A new star in 2019 : Innuos Statement Server

[asiufy]

Sorry Alex, but IMHO users of a product that has no documentation at all are Beta users. The machine is plug and play if you use it just to perform the very basic functions - not my case.

And yes I tried engage/disengage the software in two machines - at a point everything hanged-up and I had to shoot down and power-on everything to have music again. Not my idea of a fast comparison.

As you I was told that they are working on a new website, with a dedicated documentation and support area. I am sure that they will have it, but I am commenting on the current status of the product. Surely using advanced google and looking for discussions on the Zen machines I am solving most of the problems I find. And I do not care about comparisons with the Aurender, I compare it with an installation of Roon on my Windows10 server.

But yes, the most important part, it sounds really excellent in the Vivaldi using the streamer output.

Francisco,

"Beta user" mean people are using pre-release software, or software that's not fully realized and/or tested. That is simply not true with the Statement, or any Innuos product, so that's why I opposed your use of this term. This is a widely used term in software (a field which I used to work for many years), and while it does not have a negative connotation per se, it can be construed as negative, thus my objection to its use here.

I've been through a lot of server products, and the recurring theme was: not existent or very poor documentation. It's way more common than you think, from the $1000 machine all the way to the many $$$ ones. Since I know my way around these things, and our clients will always have service, I don't consider that a negative, at all.

Now, support from the manufacturer is most important in these cases, and Innuos (along with other brands, like Aurender) have been exemplary. You shouldn't need to Google for help if you have a dealer that's taking care of you, or if you purchase direct from a manufacturer, his own direct support.

In the store, we had at some point FOUR Roon Cores, each in a different machine, and I'd switch between them to do comparisons. FOUR. Using the same license. Obviously Roon wouldn't let more than one run simultaneously, but it's smart enough to ask me to disable the other licenses. And that I did. Back and forth multiple times.

For more "normal" comparisons, between two Roon Cores, such as your case, perhaps what you can do is keep an iPad connected to the Innuos Roon Core, and a 2nd device, an iPad, a phone, a laptop, connected to the Windows PC Roon Core. That way, you keep both machines as Roon Core, and use different devices to control them, engaging/disengaging them as needed.

I don't recommend you re-purpose the Windows PC as a Remote to the Innuos Roon Core, as it'll just take more of your time to switch back/forth between them.

I hope this helps
enjoy!
Alex

 

[microstrip]

Excellent, let me see if I can help!


The whole "my.innuos.com" thing is an ingenious way to locate your device, *from another device on your local network*. What does that mean? That it'll only work if whoever's accessing it already has access to your local network.
Want to give it a try? On your mobile, while connected to your local network (via Wi-Fi), "my.innuos.com" will work just fine. Now, disable Wi-Fi, and just use your phone's data network. That way you will NOT be on your local network anymore. Then try "my.innuos.com" again. Your Statement won't be visible anymore, even though it's still there, operational, maybe even playing music. It's just that your phone is not on your local network anymore, and so the Innuos is off-limits to it.
As it is, there is no access control built into the Innuos software, and that's for simplicity' sake. That's part of what makes it easy to set up and use (one less password).

Basically you are saying that anyone one with access to the Wi-Fi or home network can access and control the server and there is no way to protect it.

What do you mean here? Roon has a myriad of options and configurations. The first thing I'd do is disable all sorts of DSP within Roon, by going into Settings > Audio > <locate your dCS here> > gears > DSP Engine.

I am addressing the three ways of using Roon found in the Statement configuration menu, not the Roon configuration.

If you've already purchased them in FLAC, go ahead and convert them using an app on your Mac or PC, like dBpowerAmp or XLD, then store the AIFF files (preferable to WAV) and ditch the FLAC.
If you're buying new albums, I recommend you purchase either WAV or AIFF files, not FLAC.
And, as I said, when ripping, go with the WAV option with the Innuos software.

Hope this helps!

I would expect that such expensive server would not be sensitive to FLAC or WAV format. All my HI-Rez library is FLAC and I do not consider converting them one by one. I will have to look for some utility that converts all the disk FLAC content to WAV.

Thanks!

 

[asiufy]

Basically you are saying that anyone one with access to the Wi-Fi or home network can access and control the server and there is no way to protect it.

If this is a concern to you, access control can be implemented in your own local network, in the router. For example: you can configure it so the Innuos can only "talk" to the IP of your DAC and your iPad.

I am addressing the three ways of using Roon found in the Statement configuration menu, not the Roon configuration.

Use Roon Core.

I would expect that such expensive server would not be sensitive to FLAC or WAV format. All my HI-Rez library is FLAC and I do not consider converting them one by one. I will have to look for some utility that converts all the disk FLAC content to WAV.

Thanks!

This has ZERO to do with the server, and it's jus an inherent flaw of the FLAC format. The server just pumps the bitstream to the DAC.
The bits need to be "unpacked" (by Roon) from FLAC to WAV, in order to be sent to the DAC. This extra "unpacking", done via software, is what's responsible for the loss in performance. That's why disabling the DSP code in Roon makes such a massive difference, even if all its options are switched off.
Again, this is not a fault of "such expensive server", and it'd be nice if you refrained from assuming fault with the product, without proper knowledge of what's going on.
If anybody claims their server is not "sensitive to FLAC or WAV format", be suspicious, and I'd be curious to hear one that isn't, as I haven't met one yet.
I believe someone has posted a thread here the other day about doing just such conversion, FLAC to another format. If you have a Mac, XLD is a good app to do such conversion. For PCs, there are many options as well, like dBPowerAmp.

 

[microstrip]

If this is a concern to you, access control can be implemented in your own local network, in the router. For example: you can configure it so the Innuos can only "talk" to the IP of your DAC and your iPad.

Almost all devices connected to WiFi, even my NAS, have a password control and can be safely used in a network. I am not going to fiddle with the router just for listening to music ...

This has ZERO to do with the server, and it's jus an inherent flaw of the FLAC format. The server just pumps the bitstream to the DAC.
The bits need to be "unpacked" (by Roon) from FLAC to WAV, in order to be sent to the DAC. This extra "unpacking", done via software, is what's responsible for the loss in performance. That's why disabling the DSP code in Roon makes such a massive difference, even if all its options are switched off.
Again, this is not a fault of "such expensive server", and it'd be nice if you refrained from assuming fault with the product, without proper knowledge of what's going on.
If anybody claims their server is not "sensitive to FLAC or WAV format", be suspicious, and I'd be curious to hear one that isn't, as I haven't met one yet.
I believe someone has posted a thread here the other day about doing just such conversion, FLAC to another format. If you have a Mac, XLD is a good app to do such conversion. For PCs, there are many options as well, like dBPowerAmp.

Sorry Alex, as far as I know this is an open forum where we freely debate our opinions on products, as far as I know you are not a WBF moderator. Considering that both formats have the same bit exact content I expect that a top server will sound the same on both formats, and I will express my opinions freely on WBF. Needing to change the format of the files to get the best sound is surely an issue with the product and should be discussed - with this or any other server.

And yes, someone also questioned about a way of doing these conversions in batch and all he got was vague answers about programs doing conversions, not a defined way of doing it. It is why I asked again, hoping someone would give a solid advice. I do not see anyone wanting to do it on thousands of recordings one by one.

 

[asiufy]

Francisco,

I'm just saying that I have not heard a server where a FLAC file didn't sound different from a WAV yet, and, as such, placing blame on the "expensive server" might be wrong, and ask you to reflect on that.

Both formats do NOT have the exact same bit content, as they are stored. As I said, FLAC is "packed", and needs "unpacking" in order to obtain the exact same bit content as the WAV/AIFF. That "unpacking" process is likely responsible for the diminished performance. And besides, I thought we were way past the "bits are bits" in this forum.

It'd be interesting for you to actually carry that comparison, and let us know what you hear, as obviously my opinion is just that, an opinion, and you might find the two files indistinguishable, which will render this whole point moot. And, depending on your finding, converting FLAC to WAV/AIFF might be in your best interest, regardless of which server you use.

I don't agree it's an issue with the server, and I never placed blame on this or that server, whenever I heard differences between FLAC and WAV/AIFF. Actually, I consider it a BENEFIT that I can hear a difference, as the hardware is able to show exactly the impact of the software in the overall performance. I've played with other "expensive servers" where the sound changed DRAMATICALLY (for the better) just by disabling certain things within the operating system. Is this a fault of the server? I don't think so. Same thing with Roon. It ships with DSP processing on by default, and it does affect performance. It's a piece of software, affecting the performance of a server. Is this a fault of the server? I don't think so.

I do remember posting in the other thread that I use and recommend XLD for the conversion task, and it's a very simple, straightforward process. Just drag the folder(s) where the FLAC files are, and XLD will convert them to a different format, in the original folder(s), and delete the FLAC files (optionally). You don't have to do it one by one, as you can drag multiple folders. I don't know what better advice can be given. I remember somebody else also recommended using JRiver, as it's also able to automatically do the work.

 

[microstrip]

Francisco,

I'm just saying that I have not heard a server where a FLAC file didn't sound different from a WAV yet, and, as such, placing blame on the "expensive server" might be wrong, and ask you to reflect on that.

Both formats do NOT have the exact same bit content, as they are stored. As I said, FLAC is "packed", and needs "unpacking" in order to obtain the exact same bit content as the WAV/AIFF. That "unpacking" process is likely responsible for the diminished performance. And besides, I thought we were way past the "bits are bits" in this forum.

It'd be interesting for you to actually carry that comparison, and let us know what you hear, as obviously my opinion is just that, an opinion, and you might find the two files indistinguishable, which will render this whole point moot. And, depending on your finding, converting FLAC to WAV/AIFF might be in your best interest, regardless of which server you use.

I don't agree it's an issue with the server, and I never placed blame on this or that server, whenever I heard differences between FLAC and WAV/AIFF. Actually, I consider it a BENEFIT that I can hear a difference, as the hardware is able to show exactly the impact of the software in the overall performance. I've played with other "expensive servers" where the sound changed DRAMATICALLY (for the better) just by disabling certain things within the operating system. Is this a fault of the server? I don't think so. Same thing with Roon. It ships with DSP processing on by default, and it does affect performance. It's a piece of software, affecting the performance of a server. Is this a fault of the server? I don't think so.

I do remember posting in the other thread that I use and recommend XLD for the conversion task, and it's a very simple, straightforward process. Just drag the folder(s) where the FLAC files are, and XLD will convert them to a different format, in the original folder(s), and delete the FLAC files (optionally). You don't have to do it one by one, as you can drag multiple folders. I don't know what better advice can be given. I remember somebody else also recommended using JRiver, as it's also able to automatically do the work.

By "bit exact" we usually do not refer to bits in the files, but to the bits being input at the DAC. It means that there is no content manipulation and the DAC sees the same bits.

The DAC will get the same bits either using a FLAC or a WAV. But as you say it will sound different. I do not consider that as a benefit - IMHO the server should deal with this issue. And we can be sure that soon someone will have a server that does it. It is the only thing I am addressing and bugs me, I will move the question to another thread in WBF. And yes, the problems you refer with other servers are not benefits!

You are mixing a lot of different issues, that are not related to the question I raise. We can forget Roon and DSP, they are not related to this question.

And no, we are past the "bits are bits" in WBF - but we are still looking for proper explanations why "bits are not bits". Some of us like to learn about digital audio and consider that raising questions is a way of doing it.

BTW, thanks for the suggestions, I had researched them - but XLD is a Mac utility, I am an Windows user, and besides it seemed to me it will not do batch conversions - it would need complicated scripts. JRivers ask us to select the tracks we want to convert or create playlists - it does not carry full conversion of the library in one step. At this point most people have large FLAC libraries, I would be happy to buy such tool if it existed.

 

[asiufy]

But that's exactly what I pointed out, in order to get "the exact same bits", there's processing involved when using one of the formats, while there's no processing with the other. And that's the explanation you wanted, and the potential source of the performance delta.
Perhaps one day the hardware can be made immune to those software differences, that little extra processing. But, to this day, I don't know one. And no, I'm not mixing different issues. Software is software, and its contribution to overall performance is what we're talking about here. DSP or simple FLAC unpacking requires extra software, and that, AFAIK and IMHO, has a penalty in performance, regardless of the hardware it's running on.
Roon itself has a sound. There are some server manufacturers that won't use Roon, as they consider JRiver to be a better sounding solution.
TBH, the way I see it, there will always be a difference in performance with different software. There will never be a piece of hardware that will make software differences irrelevant, as far as sound quality/performance. That's why it's so important for the *software* to be optimized at this point. Assuming (like I believe) the hardware can't be made immune to whatever the software's doing, might as well make the software impose as little penalty in SQ as possible!
For Windows, I've used this app once, a few years ago, and it did the job:

https://www.dbpoweramp.com/dmc.htm

It does say it'll do "batch convert".

 

[microstrip]

But that's exactly what I pointed out, in order to get "the exact same bits", there's processing involved when using one of the formats, while there's no processing with the other. And that's the explanation you wanted, and the potential source of the performance delta.

If a little bit of extra light processing - and we are not addressing heavy computation, such as filtering or DSP, just changing the format of uncompressed files - affects sound quality, what can we expect from running the Roon core in the background? Or ripping in the background? I have tried ripping while listening and could not detect any apparent difference in sound quality , I will have to listen with more care - it was a quick test.

Perhaps one day the hardware can be made immune to those software differences, that little extra processing. But, to this day, I don't know one. And no, I'm not mixing different issues. Software is software, and its contribution to overall performance is what we're talking about here. DSP or simple FLAC unpacking requires extra software, and that, AFAIK and IMHO, has a penalty in performance, regardless of the hardware it's running on.
Roon itself has a sound. There are some server manufacturers that won't use Roon, as they consider JRiver to be a better sounding solution.
TBH, the way I see it, there will always be a difference in performance with different software. There will never be a piece of hardware that will make software differences irrelevant, as far as sound quality/performance. That's why it's so important for the *software* to be optimized at this point. Assuming (like I believe) the hardware can't be made immune to whatever the software's doing, might as well make the software impose as little penalty in SQ as possible!

I am more optimist than you. I accept that sometime someone will create a black box with with one input and one output that will "clean" any exact digital data from server interference. But I am not sure we will like the sound it produces more than the sound of a SACD spinner ...

For Windows, I've used this app once, a few years ago, and it did the job:

https://www.dbpoweramp.com/dmc.htm

It does say it'll do "batch convert".

I am a registered user of dbPoweramp. It does not do what I want (and many other people require). BTW, do you know what database is used by the Statement to check for ripping accuracy? I find contradictory information in the net.

As far as I know dBpoweramp supports uncompressed FLAC - equivalent to WAV with better tagging. Do these files also sound inferior to WAV in the Statement?

Thanks for trying to answer my questions. Servers are an hard subject - I have learned that they are supposed to be listened, not to be inquired ... And the most critical question was concerning playback sound quality using NAS or the statement SDD!

 

[Taiko Audio]

That's a great start.
How about pictures of internals?
How about pictures of an actual EVO?
How about being able to order on the web shop?
Options and costs?

Thank you Koegz for the internal photo's.
We can enable webshop purchases if you really want to but we prefer to quote by phone and/or email for products in this pricerange.
I feel very uncomfortable discussing the SGM in this thread, so how about we continue in this thread which also contains actual EVO pictures:
The 1st SGM Server EVO Edition in Hong Kong

 

[asiufy]

If a little bit of extra light processing - and we are not addressing heavy computation, such as filtering or DSP, just changing the format of uncompressed files - affects sound quality, what can we expect from running the Roon core in the background? Or ripping in the background? I have tried ripping while listening and could not detect any apparent difference in sound quality , I will have to listen with more care - it was a quick test.

I know. And that's where it gets tricky, because you can't solve the problem by throwing more CPU power at it, because it's not how heavy/intensive the process is that matters, but the nature of it (apparently). And again, that's why it's all in the software, both the playback, as well as the foundational (operating system) software.

I am more optimist than you. I accept that sometime someone will create a black box with with one input and one output that will "clean" any exact digital data from server interference. But I am not sure we will like the sound it produces more than the sound of a SACD spinner ...

It could be! What's important is to realize we're still in the very early stages of development of this technology. There's so much left to learn still, in a field where most companies never had much experience in (software).
Like anything in audio, there will be many ways to skin that particular cat, with the usual different results, and some will prefer this way, and others will prefer that other way, etc. That's why you see folks approaching it with massive hardware, and others with minimal hardware, custom Linux versus tweaked Windows, etc.

I am a registered user of dbPoweramp. It does not do what I want (and many other people require). BTW, do you know what database is used by the Statement to check for ripping accuracy? I find contradictory information in the net.

As far as I know dBpoweramp supports uncompressed FLAC - equivalent to WAV with better tagging. Do these files also sound inferior to WAV in the Statement?

Yes. The Innuos itself rips to uncompressed FLAC, and Innuos' (the company) position is that they don't sound different. My ears tell me otherwise

I don't know which database they use for accuracy. But if you contact support@innuos.com, they'll tell you I know that I've been pleasantly surprised by their rips, specially with the odd scratched/slightly defective CD...

Oh, shame about dBpowerAmp, I seem to remember it worked fine for what I needed it then...

Thanks for trying to answer my questions. Servers are an hard subject - I have learned that they are supposed to be listened, not to be inquired ... And the most critical question was concerning playback sound quality using NAS or the statement SDD!

I know, as I said, this is such a new field, that I understand if people can't help but be intrigued and have questions. But at the end of the day, it needs to sound good!

Regarding the NAS x internal drive playback, I will be able to give this a try soon enough, and will report back. I'm curious how it'll be with the Statement...

 

[Taiko Audio]

@microstrip:

We have done quite a bit of experimenting on WAV vs (compressed) FLAC sound quality differences.
The obvious difference is: uncompressed WAV causes more disk activity (larger file), where decompressing uses more CPU.
On fast CPU servers with SATA SSD storage compressed FLAC usually sounds better then uncompressed WAV (or any uncompressed vs compressed format), on slow CPU's it's the other way around.
On fast CPU servers with direct CPU attached PCIe NVMe storage the difference becomes undetectable.
FWIW

 

[Kingsrule]

I asked about batch conversion of FLAC to WAVE and converted a dozen or so files
After a lot of back and forth I agree with Emile, FLAC sounds better most of the time, otherwise its a draw.

 

[asiufy]

Interesting! That's why I'm curious to hear what Francisco's thoughts are... I've never, once, had a FLAC sound better. Same or similar, yes. Better, never.

 

[Kingsrule]

Alex

How about suggesting a track or 2 that we all probably have and then we can compare and talk about the differences, if any

 

[asiufy]

I'm currently in Brazil, thus unable to do any comparisons of my own! That's why I requested Francisco's feedback, and now yours too...
But I'll be back to the store in 10 days, so I will get around to that, as well as comparing sound from Statement's internal drive x NAS.

 

[CKKeung]

... comparing sound from Statement's internal drive x NAS.

The internal drives inside Statement have much advantages over those stored in a NAS.
Unless a lot of tweakings have been applied to the NAS/network path, my guess is that the internal drives will win easily.
There are too many factors in the NAS/network path that can deteriorate the final sonic performance.
This is also the reason why streaming is the future but currently it cannot fight with internal files in terms of sonic quality.

 

[Taiko Audio]

The internal drives inside Statement have much advantages over those stored in a NAS.
Unless a lot of tweakings have been applied to the NAS/network path, my guess is that the internal drives will win easily.
There are too many factors in the NAS/network path that can deteriorate the final sonic performance.
This is also the reason why streaming is the future but currently it cannot fight with internal files in terms of sonic quality.

I agree, this matches our experience. Optical fibre networking helps streaming (helps local storage playback aswell though), but still internal file storage usually wins. If you bypass the internal SATA controllers + motherboard chipset and use direct CPU attached storage the difference becomes even larger. You need modern Xeon cpu's for that though which have integrated disk controllers and PCIe x16 slots which are directly attached to the CPU, not by means of PCI multiplexers connected to the motherboard chipset.

 

[Kingsrule]

How would u get 16Tb of music into internal storage?

 

[Taiko Audio]

How would u get 16Tb of music into internal storage?

4*4TB ssd's or 8*2TB m2 PCIe drives.

 

[Kingsrule]

And u can fit this many drives in your server?