|
Post by indiehouse on Mar 18, 2021 18:20:38 GMT -6
The consensus between the DA of these two interfaces seems to suggest they are comparable. What about the AD?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 18, 2021 18:50:31 GMT -6
They sound totally different. I don’t know how you can qualify them because it will come down to preference.
|
|
|
Post by kcatthedog on Mar 18, 2021 19:46:07 GMT -6
Ask wiz didn’t he ditch his motu for Apollo X ?
|
|
|
Post by svart on Mar 19, 2021 10:29:25 GMT -6
Apollos all have a "sound". You either don't mind it or you hate it.
Motu is just plain. Nothing particularly bad or outstanding about it.
|
|
|
Post by veggieryan on Mar 19, 2021 10:37:52 GMT -6
Apollo X is a huge improvement over the previous 2 Apollo generations. Interestingly they have a seperate clock for each sample rate as I understand it which makes it sound similar at all rates. The latest Motu stuff has a noticeably worse clock and more "meh" sound in general. With a good external clock they both sound better. Personally its hard to justify a Motu when the Apollo X includes LUNA which is already my favorite DAW by far even in its early stages. I notice it's faster/easier to get an ear pleasing sound working in LUNA. I like the analog style workflow, tape/console gang controls and the VU meters.
|
|
|
Post by phantom on Mar 19, 2021 11:06:22 GMT -6
Apollo X is a huge improvement over the previous 2 Apollo generations. Interestingly they have a seperate clock for each sample rate as I understand it which makes it sound similar at all rates. The latest Motu stuff has a noticeably worse clock and more "meh" sound in general. With a good external clock they both sound better. Personally its hard to justify a Motu when the Apollo X includes LUNA which is already my favorite DAW by far even in its early stages. I notice it's faster/easier to get an ear pleasing sound working in LUNA. I like the analog style workflow, tape/console gang controls and the VU meters. I prefer MOTU AVBs than an Apollo X, soundwise. About Luna, I couldn't care less.
|
|
|
Post by kcatthedog on Mar 19, 2021 11:51:04 GMT -6
The x’s have two crystal clocks set for 44.1 and 48 and then their respective multiples. The idea is to reduce digital artifacts, if you record at either of the roots’ multiples and then reduce the sampling rate.
|
|
|
Post by svart on Mar 19, 2021 11:59:44 GMT -6
The x’s have two crystal clocks set for 44.1 and 48 and then their respective multiples. The idea is to reduce digital artifacts, if you record at either of the roots’ multiples and then reduce the sampling rate. Not sure what you mean. Using oscillators at much higher frequencies and dividing down is by far the best way because it divides the jitter and noise as well.
|
|
|
Post by Johnkenn on Mar 19, 2021 12:13:52 GMT -6
Apollo X is a huge improvement over the previous 2 Apollo generations. Interestingly they have a seperate clock for each sample rate as I understand it which makes it sound similar at all rates. The latest Motu stuff has a noticeably worse clock and more "meh" sound in general. With a good external clock they both sound better. Personally its hard to justify a Motu when the Apollo X includes LUNA which is already my favorite DAW by far even in its early stages. I notice it's faster/easier to get an ear pleasing sound working in LUNA. I like the analog style workflow, tape/console gang controls and the VU meters. I prefer MOTU AVBs than an Apollo X, soundwise. About Luna, I couldn't care less.
|
|
|
Post by Johnkenn on Mar 19, 2021 12:15:50 GMT -6
AD is the hardest for me to determine because I'm not tracking a bunch of cumulative stuff...Just vocals and overdubs. Once it gets to a certain point - who the hell knows?
|
|
|
Post by kcatthedog on Mar 19, 2021 12:27:25 GMT -6
svart, if sample at a higher rate, at some points math has to be done to reduce the sampling rate. UA explained that this can create digital artifacts that will contribute to jitter etc.? So, instead of having one clock sampling at various rates and the different maths being done with their various artifacts, the idea is that multiples , eg, 96 down to 48 creates less jitter and digital artifacts then say 96 down to 44? I have no idea how actually significant this is, you would know better thsn I, but I believe the X Apollo’s have the best s/n and jitter specs yet and that the 2 crystal clocks were a contributing factor ?
|
|
|
Post by popmann on Mar 19, 2021 12:38:25 GMT -6
That sounds misunderstood. A jitter percentage will be more apparent the more time each “tick” represents.
But, drew is welcome t9 get technical—as 8 think theyre all us8ng resampling ADA now....so you ARE sampling at 192 if you choose 48....the chip is just doing an SRC. So, how does that mess with my understanding of PCM? Maybe, but part of the benefit of HD is that the inherent “error” in different facets is a smaller scale. When does the second sample fire? A clock is THE PCM clock....you cant look BaCK in time and divide like you maybe can with buffered code.
|
|
|
Post by kcatthedog on Mar 19, 2021 12:46:50 GMT -6
I’m probably not explaining it correctly:) Hopefully, Drew chimes in !
|
|
|
Post by Drew @ UA on Mar 19, 2021 13:06:34 GMT -6
All you need to know.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 19, 2021 13:27:49 GMT -6
Apollos all have a "sound". You either don't mind it or you hate it. Motu is just plain. Nothing particularly bad or outstanding about it. Yep. The MOTUs have been warm, clear, and inoffensive for a while. They don’t have mushy bass, harsh treble, or a screwed up stereo image. Generic but well done analog stages Usually with JRC 4580 and LME opamps. They’re as workman as you can get without finding junk inside, some that you can’t even buy from reputable electronics distributors here. They use a Direct Digital Synthesis clock like RME does now but MOTU multi-channel interfaces are better sounding than the Firefaces. motu.com/products/motuaudio/896mk3/clock.html
|
|
|
Post by svart on Mar 19, 2021 13:34:08 GMT -6
svart , if sample at a higher rate, at some points math has to be done to reduce the sampling rate. UA explained that this can create digital artifacts that will contribute to jitter etc.? So, instead of having one clock sampling at various rates and the different maths being done with their various artifacts, the idea is that multiples , eg, 96 down to 48 creates less jitter and digital artifacts then say 96 down to 44? I have no idea how actually significant this is, you would know better thsn I, but I believe the X Apollo’s have the best s/n and jitter specs yet and that the 2 crystal clocks were a contributing factor ? That's not what I'm talking about here, but oversampling converters are regarded as much better than non-oversampling converters. I'll get to that. I'm talking about the crystals themselves. Crystals are not oscillating devices on their own and generally need to be stimulated by a vibrator circuit, kinda like plucking a guitar string to get a specific note. Oscillators are generally powered devices that output a frequency without external stimulus. There's multiple types of clocking circuits that have shown up for audio gear: 1. Some devices have a single master crystal/oscillator centered around a common frequency unrelated to clocking frequencies but are used as a reference frequency for PLLs to generate a higher frequency for division down to the wanted frequency, or directly to the wanted frequency. This method has a medium amount of jitter and noise. This was one of the most common methods for consumer gear. 2. Some devices have multiple crystals at higher multiples of the wanted frequency which are then divided directly to the wanted frequency. My 15 year old SSL converters used this method of clock generation. It's simple and elegant but has drawbacks with temperature fluctuations. Crystals are not low jitter devices, especially when not used with a feedback loop, and are extremely sensitive to temperature. 3. Some devices use very high frequency oscillators (not crystals) and are divided through logic or used directly by oversampling converters. My "Svartbox" converters used a dual oscillator that supplied 512x frequencies to the AD and encoding ICs with the AD source being divided by 2 resulting in 256fs and halving the already low 300fs jitter.. The converter IC's themselves are internally dividing and oversampling natively. Oversampling is great because it relaxes the Nyquist requirements and multiple samples of uncorrelated noise do not add while correlated signals do, increasing SNR. Digital filters in the AD chip are more steep and efficient than analog filters and digital downsampling is quite easy if the output is an even division of the input frequency. Doing this when you sample at higher multiples of the intended sampling frequency, filtering and then downsampling is the definition of "noise shaping" and has much less overall quantization noise. So I don't personally believe using "Crystal" sources is better. In fact I think it's worse. Using "two clocks" or simply two sources at multiples of the intended frequency and dividing down is arguably the best option without resorting to some expensive PLL solutions.
|
|
|
Post by svart on Mar 19, 2021 13:38:46 GMT -6
Apollos all have a "sound". You either don't mind it or you hate it. Motu is just plain. Nothing particularly bad or outstanding about it. Yep. The MOTUs have been warm, clear, and inoffensive for a while. They don’t have mushy bass, harsh treble, or a screwed up stereo image. Generic but well done analog stages Usually with JRC 4580 and LME opamps. They’re as workman as you can get without finding junk inside, some that you can’t even buy from reputable electronics distributors here. They use a Direct Digital Synthesis clock like RME does now but MOTU multi-channel interfaces are better sounding than the Firefaces. motu.com/products/motuaudio/896mk3/clock.htmlDDS sources are generally cleaner and better than PLL sources. I design with DDS devices all the time and they're great, but like most things, they require a reference clock whose quality determines the base quality of the DDS output. But with DDS devices, the output can be switched in frequency easily and the overall performance stays the same from frequency to frequency.
|
|
|
Post by popmann on Mar 19, 2021 13:51:36 GMT -6
What I TOOK from it was that they have two physical crystals. One 192 crystal and a 174 crystal inside....that should improve only the 44/88/174 quality which used to have to do uneven math from a 96, 192, or 384 clock (depending on the unit's max).
|
|
|
Post by kcatthedog on Mar 19, 2021 13:53:49 GMT -6
Thx Chris, UA’s approach to what it calls crystal clocking is explained in the video Drew posted.
|
|
|
Post by stratboy on Mar 19, 2021 14:19:49 GMT -6
Leaving aside the excellent technical discussion, I switched from MOTU to X series Apollo, just like Wiz. I sold my 16a/8a combo and bought an x6 and Twin x Duo. I did this after a lot of thought and research because I was moving, drastically downsizing my space and moving from a studio model to more of a writing room approach. I was already invested in the UA platform with two Satellite Duos, so when UA finally caught up on the converter front with the X series, it was much easier to let go of the MOTU platform. I loved those things. I’ve posted elsewhere about my converter journey so I won’t go over it again here, but for the money, the 16a was the best value I’ve ever owned.
The Duo gives me portability and the x6 adds I/O and DSP, so that hit my studio design goals; the A/D is really good, especially with Unison preamps and the D/A is just fine on the outputs I use. However, with both the MOTU setup and now the Apollo, I used(use) my Svartbox for all my 2-bus monitoring. It sounds great, drives my monitoring system in a really detailed way, and makes the MOTU/Apollo question a little less relevant for me. Thanks again, Svart, for making such a good box!
With all that said, I agree with others here that the Apollo and MOTU sound a little different, but I’ve gotten very satisfying results from both. No regrets with either platform. If you don’t want or need UA plugs, I highly recommend MOTU. And if you can find a Svartbox, it’s a steal, unless you are already in Burl//Hilo/high end DAC territory.
|
|
|
Post by svart on Mar 19, 2021 14:35:56 GMT -6
Thx Chris, UA’s approach to what it calls crystal clocking is explained in the video Drew posted. I see. It sounds like #2 in my earlier post. It was used in my old SSL converters to success. I never did measure the jitter/phase noise but I never felt like it was a problem either. Also, 4.5-6pS.. I wouldn't call that super low for internally derived clocks, but then again in my RF work I'm usually worrying about clocking with femptosecond jitter so I might be biased. I certainly wouldn't tout it as groundbreaking or novel by any stretch.
|
|
|
Post by svart on Mar 19, 2021 14:39:48 GMT -6
What I TOOK from it was that they have two physical crystals. One 192 crystal and a 174 crystal inside....that should improve only the 44/88/174 quality which used to have to do uneven math from a 96, 192, or 384 clock (depending on the unit's max). Yes, it's most likely two crystals (crystal is a common misnomer for true oscillators as well though) at some higher multiple. I highly doubt they divide these down to sampling frequencies directly and would be someone a nice steak lunch that they feed the oscillators into the converter ICs directly and use the internal oversampling dividers. That would yield the absolute best jitter performance and noise rejection.
|
|