clinical' or 'cold' recordings) - fortunately the tech improved fairly quickly.

Streaming services uses packets of data. As demonstrated in the video, packetized data is totally impervious to jitter. The difference in quality is compression.

Streaming audio is put into a local buffer. The audio is played directly from the audio buffer which is local memory and, therefore, no more susceptible to jitter than a local file (which, incidentally, is also streamed, just from a local media rather than a remote one). "Buffering" occurs when the data stream isn't fast enough to keep the local buffers full. In that case you will get obvious dropouts or pauses in the stream. If the stream is fast enough (which, for streaming audio, that's normally the case) then you won't get buffering. Either way, you won't get jitter.

streaming services use lossy compression spotify for example uses the vorbis codec

@@djentledjosh which has nothing to do with jitter.

@@KeithMilner correct i was responding to the top comment which said that network jitter could cause quality issues with streaming services which is not the case

Geiler Kommentar! Danke 😃

The HDSPe are fine cards, they could be a little problematic with certain OS and DAW combinations but are more widely compatible nowadays. They're widely used in pro audio and broadcast environments (where I encounter them) where uncoloured sound, good noise floor, support for house clocking, AES I/O and 24/7 operation stability are needed.

It’s „almost“ pointless to test RME stuff since their own technical documentation is pretty much complete and they’re open (and honest!) about performance specs.

Both are great for the price point, buy whichever suits your requirements the most.

Clock disparity between USB host and device does not result in erroneous (misclocked) data transmission because USB uses NRZI encoding to interleave clock and data. Bit errors do occur, isochronous transfer just like any includes a CRC checksum, which allows one bit error correction per 1ms window. Sometimes, devices do not implement CRC checking for incoming data, but with modern USB processors, you expect that all of these current devices do. At higher error rates, the error leaks into the output. There is indeed no retransmission possibility with iso, because the transmission and processing time window is already spent, but if you had errors at mere fullspeed (12Mbaud) that were uncorrectible more frequent than once every few days, you'd be THIS close to the device regularly falling off the bus, so you don't have to worry about it. And interpolation... you're just throwing words around at this point.

Shouldn't be, the signal off the CD is extensively buffered and reclocked for the purpose of Reed Solomon correction, and if it had read errors that exceed the Reed Solomon threshold, you'd be hearing bursts of static in silent sections, it gets very very obvious, or it would plain mute the window which cannot be restored, that's common as well, So CD Jitter is not something you need to consider. More likely the distortion is inherent in the bits recorded on the CD, it's how it's been recorded/mastered and "supposed" to sound like. But to be certain you could try copying it with a PC burner, also see if it throws up read errors.

Using SPDIF

If you use two interfaces, you absolutely want to sync them, otherwise you run the risk of the clocks slowly drifting apart. Simply connect the interfaces via SPDIF and select it as a clock source in on the receiving one. This way one interface is the master and the other one stays in sync with it.

I've had this problem a lot in the past. Poor motherboard design, grounding and power supply issues often contribute. I minimised by ensuring all my gear was running from a common ground. Sometimes different USB sockets run off different chipsets and some are less affected due to the PCB design and adjacency to CPU data lines or power rails. An adequately designed interface makes a big difference, I don't have any hum/interference issues with fairly common interfaces like Scarletts or Mboxes in daily use.

@@ChristopherWoods My favourite source of USB ground loop is the front panel PCB in many PC ATX enclosures. It's common for it to erroneously connect audio ground and USB ground together. The resulting loop antenna comes out so generous, that it collects unproductive current from any of the numerous sources of PC activity.

@@ChristopherWoods Perhaps I should invest in a more premium motherboard for my next build then. It's a shame there aren't "audiophile" grade lines of motherboards as well as all the "UBER LEET GAMER X EDITION" boards haha...

@@AJDOLDCHANNELARCHIVE It can be a real crapshoot. I've had high-end model motherboards with supposed 'premium' onboard sound which have performed very poorly, and likewise perfect budget systems. Sometimes the USB chipset in use can make a difference (quite a few boards have multiple USB chipsets running different pairs of ports).

​@@SianaGearz I've shared in that pain 😅

You sure about this? I would assume that the UA is not too much behind the RME. You can test it by outputting a tone of 1/4 the sample rate with the UA and record it with the RME. You can use pretty much any FFT to analyze.

Check out multitone by distortaudio. Its a free audio analyzer that has jitter test built into it. Or do it manually with REW

A word clock synchronizes the clocking of two devices. Without it, one interface will always run a bit slower. No two crystals are exactly the same frequency, and they vary over time, temp and voltage, the world is imperfect. My guess is that you are using both interfaces at the same time when noticing the glitching.

It’s not glitching. Only difference is running converters from internal clock vs external clock via word clock from apogee or RME. The UA Apollo playback D to A is the noticeable difference. Going to take measurements to show using Jtest samples. Also looked at difficulty of electronically creating a PLL from a word clock input and it doesn’t seem that hard to be honest. On the scope the PLL I created looks pretty even with 10x over sampling. Maybe all this talk about “an excellent crystal osc is easier than an excellent PLL are based more on theory than practice?

Multitone looks like PC only. Will try to get access to one but prob won’t work with the Apollo cos it’s thunderbolt 2.

If you record 120 tracks live you already have a high class interface like RME, where jitter really is the least of any issues.

The question was asked in the context of the educational aspect of Julian's video. His videos are very informative for a detailed analysis of the topic, so I miss a mention of the impact of jitter on layering recorded audio material. After 2010, almost every low-end recording device proudly featured the label "professional." This misleading marketing "educated" an entire generation, which routinely recorded brass sections, string quartets, vocal groups, but also drums and percussion on affordable "professional" devices with even "more professional" condenser microphones. Processing such material and explaining this misconception to people was an almost daily nightmare for a real professional sound engineer. I admit, I might be a bit oversensitive in this context.@@anonymus8760

Already have the IXO22 on my test bench ;)