Post Reply 
dsf to wav : what CPU power ?
08-12-2016, 20:56 (This post was last modified: 08-12-2016 21:47 by chip.)
Post: #41
RE: dsf to wav : what CPU power ?
Hi all,
Sonore ask me to perform some additional test so I just posted this following on their forum.

Hi,
Music Library
MinimServer 0.8.4 update 88
MinimStreamer 0.5.2
stream.converter ffmpeg314
Bubble UPnP Server Version 0.9-update21

Control Point
Samsung Galaxy Tab S2 tablet
BubbleUPnP version 2.8.3.1

Renderer
Sonore Signature Rendu shown as Audio Render-143 (OpenHome)

DAC
PS Audio DirectStream DAC

NativeDSD downloads used for testing
Sarah Vaughn - Live at Rosy's
DSD64 - 09 A-Tisket, A-Tasket.dsf
DSD128 - 09 A-Tisket, A-Tasket.dsf

MinimStreamer stream.transcode values and results from 2016-12-08 testing requested by Sonore

MinmStreamer stream.transcode value <empty, no values> and restart server
DSD64 file
PS Audio displays - Input: DSD Rate: DSD64 1 bit
BubbleUPnP displays - DSF | 2822.4 kHz | 1 bit
DSD128 file - small spike before playback
PS Audio displays - Input: DSD Rate: DSD128 1 bit
BubbleUPnP displays - DSF | 5644.8 kHz | 1 bit
This plays back as expected with only a small spike when changing from DSD64 to DSD128.

Clear BubbleUPnP playlist
MinmStreamer stream.transcode value "dsf:dopwav" and restart server
DSD64 file - small spike before playback, switching from DSD128 to DSD64
PS Audio displays - Input: DOP Rate: DSD64 1 bit
BubbleUPnP displays - WAV | 176.4 kHz | 24 bits
DSD128 file - small spike before playback, switching from DSD64 to DSD128
PS Audio displays - Input: DOP Rate: DSD64 1 bit
BubbleUPnP displays - WAV | 352.8 kHz | 24 bits
Playback of this track sounds like it is at half speed
The PS Audio does not playback 352.8 kHz files - incorrect see next post
The time slider on BubbleUPnP moves to 100% but timer keeps counting until 3:39 but the track is only 1:49 long
The DSD64 file plays back as expected but the DSD128 file does not playback properly

Clear BubbleUPnP playlist
MinmStreamer stream.transcode value "dsf:wav24;176" and restart server
DSD64 file
PS Audio displays - Input: PCM Rate: 176.4k 24 Bits
BubbleUPnP displays - WAV | 176.4 kHz | 24 bits
DSD128 file
PS Audio displays - Input: PCM Rate: 176.4k 24 Bits
BubbleUPnP displays - WAV | 176.4 kHz | 24 bits
Both playback as expected without spikes


Is this the test you wanted me to perform?
Do my results help you identify the issue?

Please advise,
Chip
Find all posts by this user
Quote this message in a reply
08-12-2016, 21:41
Post: #42
RE: dsf to wav : what CPU power ?
Hello,
I just went over to the PS Audio website and it says the DirectStream DAC will play:
Sample Rates
I2S,S/PDIF, and USB -- 44.1kHz to 352.8kHz 16bit, 24bit, DSD 64, DSD 128
TOSLINK – 44.1kHz to 96kHz 16bit, 24bit
XLR (AES/EBU) -- 44.1kHz to 192kHz 16bit, 24bit, DSD 64
I am using the I2S interface from my SSR to the DS DAC so my statement in the previous post:
The PS Audio does not playback 352.8 kHz files
is wrong!

Now I am more confused about why the DSD128 file did not playback correctly.
I then tried the file again and had the same results.

Sorry about my confusion,
Chip
Find all posts by this user
Quote this message in a reply
09-12-2016, 10:30
Post: #43
RE: dsf to wav : what CPU power ?
Hello,

Just received this from Sonore to explain the problem.

"For DSD128 to stream via DoP you need 352.8 kHz support. The SSR does not currently support 352.8 kHz or 384 kHz. It is possible to update it to support up to 384 kHz, but it's expensive and we have not offered the update. The PS audio DAC supports 352.8 kHz, but not 384 kHz."

So it looks like the Sonore can not process DSD128 to dopwav since it creates a 352.8 kHz data stream.

Thanks all,

Chip
Find all posts by this user
Quote this message in a reply
09-12-2016, 18:47
Post: #44
RE: dsf to wav : what CPU power ?
Hi,
I asked Sonore if they could add the audio/x- MIME types but they are not sure but did say they would try to look into it.
from Sonore "It might be possible, but I have to look into it."

I would be interested in what MinimStreamer and BubbleUPnP might be able to do so if you guys have an idea I would be a willing tester for you.

Thanks for all your assistance and the education I have received from you!

Please let me know what you think,
Chip
Find all posts by this user
Quote this message in a reply
09-12-2016, 22:30
Post: #45
RE: dsf to wav : what CPU power ?
It is not possible for MinimStreamer to change the MIME types for the streams that it sends because these types are required by some renderers and each stream can only have a single MIME type.
Find all posts by this user
Quote this message in a reply
10-12-2016, 11:13
Post: #46
RE: dsf to wav : what CPU power ?
I understand what you are saying now that you have given me such a thorough education. Thanks again for that.

I think the best option would be for Sonore to advertise compatibility with the audio/x- MIME types since that would insure compatibility with programs that use either. I am anxious to hear if they are willing to do that.

The second best option would be for BubbleUPnP to recognize the audio/x- MIME types as an alias as mentioned by the developer.
from bubbleguum:
"For some mime-types, BubbleUPnP add missing aliases (eg if a renderer advertises audio/flac, BubbleUPnP will add audio/x-flac. But not for DSD mime-types where BubbleUPnP only recognizes the x- versions. I could probably add the non x- versions as aliases as I don't think it would cause any problems."

Do you think my ideas are correct or did I miss something?

Your humble student,
Chip
Find all posts by this user
Quote this message in a reply
10-12-2016, 11:49
Post: #47
RE: dsf to wav : what CPU power ?
You have not missed anything and I think this is a very good summary.
Find all posts by this user
Quote this message in a reply
10-12-2016, 12:08
Post: #48
RE: dsf to wav : what CPU power ?
(10-12-2016 11:13)chip Wrote:  I understand what you are saying now that you have given me such a thorough education. Thanks again for that.

I think the best option would be for Sonore to advertise compatibility with the audio/x- MIME types since that would insure compatibility with programs that use either. I am anxious to hear if they are willing to do that.

The second best option would be for BubbleUPnP to recognize the audio/x- MIME types as an alias as mentioned by the developer.
from bubbleguum:
"For some mime-types, BubbleUPnP add missing aliases (eg if a renderer advertises audio/flac, BubbleUPnP will add audio/x-flac. But not for DSD mime-types where BubbleUPnP only recognizes the x- versions. I could probably add the non x- versions as aliases as I don't think it would cause any problems."

Do you think my ideas are correct or did I miss something?

Your humble student,
Chip


I will add the aliases. It should make BubbleUPnP pickup the DSF/DSD stream when playing to the Sonore as an UPnP AV renderer.
But Sonore should really add the correct x- mime-types, instead of using invented ones.
Find all posts by this user
Quote this message in a reply
10-12-2016, 12:22 (This post was last modified: 10-12-2016 12:22 by bubbleguuum.)
Post: #49
RE: dsf to wav : what CPU power ?
I will also make BubbleUPnP recognize audio/x-dsd (and the alias audio/dsd). This is purely for display in the Now Playing screen for the codec which will display as "DSD".
Does MinimServer ever send this mime-type and in which case ?
Find all posts by this user
Quote this message in a reply
10-12-2016, 12:39 (This post was last modified: 10-12-2016 13:09 by simoncn.)
Post: #50
RE: dsf to wav : what CPU power ?
(10-12-2016 12:22)bubbleguuum Wrote:  I will also make BubbleUPnP recognize audio/x-dsd (and the alias audio/dsd). This is purely for display in the Now Playing screen for the codec which will display as "DSD".
Does MinimServer ever send this mime-type and in which case ?

The history of audio/dsd is that JRiver sends it for both DSF and DFF, leaving it to the renderer to work out which of these it is. Some other products (renderers and possibly servers) have imitated this. MinimServer always uses the specific audio/x-dsf and audio/x-dff types. never audio/x-dsd.

For aliasing purposes in the control point, I think the following should work:

1) A renderer that claims to support audio/dsd or audio/x-dsd should be considered compatible with the following server types:

audio/dsd or audio/x-dsd
audio/dsf or audio/x-dsf
audio/dff or audio/x-dff

This is not required for the Sonore but is required for renderers such as the Melco N1, which advertises audio/dsd and audio/x-dsd only.

2) A renderer that claims to support audio/dsf or audio/x-dsf should be considered compatible with the following server types:

audio/dsf or audio/x-dsf

3) A renderer that claims to support audio/dff or audio/x-dff should be considered compatible with the following server types:

audio/dff or audio/x-dff
Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump:


User(s) browsing this thread: 1 Guest(s)