Re: [LAU] Sound quality of the Focusrite Scarlett 18i20 2nd Gen

List overview All Threads
Download

newer

older

Linux audio on Intel NUC

Re: [LAU] every first startup of...

tom haddington

7 Jan 2017 7 Jan '17

9:45 p.m.

Hello, Ralf-- I hoped that just the headphone amps of the Focusrite

...

are weak,

i have a Focusrite 2i2. A few years ago, i sat down with a salesman to audition headphones for it. i called ahead before the session to let him know what phones interested me (an expensive mixing pair), and he told me the 2i2 didn't have strong enough drivers to run them. He eventually sold me a pair of Focal Spirit Professionals, which sounded fantastic--to me, at least. They were half the price of the pair i first mentioned, too, so he's an excellent salesman. Anyway, i wouldn't have suspected the same issue with a larger model. The problem seems to run in the family, i guess. I suspect the Presonus is the better choice.

...

i'm not the most experienced guy for this opinion, but i have to concur. i also have to apologize for not chiming in sooner. i own the Presonus of your original posts and have had success with it--for my limited projects, anyway. It came packaged with a plugin host i bought, and the company sent the Presonus with one of these <https://www.amazon.com/CableWholesales-Speed-Desktop-Powered-Multi/dp/B003V1QNQU/ref=sr_1_4?ie=UTF8&qid=1483821043&sr=8-4&keywords=multi+TT+usb+interface>. i've used it to plug the Presonus into my computer when i want to record and have never had an issue. The MIDI connections even show up in Patchage. Again, my apologies for not chiming in sooner. i don't own an iPad and didn't think i could help with that issue. Keep trying Ralf, though i won't blame you for not trying the Presonus, again. You'll come across a solution, soon. Tom

Attachments:

attachment.html (text/html — 2.1 KB)

Show replies by date

Ralf Mardorf

8 Jan 8 Jan

12:20 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

Hi Tom, thank you for reporting your experiences. Until now I only tested the outputs. Not only the headphone outputs sound bad, the analog line outputs sound disgusting, too. The sound is comparable to turning on Dolby B on a cheap tape recorder, without pumping, but with cutting highs and in addition with boosting bass or like having dirt in the ear or having hearing issues caused by a cold. All analog outputs are unusable crap. The companies claim that it's a professional device is a lie, it's even not a home recording or cheap hifi device. I experienced a new issue, that might be an issue for all USB devices on my machine. I didn't do that much tests with the Presonus as I already did with the Focusrite. A few times Qtractor as well as Ardour had issues to read WAV data from the hard disk. Fortunately Ardour informs about the reason, Qtractor doesn't. One time I needed to restart the computer to get rid of it. I tested two SATA HDDs, so it's seemingly not a broken HDD, but most likely an issue related to USB. It seems not to be related to an IRQ... [rocketmouse@archlinux ~]$ rtirq status PID CLS RTPRIO NI PRI %CPU STAT COMMAND 126 FF 90 - 130 0.2 S irq/19-ehci_hcd 130 FF 90 - 130 0.0 S irq/16-ohci_hcd 152 FF 89 - 129 0.0 S irq/17-ohci_hcd 154 FF 88 - 128 0.0 S irq/18-ohci_hcd 155 FF 88 - 128 0.0 S irq/17-ohci_hcd 157 FF 87 - 127 0.0 S irq/18-ohci_hcd 274 FF 85 - 125 0.0 S irq/18-snd_hdsp 32 FF 50 - 90 0.0 S irq/9-acpi 48 FF 50 - 90 0.0 S irq/8-rtc0 78 FF 50 - 90 0.0 S irq/26-radeon 115 FF 50 - 90 0.0 S irq/1-i8042 129 FF 50 - 90 0.0 S irq/22-firewire 132 FF 50 - 90 0.0 S irq/22-ahci[000 142 FF 50 - 90 0.0 S irq/14-pata_ati 143 FF 50 - 90 0.0 S irq/15-pata_ati 261 FF 50 - 90 0.0 S irq/7-parport0 436 FF 50 - 90 0.0 S irq/27-enp3s0 3 TS - 0 19 0.0 S ksoftirqd/0 17 TS - 0 19 0.0 S ksoftirqd/1 ...ahci doesn't share its IRQ with USB. However, I'll continue with some fine tuning. [rocketmouse@archlinux ~]$ sudo modprobe -r pata_atiixp [rocketmouse@archlinux ~]$ sudo modprobe -r firewire-ohci; sudo modprobe -r firewire_core [rocketmouse@archlinux ~]$ sudo modprobe -r ppdev; sudo modprobe -r lp [rocketmouse@archlinux ~]$ cd /sys/bus/pci/drivers/ohci-pci [rocketmouse@archlinux ohci-pci]$ printf "0000:00:13.2" | sudo tee unbind >/dev/null [rocketmouse@archlinux ohci-pci]$ printf "0000:00:13.4" | sudo tee unbind >/dev/null [rocketmouse@archlinux ohci-pci]$ rtirq status PID CLS RTPRIO NI PRI %CPU STAT COMMAND 126 FF 90 - 130 0.2 S irq/19-ehci_hcd 130 FF 90 - 130 0.0 S irq/16-ohci_hcd 152 FF 89 - 129 0.0 S irq/17-ohci_hcd 155 FF 88 - 128 0.0 S irq/17-ohci_hcd 274 FF 85 - 125 0.0 S irq/18-snd_hdsp 32 FF 50 - 90 0.0 S irq/9-acpi 48 FF 50 - 90 0.0 S irq/8-rtc0 78 FF 50 - 90 0.0 S irq/26-radeon 115 FF 50 - 90 0.0 S irq/1-i8042 132 FF 50 - 90 0.0 S irq/22-ahci[000 261 FF 50 - 90 0.0 S irq/7-parport0 436 FF 50 - 90 0.0 S irq/27-enp3s0 3 TS - 0 19 0.0 S ksoftirqd/0 17 TS - 0 19 0.0 S ksoftirqd/1 The Presonus and Focusrite cost < 400 EUR, alternatively there are just 2 Behringer for < 300 EUR and 1 RME for > 600 EUR. I don't test the Behringer devices, because I read customer reviews mentioning power supply failures for one of those devices and I had to replace caps of my Behringer mixer's switching power supply and the power supply of my ADA8000 gets boiling hot. I guess the only reason that the ADA's power supply didn't break is, that I rarely used it. Apart from this I had to clean all knobs of my Behringer mixing console. I'm living close to a main street and dirt is a serious issue, however, all of my gear suffers from the same dirt, but just the Behringer mixer needed a radical cleaning in a shorter space of time. The RME is too expensive. If all ADAT channels of my RME HDSPe AIO of my Linux PC would work, I would buy and test the RME USB audio interface, but because they don't, there's a Linux driver issue or I'm to stupid to understand how to use it, I need to buy two devices, one for the iPad and a second for the Linux PC, so the RME interface is too expensive for me at the moment. I continue testing the Focusrite now, assuming there should be no additional issues, at least no issues for the iPad, I'll stay with the Focusrite and use it for MIDI and ADAT. I still need to test the analog inputs, perhaps they are usable. I also will call my dealer, too ensure that the output sound isn't bad, because something is broken. The analog outputs of the RME HDSPe AIO and the Behringer ADA8000 do not sound that bad, so it's not some technical issue with my PC, mixer, amp or ears. The culprits are the analog outputs of the Focusrite. My two TerraTec EWX 24/96 cards aren't mounted, but I'm sure that even those cards don't sound that bad, especially since even my cheap hifi CD player does sound better, than the Focusrite. The headphone amp of the Presonus wasn't very good, but it didn't colour the sound that much and unfavourable as the Focusrite does. Unfortunately I didn't test the sound quality of the Presonus outputs. Regards, Ralf

Georg Krause

12:49 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

Hello Tom, Hello Ralf, Hello List! I own a Scarlett 18i6, which is a little bit older (3 Generations older than the device you are talking about) and i think the outputs are quite good. So i am a little surprised there are such issues. Maybe the hardware is getting worse over the time? Anyway, thank you for your intensive testing and your feedback on the linux support, someone who thinks about bying a new device will thank you too ;) I hope you have luck on your further testing, have a nice sunday! ~Georg Am 08.01.2017 um 12:20 schrieb Ralf Mardorf:

...

Ralf Mardorf

5:39 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 12:49:41 +0100, Georg Krause wrote:

...

I own a Scarlett 18i6, which is a little bit older (3 Generations older than the device you are talking about) and i think the outputs are quite good.

I've got the impression that my hearing isn't that good today. However, the Focusrite Scarlett 18i20 2nd Gen, as well as the Behringer ADA8000 directly connected to my amplifier, even today, definitively, without doubts, sound disgusting compared to the RME HDSPe AIO, but I can't hear anymore a considerable difference between the Focusrite and the ADA8000 connected to the Focusrite. OTOH the test was not really neutral, since the RME's outputs had much more gain, which likely affect the amps inputs. OTOH, if I connect all of them to my Behringer mixing console, then even the audible difference to the RME is marginal for me at the moment, if audible at all and part from this there's no mixer to control the Focusrite's analog and ADAT outputs and since the source was the same for the RME, as well as the Focusrite, I would need to lower the output levels for the RME, but actually the chosen levels were "normal"/"optimal". The hifi CD-Player doesn't sound as good as the RME, but while the colouration of the Focusrite and the ADA8000 are similar, the CD-Player has got it's own colouration, maybe a little bit closer to the RME, but with too much muddy bass. To compare my own recordings with CDs from bands, I should rip the CDs in the future and listen to the RME card only. The Focusrite and the ADA8000 are missing nuances by all frequencies, the bass is muddy too, but for the Focusrite it seems to be not that much boosted, as I claimed before. All tests were done with an AKG 240 DF headphone. I just listened very short to two different kinds of speakers, but couldn't form an opinion. I need to do a third test in some days. At the moment I'll test use the Focusrite without comparing the sound quality. After that I shouldn't listen to sound for at least one day and compare different songs with similar output gain connected to the amp's inputs. IMO it's not only important to care about the volume of the headphone and speaker outputs of the amp, but also what goes into the amp. Btw. the only available control by alsamixer is "Clock Source". ┌────────────────────────────────────────────── AlsaMixer v1.1.3 ───────────────────────────────────────────────┐ │ Card: Scarlett 18i20 USB F1: Help │ │ Chip: USB Mixer F2: System information │ │ View: F3: Playback F4: Capture F5:[All] F6: Select sound card │ │ Item: Clock Source [Internal] Esc: Exit │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ │ ┌──┐ Internal ┌──┐ ┌──┐ │ │ │OO│ │OO│ │OO│ │ │ └──┘ └──┘ └──┘ │ │ ADAT Validity < Clock Source >Internal Validity S/PDIF Validity │

Len Ovens

6:18 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017, Ralf Mardorf wrote:

...

Btw. the only available control by alsamixer is "Clock Source".

That should mean there are physical controls on the interface which may have more that 128 actual levels and don't require a mouse/kb to operate. It does make it harder for them to be remotely controlled. I am not sure if this feature is desirable or not, but the cost of having both local physical control and remote control is much higher... encoders are similar priced to pots (assuming 128 div) but LED rings are more and the CPU needs more GPIO and or glue circuitry... and the manufacture has to compete with those using pots only. As for these companies worrying about their name.. - most people are deaf. - people buy what they are told to - these companies have probably long since been switched over from being run by technical people to accountants - Or they have been bought out by some holding company Long gone are the days when a schematic was included with everything. the specification list is as misleading as ever. This is not addressed specifically towards Focusrite, but consumer audio stuff in general. Any audio interface less than $1k is probably consumer not profesional. This is particularily true of those containing mic pres or headphone outputs. Sorry for the rant... -- Len Ovens www.ovenwerks.net

Ralf Mardorf

7:04 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 09:17:58 -0800 (PST), Len Ovens wrote:

...

On Sun, 8 Jan 2017, Ralf Mardorf wrote:

Btw. the only available control by alsamixer is "Clock Source".

[snip]

There's one volume pot for one stereo output pair, by default routed to outputs 1 and 2. There are 8 level pots, one for each analog input. And there are two volume pots, one for each headphone output. There's a mixer available for Mac, not iOS and Windows >=7, so not for XP. I guess for iOS is a remote software available, that could be used, if the 18i20 is connected to a Mac or Win PC, it at least is available for the Presonus and IIRC I read the same about the Focusrite. For Linux and iOS it's only possible to chose the channels with the defaults or with settings that were changed using the Mac or Windows mixer, aka "control" for the Focusrite. It does store the changes, so the Focusrite could be used even without a computer, to mix and rout signals. Most likely I don't install Windows >=7. Regards, Ralf

Fons Adriaensen

2:04 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, Jan 08, 2017 at 12:20:02PM +0100, Ralf Mardorf wrote:

...

thank you for reporting your experiences. Until now I only tested the outputs. Not only the headphone outputs sound bad, the analog line outputs sound disgusting, too. The sound is comparable to turning on Dolby B on a cheap tape recorder, without pumping, but with cutting highs and in addition with boosting bass or like having dirt in the ear or having hearing issues caused by a cold. All analog outputs are unusable crap. The companies claim that it's a professional device is a lie, it's even not a home recording or cheap hifi device.

I find it a bit hard to believe that Focusrite, who have some reputation to take care of, would produce anything that bad. Anyway, if any of this is true then even the most amateuristic measurements should clearly show up the problem. So if you want to be taken seriously, produce some measurements, it's easy enough. BTW, how did you test the line outs ? Connected them so some unbalanced input ? If the line outs are truly balanced (as per the specs) that would short-circuit the negative signal. Anything can happen in that case. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow)

Ralf Mardorf

6:34 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 13:04:10 +0000, Fons Adriaensen wrote:

...

So if you want to be taken seriously, produce some measurements, it's easy enough.

How should it be easy to measure at home? It likely is possible to get an audio source for the measurement, but what should I do to measure the output? Connect it to some obscure input :D? Do you want to be taken seriously ;)?

...

BTW, how did you test the line outs ? Connected them so some unbalanced input ? If the line outs are truly balanced (as per the specs) that would short-circuit the negative signal. Anything can happen in that case.

The user guide says that TRS jacks could be used for balanced and TS jacks could be used for unbalanced connections. So seemingly the negative signal and ground of the TRS sockets were connected, when I used TS jacks for an unbalanced connection. If this should be handled by the device in a way that anything could happen and they should have some reputation to take care of, I wonder why it's mentioned by the user guide. I could test connecting TRS to TRS the Focusrite with my Behringer mixer, but my amp provides unbalanced inputs only and the Behringer mixer doesn't sound that good. Regards, Ralf

Robin Gareus

6:45 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On 01/08/2017 02:04 PM, Fons Adriaensen wrote:

...

I find it a bit hard to believe that Focusrite, who have some reputation to take care of, would produce anything that bad.

s/have/had/ I suppose, and mostly due to their pro equipment, Saffire etc which are still good quality (citation needed). The Scarlett series however is pretty much crap. Just terminate the input with 200 Ohm, start jnoisemeter and start wondering which award those "award winning preamps" won. flat/RMS is about -61dBFS at max gain. Don't even bother to calibrate it, those input-gain knobs have a rubber effect that randomly wanders a couple of dBs every now and then, even when the knob is at the rails. The same happens on the output.

...

So if you want to be taken seriously, produce some measurements, it's easy enough.

measurements are pretty much moot; at least for the 18i6 here. That mostly functions as book-end these days. I can't even bring myself to give it to someone. Since I only have one device, it could well be that it's some kind of manufacturing error, then again Focusrite QA should have detected this before shipping, maybe. ciao, robin

Ralf Mardorf

8:25 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 2017-01-08 at 18:45 +0100, Robin Gareus wrote:

...

measurements are pretty much moot; at least for the 18i6 here. That mostly functions as book-end these days. I can't even bring myself to give it to someone.

It's not a useless interface, since it has got ADAT input. There might be all kinds of alternatives to connect different computers or other devices by networks, but ADAT, as well as S/PDIF, AES/EBU and what ever else of this kind, aren't that fussy, as solutions that require special network software, for different operating systems. If we get a digital interface, especially ADAT, we easily could connect different devices by just plugging optical cables from one to another device. It's better to have ADAT I/Os, but a device that only has got ADAT inputs is useful, too. Regards, Ralf

Robin Gareus

8:53 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On 01/08/2017 08:24 PM, Ralf Mardorf wrote:

...

On Sun, 2017-01-08 at 18:45 +0100, Robin Gareus wrote:

measurements are pretty much moot; at least for the 18i6 here. That mostly functions as book-end these days. I can't even bring myself to give it to someone.

It's not a useless interface, since it has got ADAT input.

only if you have some ADAT or S/PDIF devices to connect. I do pull it off the shelf on occasion for testing MIDI sync or some non-critical things where the device is not the bottleneck in the signal-chain. It's a good a soundcard as any for modern Pop/Rock and even better since it includes a hardware mixer. on that note:

...

Missing alsamixer controls aren't a problem for me.

..except that the default routing inside the device may affect what you're hearing. OOTB it patches analog inputs through to the outputs ciao, robin

Georg Krause

9:02 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

Am 08.01.2017 um 20:53 schrieb Robin Gareus:

...

Missing alsamixer controls aren't a problem for me.

..except that the default routing inside the device may affect what you're hearing. OOTB it patches analog inputs through to the outputs

Not just that, my focusrite sounds crappy when the outputs are "too loud" in alsamixer. means over 80% its not usable. so maybe that is an issue, Ralf?

Johannes Kroll

9:25 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 21:01:45 +0100 Georg Krause <mail(a)georg-krause.net> wrote:

...

Am 08.01.2017 um 20:53 schrieb Robin Gareus: >> Missing alsamixer controls aren't a problem for me. > ..except that the default routing inside the device may affect what > you're hearing. OOTB it patches analog inputs through to the outputs

...which would mean that you if you listen to an input, you mix it with a copy of itself delayed by the buffer size. That would definately sound weird, and could explain Ralf's experience.

...

Not just that, my focusrite sounds crappy when the outputs are "too loud" in alsamixer. means over 80% its not usable. so maybe that is an issue, Ralf?

That's definately an issue with my 18i8 first-gen as well. When I set an output control to more than 0db, the output doesn't clip -- it integer-wraps, which causes really evil distortion. I confirmed that with an oscilloscope. I'm not sure whether this is a Linux driver issue, or bad design. It shouldn't be possible to turn volume up beyond the 0db setting if that causes such distortion.

Ralf Mardorf

9:26 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 21:01:45 +0100, Georg Krause wrote:

...

Ok, this could be nasty, assuming I should use the analog I/Os for some reason, so I at least should get rid of passing inputs through to outputs.

...

Not just that, my focusrite sounds crappy when the outputs are "too loud" in alsamixer. means over 80% its not usable. so maybe that is an issue, Ralf?

I don't know if there is a less optimal bias point for the analog domain, at least the audio source didn't cause a signal > 0 dBFS. To patch the kernel, do I need to search an alsa / snd_usb_audio source file to add the ID or is it somewhere else? ... [rocketmouse@archlinux ~]$ aplay -l **** List of PLAYBACK Hardware Devices **** card 0: HDSPMx579bcc [RME AIO_579bcc], device 0: RME AIO [RME AIO] Subdevices: 1/1 Subdevice #0: subdevice #0 card 3: USB [Scarlett 18i20 USB], device 0: USB Audio [USB Audio] Subdevices: 1/1 Subdevice #0: subdevice #0 [rocketmouse@archlinux ~]$ lsusb Bus 001 Device 017: ID 1235:8201 Focusrite-Novation Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 005 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub Bus 003 Device 002: ID 192f:0916 Avago Technologies, Pte. Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub ... and the ID would be "ID 1235:8201"? Regards, Ralf

Johannes Kroll

9:46 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 21:26:49 +0100 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote:

...

To patch the kernel, do I need to search an alsa / snd_usb_audio source file to add the ID or is it somewhere else? ...

First, look here: https://github.com/smilingthax/alsa-driver_scarlett/commit/7a77707b66ee476d… (This patch is now inside the mainline kernel; I did not write it) If your second-gen has identical controls to the first-gen 18i20, then you would have to simply insert your USB ID there and it would just work. That is, in sound/usb/mixer.c around line 2395, create another 'case' - case USB_ID(0x1235, 0x8201): /* Focusrite Scarlett 18i20 2nd-gen */ And in sound/usb/scarlettmixer.c after line 1226, insert case USB_ID(0x1235, 0x8201): info = &s18i20_info; break; Line numbers might not be 100% identical depending on your kernel version. That should treat your device like a first-gen 18i20, if I didn't miss anything. No guarantee it will work, but it might.

Ralf Mardorf

11:06 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 21:25:35 +0100, Johannes Kroll wrote:

...

...which would mean that you if you listen to an input, you mix it with a copy of itself delayed by the buffer size. That would definately sound weird, and could explain Ralf's experience.

No, since on Linux I used Ardour and directly connected the stereo track's outputs by jackd to playback_n. I didn't test any of the 18i20's inputs. [rocketmouse@archlinux ~]$ pacman -Q ardour5 jack2 ardour5 5.5-1 jack2 1.9.10.r202.gf2ece24-1 On the iPad I used Cubasis 2 with a single stereo audio track and directly connected Cubasis 2 to the Focusrite outputs. The original WAV is ripped from Crowded House's CD "Woodface", track 5, "weather with you", a 44.1 KHz, 16 bit file and then imported by Ardour at 44.1 KHz, 32 bit float and by Cubasis at 48 KHz, 24 bit.

...

Not just that, my focusrite sounds crappy when the outputs are "too loud" in alsamixer. means over 80% its not usable. so maybe that is an issue, Ralf?

...

0 dB related to what? The audio files were played with an output

< 0 dBFS and perhaps at max. 0 dBFS here. Does the Focusrite mixer control the analog domain? Does it increase a digital signal that comes from the DAW by the digital domain? On Sun, 8 Jan 2017 21:46:40 +0100, Johannes Kroll wrote:

...

First, look here: https://github.com/smilingthax/alsa-driver_scarlett/commit/7a77707b66ee476d… (This patch is now inside the mainline kernel; I did not write it)

It's neither part of the latest kernel supported by the rt project, nor by the current mainline rc, https://www.kernel.org/ : [rocketmouse@archlinux tmp]$ wget -q https://www.kernel.org/pub/linux/kernel/v4.x/linux-4.9.tar.gz [rocketmouse@archlinux tmp]$ tar -xf linux-4.9.tar.gz [rocketmouse@archlinux tmp]$ grep USB_ID -n linux-4.9/sound/usb/mixer.c 850: case USB_ID(0x0763, 0x2030): /* M-Audio Fast Track C400 */ 851: case USB_ID(0x0763, 0x2031): /* M-Audio Fast Track C600 */ 878: case USB_ID(0x0763, 0x2081): /* M-Audio Fast Track Ultra 8R */ 879: case USB_ID(0x0763, 0x2080): /* M-Audio Fast Track Ultra */ 898: case USB_ID(0x0471, 0x0101): 899: case USB_ID(0x0471, 0x0104): 900: case USB_ID(0x0471, 0x0105): 901: case USB_ID(0x0672, 0x1041): 915: case USB_ID(0x046d, 0x09a4): 925: case USB_ID(0x046d, 0x0807): /* Logitech Webcam C500 */ 926: case USB_ID(0x046d, 0x0808): 927: case USB_ID(0x046d, 0x0809): 928: case USB_ID(0x046d, 0x0819): /* Logitech Webcam C210 */ 929: case USB_ID(0x046d, 0x081b): /* HD Webcam c310 */ 930: case USB_ID(0x046d, 0x081d): /* HD Webcam c510 */ 931: case USB_ID(0x046d, 0x0825): /* HD Webcam c270 */ 932: case USB_ID(0x046d, 0x0826): /* HD Webcam c525 */ 933: case USB_ID(0x046d, 0x08ca): /* Logitech Quickcam Fusion */ 934: case USB_ID(0x046d, 0x0991): 1503: case USB_ID(0x08bb, 0x2702): 1509: case USB_ID(0x1130, 0xf211): [rocketmouse@archlinux tmp]$ ls -lh linux-4.9/sound/usb/sc* ls: cannot access 'linux-4.9/sound/usb/sc*': No such file or directory [rocketmouse@archlinux tmp]$ ls -lh linux-4.9/sound/usb/s* -rw-r--r-- 1 rocketmouse rocketmouse 21K Dec 11 20:17 linux-4.9/sound/usb/stream.c -rw-r--r-- 1 rocketmouse rocketmouse 286 Dec 11 20:17 linux-4.9/sound/usb/stream.h [rocketmouse@archlinux tmp]$ wget -q https://cdn.kernel.org/pub/linux/kernel/v4.x/testing/linux-4.10-rc2.tar.xz [rocketmouse@archlinux tmp]$ tar -xf linux-4.10-rc2.tar.xz [rocketmouse@archlinux tmp]$ grep USB_ID -n linux-4.10-rc2/sound/usb/mixer.c 850: case USB_ID(0x0763, 0x2030): /* M-Audio Fast Track C400 */ 851: case USB_ID(0x0763, 0x2031): /* M-Audio Fast Track C600 */ 878: case USB_ID(0x0763, 0x2081): /* M-Audio Fast Track Ultra 8R */ 879: case USB_ID(0x0763, 0x2080): /* M-Audio Fast Track Ultra */ 898: case USB_ID(0x0471, 0x0101): 899: case USB_ID(0x0471, 0x0104): 900: case USB_ID(0x0471, 0x0105): 901: case USB_ID(0x0672, 0x1041): 915: case USB_ID(0x046d, 0x09a4): 925: case USB_ID(0x046d, 0x0807): /* Logitech Webcam C500 */ 926: case USB_ID(0x046d, 0x0808): 927: case USB_ID(0x046d, 0x0809): 928: case USB_ID(0x046d, 0x0819): /* Logitech Webcam C210 */ 929: case USB_ID(0x046d, 0x081b): /* HD Webcam c310 */ 930: case USB_ID(0x046d, 0x081d): /* HD Webcam c510 */ 931: case USB_ID(0x046d, 0x0825): /* HD Webcam c270 */ 932: case USB_ID(0x046d, 0x0826): /* HD Webcam c525 */ 933: case USB_ID(0x046d, 0x08ca): /* Logitech Quickcam Fusion */ 934: case USB_ID(0x046d, 0x0991): 935: case USB_ID(0x046d, 0x09a2): /* QuickCam Communicate Deluxe/S7500 */ 1504: case USB_ID(0x08bb, 0x2702): 1510: case USB_ID(0x1130, 0xf211): [rocketmouse@archlinux tmp]$ ls -lh linux-4.10-rc2/sound/usb/sc* ls: cannot access 'linux-4.10-rc2/sound/usb/sc*': No such file or directory [rocketmouse@archlinux tmp]$ ls -lh linux-4.10-rc2/sound/usb/s* -rw-r--r-- 1 rocketmouse rocketmouse 21K Jan 1 23:31 linux-4.10-rc2/sound/usb/stream.c -rw-r--r-- 1 rocketmouse rocketmouse 286 Jan 1 23:31 linux-4.10-rc2/sound/usb/stream.h [rocketmouse@archlinux tmp]$ Regards, Ralf

Len Ovens

11:24 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

This one's internal mixer can be controlled from Linux: http://motu.com/products/avb/624 It is a bit pricy though. It does take ADAT, but does not have word clock in. Can handle as many as 256 audio lines (128 in and 128 out with an AVB extender or two) It says it is USB2 compliant (works with OSx in USB2 or 3 mode with no extra driver). I would try before buy. -- Len Ovens www.ovenwerks.net

Robin Gareus

11:25 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On 01/08/2017 11:06 PM, Ralf Mardorf wrote:

...

[rocketmouse@archlinux tmp]$ ls -lh linux-4.10-rc2/sound/usb/s*

Try .../usb/*sc* to match "mixer_scarlett.cc". it's in vanilla linux since 2014: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/sound/… ciao, robin

Ralf Mardorf

11:40 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 14:24:35 -0800 (PST), Len Ovens wrote:

...

http://motu.com/products/avb/624 It is a bit pricy though.

It's more expensive than a RME Babyface Pro. On Sun, 8 Jan 2017 23:25:16 +0100, Robin Gareus wrote:

...

On 01/08/2017 11:06 PM, Ralf Mardorf wrote:

[rocketmouse@archlinux tmp]$ ls -lh linux-4.10-rc2/sound/usb/s*

Try .../usb/*sc* to match "mixer_scarlett.cc". it's in vanilla linux since 2014: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/sound/…

Thank you! [rocketmouse@archlinux tmp]$ ls -lh linux-4.9/sound/usb/mixer_sc* -rw-r--r-- 1 rocketmouse rocketmouse 29K Dec 11 20:17 linux-4.9/sound/usb/mixer_scarlett.c -rw-r--r-- 1 rocketmouse rocketmouse 169 Dec 11 20:17 linux-4.9/sound/usb/mixer_scarlett.h [rocketmouse@archlinux tmp]$ ls -lh linux-4.10-rc2/sound/usb/mixer_sc* -rw-r--r-- 1 rocketmouse rocketmouse 29K Jan 1 23:31 linux-4.10-rc2/sound/usb/mixer_scarlett.c -rw-r--r-- 1 rocketmouse rocketmouse 169 Jan 1 23:31 linux-4.10-rc2/sound/usb/mixer_scarlett.h But sound/usb/mixer.c then still isn't the right place to add the ID. *?*

Johannes Kroll

11:29 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 23:06:54 +0100 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote:

...

On Sun, 8 Jan 2017 21:25:35 +0100, Johannes Kroll wrote:

...which would mean that you if you listen to an input, you mix it with a copy of itself delayed by the buffer size. That would definately sound weird, and could explain Ralf's experience.

No, since on Linux I used Ardour and directly connected the stereo track's outputs by jackd to playback_n. I didn't test any of the 18i20's inputs.

OK. Would still be interested in the comparison recordings you mentioned.

...

0 dB related to what? The audio files were played with an output

< 0 dBFS and perhaps at max. 0 dBFS here. Does the Focusrite mixer control the analog domain? Does it increase a digital signal that comes from the DAW by the digital domain?

I'm referring to the ALSA mixer control, which goes up to +6db.

...

On Sun, 8 Jan 2017 21:46:40 +0100, Johannes Kroll wrote:

First, look here: https://github.com/smilingthax/alsa-driver_scarlett/commit/7a77707b66ee476d… (This patch is now inside the mainline kernel; I did not write it)

It's neither part of the latest kernel supported by the rt project, nor by the current mainline rc, https://www.kernel.org/ :

I've used the Scarlett on a stock Ubuntu kernel, Mint kernel, and the AVL RT kernels. Mixer controls work on all of them, so I assumed the patch would be in mainline by now. You can try applying this patch: https://github.com/smilingthax/alsa-driver_scarlett/tree/patches

Johannes Kroll

11:43 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

...

> On Sun, 8 Jan 2017 21:46:40 +0100, Johannes Kroll wrote: > >First, look here: > >https://github.com/smilingthax/alsa-driver_scarlett/commit/7a77707b66ee476dbf2b87c18fde6a2ea4ded1ec?diff=split > >(This patch is now inside the mainline kernel; I did not write it) > > It's neither part of the latest kernel supported by the rt project, nor > by the current mainline rc, https://www.kernel.org/ :

The code is in there, but in a different file: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/sound/… Sorry for the confusion.

Ralf Mardorf

9 Jan 9 Jan

6:24 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 23:29:31 +0100, Johannes Kroll wrote:

...

Would still be interested in the comparison recordings you mentioned.

I couldn't use the same cables for the RME and the Focusrite test, the cables used for the Focusrite were better, newer, shorter. However, directly connecting the Focusrite by a balanced cable to itself, there isn't audible anymore, what still is audible, when connecting the RME and the Focusrite to other equipment by unbalanced cables. I've got an Ardour5 test as 179M tar.gz and could send it of list to some people, if you off-list ask me to do so. I'll sent it to you Johannes, you don't need to ask. More details, especially important in regards to the level of the Focusrite: [rocketmouse@archlinux ~]$ uname -rm 4.8.13-1-ARCH x86_64 [rocketmouse@archlinux ~]$ rtirq status PID CLS RTPRIO NI PRI %CPU STAT COMMAND 121 FF 90 - 130 0.0 S irq/19-ehci_hcd 126 FF 90 - 130 0.0 S irq/16-ohci_hcd 146 FF 89 - 129 0.0 S irq/17-ohci_hcd 151 FF 88 - 128 0.0 S irq/17-ohci_hcd 271 FF 85 - 125 0.0 S irq/18-snd_hdsp 32 FF 50 - 90 0.0 S irq/9-acpi 48 FF 50 - 90 0.0 S irq/8-rtc0 74 FF 50 - 90 0.0 S irq/26-radeon 111 FF 50 - 90 0.0 S irq/1-i8042 128 FF 50 - 90 0.0 S irq/14-pata_ati 129 FF 50 - 90 0.0 S irq/15-pata_ati 135 FF 50 - 90 0.0 S irq/22-ahci[000 249 FF 50 - 90 0.0 S irq/7-parport0 970 FF 50 - 90 0.0 S irq/27-enp3s0 3 TS - 0 19 0.0 S ksoftirqd/0 17 TS - 0 19 0.0 S ksoftirqd/1 [rocketmouse@archlinux ~]$ pacman -Q ardour5 jack2 ardour5 5.5-1 jack2 1.9.10.r202.gf2ece24-1 [rocketmouse@archlinux ~]$ cat .jackdrc_rme_hdspe_aio /usr/bin/jackd -dalsa -dhw:HDSPMx579bcc -r44100 -p256 -n2 [rocketmouse@archlinux ~]$ cat .jackdrc_scarlett18i20 /usr/bin/jackd -dalsa -dhw:USB -r44100 -p256 -n2 1. [rocketmouse@archlinux ~]$ $(cat .jackdrc_rme_hdspe_aio) >/dev/null 2>&1 & qjackctl & ardour5 >/dev/null 2>&1 05-weather_with_you/audio_out 1 -------> playback_1 05-weather_with_you/audio_out 2 -------> playback_2 capture_1 -----------------------------> rme unbalanced/audio_in 1 capture_2 -----------------------------> rme unbalanced/audio_in 2 unbalanced outputs were connected with unbalanced inputs hdspmixer controls were set to 0.0 and -∞ qjackctl did show no xruns 2. [rocketmouse@archlinux ~]$ $(cat .jackdrc_scarlett18i20) >/dev/null 2>&1 & qjackctl & ardour5 >/dev/null 2>&1 05-weather_with_you/audio_out 1 -------> playback_5 05-weather_with_you/audio_out 2 -------> playback_6 capture_3 -----------------------------> focusrite balanced/audio_in 1 capture_4 -----------------------------> focusrite balanced/audio_in 2 balanced outputs were connected with balanced inputs btw. the first time I used the Focusrite balanced instead of unbalanced the input pots of the focusrite were set to 0 this isn't the optimal level, but the only way to visually ensure that left-right-balance is correct qjackctl did show 0(2) xruns Regards, Ralf

Ralf Mardorf

6:45 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Mon, 9 Jan 2017 18:23:58 +0100, Ralf Mardorf wrote:

...

On Sun, 8 Jan 2017 23:29:31 +0100, Johannes Kroll wrote:

Would still be interested in the comparison recordings you mentioned.

I've got an Ardour5 test as 179M tar.gz and could send it of list to some people, if you off-list ask me to do so. I'll sent it to you Johannes, you don't need to ask.

Oops! "** Message is too big (Maximum size is 50.00MB) *** Error occurred while sending the message."

Ralf Mardorf

8 Jan 8 Jan

11:44 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 21:46:40 +0100, Johannes Kroll wrote:

...

That is, in sound/usb/mixer.c around line 2395, create another 'case' - case USB_ID(0x1235, 0x8201): /* Focusrite Scarlett 18i20 2nd-gen */ And in sound/usb/scarlettmixer.c after line 1226, insert case USB_ID(0x1235, 0x8201): info = &s18i20_info; break;

I assume now there only is the need to add it to sound/usb/mixer_scarlett.c?! [rocketmouse@archlinux tmp]$ grep USB_ID -n linux-4.10-rc2/sound/usb/mixer_scarlett.c 920: case USB_ID(0x1235, 0x8012): 923: case USB_ID(0x1235, 0x8002): 926: case USB_ID(0x1235, 0x8004): 929: case USB_ID(0x1235, 0x8014): 932: case USB_ID(0x1235, 0x800c): [rocketmouse@archlinux tmp]$ grep USB_ID -n linux-4.10-rc2/sound/usb/mixer_scarlett.c 920: case USB_ID(0x1235, 0x8012): 923: case USB_ID(0x1235, 0x8002): 926: case USB_ID(0x1235, 0x8004): 929: case USB_ID(0x1235, 0x8014): 932: case USB_ID(0x1235, 0x800c): [rocketmouse@archlinux tmp]$

Ralf Mardorf

9 Jan 9 Jan

1:09 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 20:53:37 +0100, Robin Gareus wrote:

...

OOTB it patches analog inputs through to the outputs

The 18i20 2nd Gen doesn't. I made a first recording. On Sun, 8 Jan 2017 20:44:27 +0100, Johannes Kroll wrote:

...

Don't you want to adjust gain, change routing etc? I do that often.

I'm usually doing this with an anlog mixing console or with hdspmixer for my HDSPe AIO card. The main usage of the Focusrite should be Linux PC iPad MIDI out ------> MIDI in ADAT in <------ ADAT out Sure, sometimes I want to record using the iPad, then there anyway no control is available. At 128 frames even software monitoring might be ok. It also would be nice to have an USB device for recording with Linux computers, but then again, 128 frames seem to allow software monitoring and usually anyway an anlog mixing console is involved, so routing and monitoring is no issue at all. Since the inputs aren't passed through to the outputs, I don't patch and test the Linux mixer for the 1 Gen. I consider to stay with the Focusrite for the iPad, but consider to buy a RME Babyface for the Linux PC, to completely replace the HDSPe AIO or to just buy e.g. an 18i6 or any other device, to get the ADAT inputs, I don't get by using the HDSPe, but to keep the HDSPe for analog recording and stereo mixing.

Fons Adriaensen

12:27 a.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, Jan 08, 2017 at 06:45:45PM +0100, Robin Gareus wrote:

...

flat/RMS is about -61dBFS at max gain.

According to the specs EIN is -127 dBu. I assume this is A-weighted (it doesn't say). Not really spectacular but not a disaster either, if true. Also according to the specs, max input without pad is +8.5 dBu, and gain range is 50 dB. So 0 dBFS at max gain should be -41.5 dBu at the input. Combining this with the given EIN, you should measure -85.5 dBFS(A). I guess that the -61 dBFS is even without DC filter, a classic mistake. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow)

Johannes Kroll

8 Jan 8 Jan

6:44 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

Hi Ralf, that sounds bad. I have a Scarlett 18i8 first-gen and never had such problems. Did you do any other testing in addition to subjective listening tests? Maybe hook up an oscilloscope to the outputs? There is a problem with my Scarlett where sometimes I get horrible crackling noise. Restarting JACK fixes this. Could this be your issue as well?

...

Btw. the only available control by alsamixer is "Clock Source".

This means that the ALSA mixer kernel code doesn't recognize the device yet and doesn't load the "scarlett mixer quirk". There should be a huge number of controls. Without being able to control the mixer matrix, the Scarlett is much less useful... It should be relatively easy to insert the product ID of the 2nd-gen into the alsa code, though. Have you patched & compiled a kernel before? Best, Johannes

Ralf Mardorf

7:32 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 18:44:39 +0100, Johannes Kroll wrote:

...

There is a problem with my Scarlett where sometimes I get horrible crackling noise. Restarting JACK fixes this. Could this be your issue as well?

No, I had this with the Presonus, but choosing another USB slot fixed the issue. Missing alsamixer controls aren't a problem for me. Btw. since my main interest is ADAT, I don't care that much about the sound of the analog IOs. FWIW the intensity of all LEDs is ok, something I'm missing for a lot of other modern devices. Often LEDs are much too bright, so they could be an issue, when close to the monitor.

...

It should be relatively easy to insert the product ID of the 2nd-gen into the alsa code, though. Have you patched & compiled a kernel before?

Yes. Thank you for the pointer. On Sun, 8 Jan 2017 18:45:45 +0100, Robin Gareus wrote:

...

The Scarlett series however is pretty much crap. Just terminate the input with 200 Ohm, start jnoisemeter and start wondering which award those "award winning preamps" won.

I read an old thread about this. Fons replied ironical "a marketing award" or something like this. The outputs have less gain, but aren't noisy. I didn't test the inputs. This btw. would be the only thing close to a measurement I could do: In regards to subjective audio quality I could record a WAV from the analog outputs to the analog inputs of the Focusrite and for comparison do the same with the RME card and the ADA8000. Then it would be possible to listen to the original WAV and the recorded WAVs or to take a look at a spectral analysis. Regards, Ralf

Johannes Kroll

8:44 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 8 Jan 2017 19:32:19 +0100 Ralf Mardorf <ralf.mardorf(a)alice-dsl.net> wrote:

...

On Sun, 8 Jan 2017 18:44:39 +0100, Johannes Kroll wrote:

There is a problem with my Scarlett where sometimes I get horrible crackling noise. Restarting JACK fixes this. Could this be your issue as well?

No, I had this with the Presonus, but choosing another USB slot fixed the issue. Missing alsamixer controls aren't a problem for me.

Don't you want to adjust gain, change routing etc? I do that often.

...

On Sun, 8 Jan 2017 18:45:45 +0100, Robin Gareus wrote:

The Scarlett series however is pretty much crap. Just terminate the input with 200 Ohm, start jnoisemeter and start wondering which award those "award winning preamps" won.

I read an old thread about this. Fons replied ironical "a marketing award" or something like this. The outputs have less gain, but aren't noisy. I didn't test the inputs.

I agree the inputs aren't great. They are pretty noisy. Probably a company-internal award - have 3 people design a preamp and "award" the one with the best price-performance ration, or something like that. ;) Pure marketing. I didn't have quite as bad an impression as you describe though.

...

This btw. would be the only thing close to a measurement I could do: In regards to subjective audio quality I could record a WAV from the analog outputs to the analog inputs of the Focusrite and for comparison do the same with the RME card and the ADA8000. Then it would be possible to listen to the original WAV and the recorded WAVs or to take a look at a spectral analysis.

So you would mix the input & output results together... But it would be possible to upload the files somewhere. I would be interested to see this, as I don't have another interface comparable to the Focusrite.

Ralf Mardorf

11:19 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

On Sun, 2017-01-08 at 20:44 +0100, Johannes Kroll wrote:

...

Or I record from the analog outputs of a CD player, but than there would be no neutral reference available. I don't want to spend time with doing this at the moment, but I could do it someday and send the files off- list. I'm neither subscribed to a site were it's possible to upload audio files, nor is it allowed in Germany to upload the music I would use as reference.

Ralf Mardorf

12:58 p.m.

New subject: Sound quality of the Focusrite Scarlett 18i20 2nd Gen

PS: Focusrite also claims that the device has got a power-up delay. That's at least untrue for the headphone output. I don't want to test it for the other outputs.

2880

days inactive

2882

days old

linux-audio-user@lists.linuxaudio.org

Manage subscription

30 comments

7 participants

tags (0)

participants (7)

Fons Adriaensen
Georg Krause
Johannes Kroll
Len Ovens
Ralf Mardorf
Robin Gareus
tom haddington