Naim UPnP servers: do they always transcode to LPCM when serving Naim streamers ?
Posted by: m0omo0 on 27 February 2012
Good morning every one,
I admit not having re-read all the doc or thoroughly searched the Forum, so maybe a daft question: do the Naim UPnP servers (mainly HDX/UnitiServe, but maybe NS0x as well) transcode non-WAV files to LPCM (as WAV I guess ?) when serving Naim streamers ?
Or, phrased differently, do Naim streamers only see uncompressed LPCM streams when served from Naim UPnP servers, regardless of the original file format ? (The LPCM stream not always being lossless of course, depending on the file format.)
The 'Specification' tab on the HDX-SSD web page (as an example) says, under 'Audio Inputs' (?):
Transcoding, proxying UPnP server. DLNA device compatible
but not sure what 'transcoding, proxying' means, and why it's an 'audio input' to the HDX.
Maybe an easy one for Phil after this rather draining week-end.
Thanks
Maurice
Hi Maurice,
On the HDX, go to System -> UPnP Settings -> Compatibility
This allows you 3 options:
Native (no transcoding0
WAV (transcoded, maintains sample rate and bit depth)
CD 44.1kHz / 16 bit (transcoded)
So you have a choice, have a look at the reference manual, on page 15.
Hope this helps.
Rich
Thank you Rich !
You're right, it's on the manual... RTFM but I think I had an old version prior to UPnP. Now the latest version is duly downloaded for reference.
I don't own any Naim server, I was just asking as a matter of interest and because I may be comtemplating the product.
Funny Naim does recommend the 'Native' setting !? Wouldn't 'Convert to WAV' offload the streamer from some additional conversion work (and then computational load detrimental to the electrical noise floor) ?
Maurice
Hi,
No Native does no conversion, therefore, should have less overhead on the server. I've never noticed any degradation in the performance of the digital out of the server when it is also streaming to a different room. (I haven't tested it though). This is the setting I use the most.
I've used the CD setting occasionally to stream to my TV, which doesn't support flac or higher bit rates. The Naim UPnP server software unfortunately doesn't support individual profiles for renderers (hint hint Naim, this would be a nice to have feature), so what ever setting you choose applies to all your devices.
I used to use the WAV setting when I was streaming Hi-Res files to PlugPlayer on my Mac (it does not support Flac), however, I now have a Qute in my study, so no need for PlugPlayer
I guess some people may want to drop to CD if you're using wireless and your streaming hi-res files.
I think the HDX is a great product, especially with the DAC. It seems to not be getting much press at the moment, with all the new streaming products. If I had my time again I would probably buy the NS01 as I bought a DAC at the same time. Then spend the money I saved on an ipad, as this is a must to control the server products.
Hope this helps
Rich
Rich, thank you for your detailed reply.
Yes, on the server, I can understand that. But, if you're streaming to a Naim streamer, why would you let the streamer do the conversion while it can be done on the server ?
I'm curious why Naim recommend this. My reasoning would be to relieve the streamer from as much work as possible. As I understand it, for a streamer processing the stream plus converting whatever format to LPCM needs more computational work than just processing a WAV stream, but I guess I'm missing something here ?
I see what you mean, I assume this is because the HDX was designed to rip CDs, so anything it rips is stored internally as WAV files. In this scenario no extra processing is done by server or renderer, and the reason Naim recommend Native. However, some users like myself also serve other file formats. Also, the HDX was designed/marketed as a source, so the reason processing load could be an issue.
There would be extra processing, but whether it is audible would depend on the system. There are quite a few threads on the forum about FLAC vs WAV and sound quality.
The settings give you flexibility, so you don't have to use the recommended settings, choose what ever fits best. A good dealer should be able to lend you a server for a weekend, so try out the different settings and let your ears decide which is best in your system.
Rich
In UK perhaps, but where I live, no way. Home trial is not possible, except for a cable or two. I was very very lucky to be able to listen to my speakers at home in my room, but it's an exception and not the rule (at least in my experience, not that I buy often anyway).
I understand the HDX is a source as well, but he UnitiServe manual says the same (it can be a source, but is not really advertised as one as you need additional hardware):
Native: Streams audio files in the same format as they are stored.
CD 44.1kHz/16 bit: Converts streamed files to CD format.
Convert to WAV: Converts files to WAV format while retaining their native sample rate and bit depth.
Note: Native stream format offers the highest potential audio quality followed by Convert to WAV and finally CD 44.1kHz/16 bit. The appropriate choice will depend on the capability of the playback hardware. Select Native initially and check that the playback hardware operates correctly with all stored file formats. If it fails to operate at any time select Convert to WAV. If it still fails occasionally select CD 44.1kHz/16 bit.
Naim just stated once again in the NDS white paper that, to their ears, FLAC decompression adverts sound quality negatively compared to WAV. So it seems to be against common sense to recommend streaming native format (except for WAV files of course) instead of converting to WAV. Or at least doing so when serving their streamers (YMMV for other streamers, as they say in the note above).
Of course, if I happen to own such a server one day as well as a Naim streamer, I will try for myself and see what I prefer. I'm just trying to understand the rationale here.
Maurice
Theres less work to be done on the serverconverting from wav as opposed to converting from flac. There was this other thread about someone having troble eith one particular flac file. One of the R&D folks explained that when the number of processing cycles go up, o does power consumption, leading to a degradtioN. N sound quality. I still store my files in flac, but havecompression set to 0! In effect using flac just for its tagging coabilities.
That's what I'm saying, except that I don't mind about the server, I'm wondering why it is best to let the decompression take place on the s t r e a m e r side !
Maybe the HDX and the NDX folks are different teams, and both care about what makes their stuff sound the best? ;-)
No idea. Maybe it's the balance between giving the network card more to process vs. the cpu itself?
Maybe the opposite is true when you're using a wired network instead of a wireless?
There is a very interesting strand over on the Vortexbox forum that seems like it might just bang a few nails in the coffin of the wav v flac debate and prompted by someone suggesting that Naim stuff works best with wav - never a good idea with the Linux boys ....
Phew
Tog