Octatrack: Does it DITHER?
Vložit
- čas přidán 5. 12. 2019
- Does the Octatrack dither? I generally recommend to have the Octatrack flex format and recorder format set to 24-bit, and use 24-bit sources for best audio quality, and this might be one of the reasons. During these tests I also encountered a 'click' artifact/bug that is a result of internal re-sampling, but after publishing the video I discovered that this seems to be due to the default filter in FX slot 1, which apparently introduces a DC offset. In my 'Reverb Tail Artifacts' video I attribute this to the reverb FX but I'm now realizing it was most likely also caused by a default filter in FX slot 1 - • Octatrack Gremlins: Mr...
There has been one Octatrack performance recording that I published to this channel where I recall that I specifically chose to have these settings at 16-bit - • Perspectives | octatrack - I felt it made the sound a bit more dry/flat/aggressive, which suited that track. In particular I often notice a switch from 16-bit to 24-bit results in much more depth and clarity in the sound of the reverbs.
www.digido.com/portfolio-item... - publicly avaiable article by Bob Katz regarding dither, very similar to what is written in 'Mastering Audio'
If you want to learn more about digital audio systems and mastering in general make sure to also check out his other free articles www.digido.com/articles/
in his dither article, Bob Katz outlines a method to test for truncation vs. dither application -
"You can verify whether your digital audio workstation truncates digital words or does other nasty things, without any measurement instruments except your ears. Obtain the disc Best of Chesky Classics and Jazz and Audiophile Test Disc, Vol. III, Chesky JD111. Track 42 is a fade to noise without dither, demonstrating quantization distortion and loss of resolution. Track 43 is a fade to noise with white noise dither, and track 44 uses noise-shaped dither (to be explained). Use Track 43 as your test source; you should be able to hear smooth and distortion-free signal down to about -115 dB. Then listen to track 44 to see how much better it can sound. Try processing track 43 with digital equalization or level changes (both gain and attenuation, with and without dither, if it’s available in your workstation) to see what they do to the sound. If your workstation is not up to par, you’ll be shocked. Use a quiet, high-gain headphone amplifier to help reveal the low level problems."
he also mentions the important role economics plays in the implementation of digital audio systems:
"Now you know the first critical secret of digital audio: wordlengths expand. If this concept is so simple, why is it disregarded by some manufacturers? The answer is in your wallet. While DSPs are capable of performing double and triple precision arithmetic (all you have to do is store intermediate products in temporary storage registers), it slows them down, and complicates the whole process. It’s a hard choice, entirely up to the DSP programmer/processor designer, who’s been put under the gun by management to fit more program features into less space, for less money. Questions of sound quality and quantization distortion can become moot compared to the selling price."
to support this channel and gain access to Elektron project files, sample packs, high-quality music downloads, and 1-on-1 consultation visit / maxmarco
donate via paypal - www.paypal.me/maxmarcomusic
SPOTIFY - open.spotify.com/artist/3fQNK...
BANDCAMP - maxmarco.bandcamp.com/
SOUNDCLOUD - / maxmarco
FACEBOOK - / maxmarcomusic
TWITTER - / maxmarcomusic
PAYPAL - www.paypal.me/maxmarcomusic
PATREON - / maxmarco - Hudba
Journalism prof once told me, "If a headline asks a question, thanswer is always 'no' or 'maybe, I'm not totally sure.'"
Thought provoking video! I am finally starting to use my OT for more than just making loops of stuff and am starting to asl similar questions about the nature of its DSP.
hah! that's good advice :) even though I'd long ago established a general practice of going 24-bit with everything when using Octatrack the dither question has been nagging at me for a while. Even though a failure to find evidence does not mean none exists I feel more comfortable saying 24-bit settings/sources are a best practice for maximizing Octatrack sound quality. I do have a foggy idea or two for other objective tests regarding this aspect of the Octatrack, not sure anything interesting will come up though
I love your videos! This is awesome, I bet the DT is dithering because it sounds flawless
That could be one of the reasons why DT can't sample real-time - it converts samples to 16-bit mono, but the audio i/o is 24-bit, correct? I remember seeing a frequency spectrum graph on Elektronauts that shows DT isn't particularly transparent with its re-sampling and adds a bit of upper mids/highs to the signal that most users seem to enjoy - a bit 'punchier' perhaps. This thread - www.elektronauts.com/t/testing-the-quality-of-re-sampling-on-the-dt/103412 - doesn't have the spectrum graph I remember seeing, but the fact that DT re-sampling brightens the signal is discussed
@@maxmarco the DT is also 48/16 which I think sounds better than 44.1/16 or 44.1/24. My experience is that 48k really just sounds a lot better/cleaner.
thinking a bit more about it, if the DT does add dither, a brightening of the signal when re-sampling could potentially be the result of using some form of noise-shaped dither
Instant sub 😊
You're an octatrack wizard.
Yeah he is :) The best elektron chanell. I think he is much etter them Cenk with all my respect.
@@LUCIDLINES_MIXED_MEDIA You watch Cenk play the octatrack and become depressed when you can't make it sing as he does ):
@@tali813 to be honest I became depressed when I am listening to Cenk. He is a nice guy but his music is boring as hell. Check DayDreamSound tips about the OT.
great inputas always! was wondering about a related thing.. I’ve just started my first project to learn this badboy. It was meant to be a throwaway thing, but now that I’ve started to get the hang of it, I wanna make use of it.
Anyways, I’ve been resampling loops in 16bit, but would much rather have the 24bit fidelity for pitching, further resampling etc. Whenever I go in to the memory setting and change the recorder bitdepth setting up to 24, and go out of the menu (promting that unsaved recording buffers will be lost) my patterns no longer make any sound. Why is that? I’m not using the recording buffers to playback, but have saved my 16bit resampled loops to free flex slots.
I have to reload my project to get audio out again. However, I can still preview samples, only not those loaded to my flex machines. Do I need to replace all the audio-files in the flex sample-pool to make it work like it should? Also noticed that my trigs on one of my flex tracks disappeared after making changes to the memory. After reloading the project my trigs were back. That was really concerning…
Is there a safe way to change bitdepth settings on an existing project?
Maybe it’s possible to start a new project and import the pattern/scene/part data etc?
I only have a few patterns prepared, one bank, one part.
many users have mentioned a bug that happens when switching the bit-depth, I've never really experienced what you or others have described though, so I can't offer any advice
@@maxmarco Maybe it was my OS version.. Just updated, maybe it will work now. Gonna test it out. Wanna be making those 24b loops from the getgo. Damn, I haven't felt this much excitement for a piece of gear for ages! letting go f my mpc live and my maschine mk3 now. Gonna get a second octa mk1 instead!
The click you were able to generate is interesting. I wonder if its due to slew rate cut over between DSP chips, a denormalised zero signal glitch or the ghost in the machine.
yeah that's a weird one, it looks like a steady DC offset - in my reverb tail artifacts video I noticed a DC offset in the re-samples, which I attributed to the reverb itself, but this result makes me think it's inherent to SRC3 re-sampling
just realized I forgot to respect the implications of the default filter in the FX slot - just tried removing the filter and resampling from a silent file and there is no click artifact. In my reverb artifacts video I also was working with an essentially default project (as I often do in these tests) so there was most likely a filter causing that DC offset also
dunno where else to ask this since i cant find a way to pm you, anyway Max did you ever come across the gotharmans little deformer 3 ?.. knowing elektron, knowing you and knowing it, i have a strong feeling this is right up your alley, have a good weekend
The notorious click captured in the wild!
yeah!! for some time I have assumed the OT doesn't apply dither, but encountering the 'internal re-sampling click' was not something I had anticipated
@@maxmarco Is there a part 2 for the break down for dummies and how this can be addressed?
do you mean the internal re-sampling click? or the lack of dither? Re: the click I don't think there's anything you can do, but the vast majority of situations it shouldn't be noticeable - I didn't even really know it was there until I had the idea to perform this test. Situations where you are re-sampling very quiet sounds or doing ambient types of things you are presumably more likely to notice it though. Re: the lack of dither, assuming you don't need or want the RAM/storage space, simply run the Octatrack with 24-bit settings and sources, and although it's not an ideal solution, if you have to work with 16-bit sources make sure they are already dithered before loading them into the Octatrack - and if you are dithering those files yourself it's probably best to go with simple rectangular or triangular dither to avoid potential artifacts/weirdness caused by noise-shaped dither
@@maxmarco Yeah I meant to prevent it...fascinating vid...thanks
just discovered the click is most likely DC offset introduced by the default filter in FX slot 1
Dddither.