Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[thediscman]

"Fixed: The player could inherit the sample rate of the last file for the next file."

Now, in real-time DSD output mode, the Audio Path doesn't indicate that it's resampling anything at all. Things seem to be okay when I'm listening to PCM files, but if I start with PCM and switch to DSD, the music plays in slow motion. If I start with DSD and switch to PCM, the music plays at hyperspeed.

What needs to happen is that every track that is not natively at the selected output rate should be resampled to the selected output rate.

• If DSDx1 has been selected, every track that is not natively 352.8k needs to be resampled to 352.8k.
• If DSDx2 has been selected, every track that is not natively 705.6k needs to be resampled to 705.6k.
• If DSDx4 has been selected, every track that is not natively 1411.2k needs to be resampled to 1411.2k.

So, for example, if I have output set to DSDx1 and I'm playing a DSDx1 track, no resampling would be required. Any other resolution would need to be upsampled or downsampled to 352.8k.

And I really, really liked seeing the resampling in the Audio Path. I hope that returns.

Thank you.

[6233638]

I'm experiencing the same problems as you.
 
However, Audio Path should only need to indicate "Encode as DSD in DoP format" since that indicates that the output is being resampled to 2.8MHz.

[thediscman]

I thought that the resampling stage in the Audio Path was an indication that audio data was being resampled to the proper PCM rate (352.2k, 705.6k or 1411.2k) prior to being "Encoded as DSD".

Of course, the output is the same whether the resampling is explicitly stated in the Audio Path or not, but I hope they'll show it.

[6233638]

Why resample in two stages if it can be done in one?

[thediscman]

Why resample in two stages if it can be done in one?

I agree with you. I only want one resampling stage. You and I have a different understanding of what "Encode as DSD in DoP Format" means. 

I don't think of "Encode as DSD in DoP Format" as being resampling. I think of it as conversion from PCM to DSD - only. I think that the PCM data has to be resampled to 64 bit-352.8k PCM (for DSDx1) or 64 bit-705.6k PCM (for DSDx2) and then after that it can be "Encode(d) as DSD in DoP Format".

I got that idea from here: http://wiki.jriver.com/index.php/DSD_Format

The DSD to PCM conversion process converts from 1-bit DSD to 64-bit PCM at 1/8th of the sample rate. The total amount of data from this conversion grows by 8x, so the process is effectively lossless / perfect.
Once you have PCM, it will be 64bit @ 352.8 kHz for DSD, and 64bit @ 705.6 kHz for DSD 2x.

So I definitely don't want two stages of resampling. I just liked seeing whole process spelled out in a bit more detail in the Audio Path. If the Audio Path don't show the resampling stage explicitly, it won't effect the sound. It's just what I'd prefer.

[6233638]

DSD is 2.8MHz
64-bit 352.8kHz is the internal format that Media Center uses when converting DSD to PCM.
 
So if you are encoding to DSD, it has to be upsampled to 2.8MHz.
I don't see the point in the 352.8kHz step if you have to upsample that to 2.8MHz anyway.

[thediscman]

DSD is 2.8MHz
64-bit 352.8kHz is the internal format that Media Center uses when converting DSD to PCM.
 
So if you are encoding to DSD, it has to be upsampled to 2.8MHz.
I don't see the point in the 352.8kHz step if you have to upsample that to 2.8MHz anyway.

I normally use DSDx2, so it would be 64-bit 705.2kHz for me.

But sticking to DSDx1, going from 64-bit 352.8kHz to 1-bit 2.8MHz isn't actually upsampling. The amount of data for for the two is identical. That's the point of the 352.8kHz step. You have the same number of bits in PCM as you would for DSD. Then all that's left is the encoding step.

Perhaps someone who knows for sure can tell us if that's the way the process works. I'm not certain.

[thediscman]

The biggest advantage to explicitly stating the sample rate conversion is the assurance that it's being done correctly.

I would not have known that version 19.0.152 was converting to the wrong sample rate if the sampling conversion was not shown in the Audio Path. It's one of the things that's happening to the signal, and "Encode to DSD in DoP Format" doesn't provide you with that information, especially given that there are 3 different DSD output rates available now.

[Matt]

I think this should be fixed with the build that just went up.  Feedback appreciated.

[6233638]

I think this should be fixed with the build that just went up.  Feedback appreciated.

The last ~20 seconds of a track play extremely quickly when going from a low sample rate to a high sample rate, or if you skip tracks during playback.
The same thing happens in reverse - the last 20 seconds or so of a high sample rate play very slowly when switching to a lower one.
 
I normally use an 0.4s gapped fade, but it's extremely obvious if you set it to something like a 10s smooth cross-fade.

[Matt]

The last ~20 seconds of a track play extremely quickly when going from a low sample rate to a high sample rate, or if you skip tracks during playback.
The same thing happens in reverse - the last 20 seconds or so of a high sample rate play very slowly when switching to a lower one.

I'm not able to reproduce this.  Tips?

[6233638]

I'm not sure what to suggest.
Memory Playback seems to have no effect.
I disabled all DSP except Output Format and it's still happening.
I'm using 1xDoP output to my DAC over ASIO. (also happens via WASAPI Exclusive)
 
All I did was create a playlist which included:
16/44.1
1/2822.4
16/44.1
24/96
 
Switching between them, especially with a long fade duration (gapped fade is even better than crossfade) causes them to speed up/slow down as the tracks change.

As you would expect, switching directly between the two 16/44 tracks is fine. It's only when switching sample rates that it's a problem.



P.S. Doing this test - since I picked a handful of tracks at random - made a great case for volume leveling. Things are horrible with it turned off!

[Matt]

I think this may be fixed next build.

Fixed: Cross-fades could play incorrectly between varying sample rates.

[thediscman]

P.S. Doing this test - since I picked a handful of tracks at random - made a great case for volume leveling. Things are horrible with it turned off!

I very much agree with what you said about volume leveling. I had no idea how much it would improve the listening experience.

Recordings that are quieter (and often consequently have more dynamic range), played back with normalized volume often sound fantastic, and much better than they would without the leveling. Of course the recording is the same, but the perception of it is different.

For anyone who hasn't tried Volume Leveling, I recommend that they give it a shot.