Re: [wsjt-devel] FT8 decoding sensitivity: WSJT-X vs. JTDX

[James S. <kd...@gm...>]

.wav files are pretty big and it may cause issues for people with limited bandwidth internet connections who participate via email.  Perhaps host the .wav files via a third party site like SoundCloud, DropBox, or the like?

73,

Jim S. 
N2ADV (ex KD2BIP)

> On May 2, 2019, at 4:27 PM, Topher Petty <ai...@gm...> wrote:
> 
> I tried to post wav files and a screenshot to back up my observations, but the message was rejected by the moderator.
> I do hope someone will look at the files, and they didn't just get lost to the aether. I'd hate to think the information I attempted to provide would go unused.
> 
> 
>> On Thu, May 2, 2019 at 1:12 PM Rebecca Milligan <reb...@co...> wrote:
>> Hi, is it possible that the clock is off on the pc generating those signals by enough to cause this same issue?  The example sent earlier showed a .8 second difference.  When it takes several times for my pc to decode a strong signal, it is usually due to their clock being off or two signals on top of each other. 
>> 
>>  
>> 
>> From: Topher Petty [mailto:ai...@gm...] 
>> Sent: Thursday, May 2, 2019 12:53 PM
>> To: WSJT software development
>> Subject: Re: [wsjt-devel] FT8 decoding sensitivity: WSJT-X vs. JTDX
>> 
>>  
>> 
>> I've noted situations similar to this.. though I'm not running JTDX. There have been multiple times where a reply was strongly visible in the waterfall, yet nothing was decoded on that AF, and it took another TX/RX cycle to get the reply to decode a message.
>> At first, I thought it was because I was operating on an RPI (Pi 3B), so I copied the configs over to my T410 (twin Xeon X5670s with 64G RAM), and the same behavior occurs. even at less than 10% CPU utilization.
>> 
>> I don't believe it's available processor power, or RAM availability. With only WSJT-X, TQSL, and Chrome (six tabs open) running, the T410 has more than enough processor (24 threads) to handle these decodes, yet they're not being decoded.
>> 
>> I'm not judging, here. I'm only bringing information to the discussion.
>> 
>>  
>> 
>> 73 de AI8W, Chris
>> 
>>  
>> 
>> On Thu, May 2, 2019 at 11:09 AM Bill Somerville <g4...@cl...> wrote:
>> 
>> On 02/05/2019 15:56, Black Michael via wsjt-devel wrote:
>> > Well yeah....but I would assume somebody would want to look at a 
>> > signal that seems to have sufficient SNR to decode.  Not enough info 
>> > in his snapshot to tell where JTDX decoded it or what else may have 
>> > happened.
>> >
>> > Seems to me a worthwhile exercise to ensure something pathological 
>> > isn't going on...all the reports of "loops" could just be overlaps, 
>> > fading, whatever....but worth a check isn't it to determine cause of a 
>> > few examples?
>> >
>> > de Mike W9MDB
>> >
>> >
>> Hi Mike,
>> 
>> absolutely agree, but I just wanted to make it clear to all that that 
>> signal strength is not the only parameter that determines successful 
>> decodes.
>> 
>> With respect to apparent anomalies, yes .WAV files are indeed valuable 
>> when linked to specific issues. For example your files provided 
>> Yesterday have highlighted a potential issue with transmissions of 
>> changed messages with respect to how quickly WSJT-X currently starts 
>> rendering the new message in the audio stream. This is something that is 
>> particularly critical with FT4 due to the transmission duration and 
>> thinking time approaching the limits of operator mental and manual agility.
>> 
>> 73
>> Bill
>> G4WJS.
>> 
>> 
>> 
>> _______________________________________________
>> wsjt-devel mailing list
>> wsj...@li...
>> https://lists.sourceforge.net/lists/listinfo/wsjt-devel
>> 
> _______________________________________________
> wsjt-devel mailing list
> wsj...@li...
> https://lists.sourceforge.net/lists/listinfo/wsjt-devel